Home > Technical Topics > Strategy Development > Pairs Trading Strategy Model

Pairs Trading Strategy Model

  1. I have been lately introduced to pairs trading. It sound like it is a great strategy to trade uncertain bear market in a trading range. I am looking for resources to learn more and develop my own strategy/model. Please help!:confused:

    Thanks
     
  2. There are articles in past issues of TASC - that's a good start...

    Also, don't forget to use the search feature on the upper right hand corner of this screen.

    nitro
     
  3. This is what I did so far:

    1. Zurich Capital report on Market Neutral Hedgefunds - up to 120% annual return with very high Sharpe Ratio. Sounds good. The question is what kind of model do they use? Is it technical model based on correlation, or is it a fundamentals based long/short model? Any thoughts?

    2. I found a site www.pairstrading.com. Sounds like these guys are doing a pure technical play based on 2 year correlation analysis. Looks promising. Subscribed to their free trial. Will see what happens.

    3. From the logistics stand point: what is the best market data source (inexpensive, reliable, clean data)? I want to start with correlation.

    Thanks for any input.

    Al
     
  4. Checked out pairs trading site, signed up for trial period to see what they do but can't get any response from them. Calls and emails have gotten me nowhere. Are they still in existence?
     
  5. Maybe one of their pairs did not revert to the mean and they blew themselves out.
     
  6. I have no problem with www.pairstrading.com so far. I continue to receive my updates daily. I use Qcharts to study the recommendations. This is the only affordable charting service that lets you plot the ratio(stock1/stock2) that I now of.

    I am trying to figure out their model. I know it is correlation based. I am able now to calculate two year correlation myself.
    Now the question is what else to use to pinpoint the exact entry and exit point? By the way, they could not have been blown out of a pair because the use stop losses - rule #1 for any trading system. Anyway, I figure that a combination of RSI and 2SD delta from 40MA should do at first. Then I'll try to add some more TA indicators.

    How can I get put/call ratios for S&P stocks? The idea of adding a sentiment indicator to a technical model is appealing.

    Will keep you posted.

    Neutral Al
     
  7. does any one know how to figure out ratios on pairs... like i have been told mrk to lly is a 5 to 4 ratio... how is this fugured out? and how often will this ratios change?? thanks in advance..Rocky
     
  8. Just do it in dollars. Ideally, you want your total dollar value hedged, or paired, in this case.

    MRK is trading at 49.24. So, for 1000 shares, you have $49,240. Now take the $49,240 and divide it by $58.05 or the price of LLY, and you get 851 shares.

    So, you can trade 1000:851 or some multiple or fraction of that. You can even do some rounding off and you'll see that's in a ratio of approximately 5:4.
     
  9. The ratio is 0.8244 at 14:22 today. AGN is a much better hedge/pair for MRK. The correlation over the past 2years is much higher. If you still want to trade MRK/LLY, I would suggest to wait for R=.8 or lower.
     
  10. Take a look at www.2hedge.com
    This is an aquaintance of me and I'm using the software with good results.

    If you have questions, PM me

    Greetings
     
  11. Neutral_Al,

    I'm just curious: Are you calculating your correlations from daily bars or some lower resolution?

    I suppose this goes back to the basic premise of pairs trading. Some traders will want to look for pairs that are correlated, or revert to mean, on longer timeframes such as by end of day, and others will want pairs that might trade faster, i.e. correlated better on hourly bars. Certainly, correlation on these smaller timeframes does reduce the size of the average trade you'd expect.

    Does anyone have any pairs experience that would say, "use daily correlation as this produces better intraday moves" or instead "use shorter timeframes for correlation to get the most tradeable pairs."

    Thanks in advance for any insights!
     
  12. Anytime I hear a new name of old thing (spread trading in this case) I wonder how long it will exist.
    There are correlations. It`s nothing new. But very few of them are reliable. Even fewer of them make any money.
     
  13. I've been watching this site for past two months. Following their recommendations I would be up 22K on 500k account but I don't have 500K account

    good luck.
     
  14. That is always a problem ...
     
  15. ChrisM,
    You are right, most of the correlations do not always work. What I am trying to do is to find a pair of stocks where the price ratio predictably oscillates around a certain moving average. I calculate 2 year correlation which tells me that if this pair of stocks behaved the same way for 2 years, there is a high probability it would revert to mean again. I do not plan to trade correlation, I just use it to increase the probability of a profitable trade.
    As for making money, it is very much possible. If hedge funds make money trading pairs and PairsTrading.com people make money, so can I.

    Al
     
  16. I am very new to Pairs Trading, even though I have been trading the conventional way.

    I have a few questions:
    After you get the correlation factors, do you look at the fundamentals of the correlations of how they relate? Do you look at each individual stocks after to gain more edge? Is the identifying criterias similar to options and the designated product?

    Also, is there a book out about this? I'm completely new to this and regardless of actual trading the strategy, I'd like to keep the topic in my shelves.
     
  17. There are two books that I know of with pair trading chapters. The first book (of which I am author) is Professional Stock Trading (ISBN: 0971853649); this covers intraday pair trading and is based on volatility hedging, e.g., using Tharp's Percent Volatility Model to equalize position size. The other book is The Complete Arbitrage Deskbook by Stephan Reverre, which covers multi-day spreads.
     
  18. Mark-

    I enjoyed your article in TASM. I found the information very useful, especially how you used volatility. Thank you for your straightforward presentation of the subject. I was able to build a trading system based on your ideas (and a few of my own) which has proven to be my most robust system over the past few months.

    For those of you thinking about spread trading based on correlation, how you treat volatility in your trade selection and execution is fundamental. If you treat spread trading as a linear distribution of data, you will likely lose money. Picking out stocks to spread and looking at a simple ratio is only part of the process.
    Would suggest reading Mark's article or some of the other books referenced here with a keen eye as to how volatility is treated.
     
  19. Neutral_Al ,

    You are right about that. The only problem is that You never know when this correlation stops working. Many of them live for 2-3 years period, some much longer.
     
  20. The solution to this is very simple but efficient.
    Take a look at www.2hedge.com

    The 3rd screen from te TBonds TNotes spread shows a screen on which you can set your trading parameters. On the right side you see (only half) your trade limit / stop.

    This means : the software takes into calculation it only enters a trade when current correlation is above xxx or does get you out of your trade when correlation drops below xxx.

    Hope this helps
     
  21. Hi WDGann,
    The ideal pairs trading system would incorporate some elements of the FA, especially the quantifiable ones. It would be much better for your good night sleep to be short a "bad" stock and "long" a good one. However, if you prioritize FA over TA and correlation, you end up with long/short portfolio. I'd like to base my model on correlation simply because the idea that the 2 stocks of choice have been doing the same thing for a significant period of time gives me comfort.
    If you are looking for a book, I just got my Pairsrading manual in the mail. Haven't read it completely yet, but it looks like they have all the formulas I need. Go to pairstrading.com to check it out.
    Hope this helps,
    Al
     

  22. Hi CHrisM,

    Correct. However, you are not interested in one particular correlation ex: ge/ msft. You trade 10, 20 at a time. One stops working, so you take a stop loss. The idea is that statistically you have higher chance of a profitable trade based on correlation. People at Pairstrading.com claim 80% and show the record. I still have to verify that for myself. The numbers though could be different because I am still not sure about entry/exit points.

    Will keep you posted,
    Al
     
  23. Al,

    I trade myself also T-Bonds, and their correlation to inflation figures make sense. But what`s the idea of ge/msft (except trading in one boat related to broad index) ? I know that market is a living organism and has no respect to any logic, however, can create things which can vanish in seconds.
    But You`re right about statistics - that`s all about.
    I`ll try this (ge/msft) myself on our machines. I`ll let You know what I find.
     
  24. ChrisM,
    I used ge/msft as a general example. Actually, the correlation between these 2 stocks is only 28% according to my calculations. The stock that correlates with GE the best is TMPW (87%). It is interesting that AOL correlates with GE @86%. If you run it on you machines, let me know if you get same results.

    Thanks,
    Al
     
  25. Hi Dave,
    I use end of the day data from TC2000 right now. It takes forever and it is all manual. I need an automatic source with real time data stream so I can try to attempt to build a daytrading system. Any thoughts?
    Here is a pair that I think is an ideal trading vehicle for 2-7 days trades: FO/ITT. Look at the daily ratio chart with 40DMA. I opened
    a position today at 13:50 long ITT short FO R=0.836.
    Let me know what you think.
    Al
     
  26. al..... there are better stocks to pair GE agains ... there just not one to one ratio.... i trade ge against Honeywell... or tyco... dont trade it against aol you will get killed...at least stay in the same sector or news will kill you... tyc and ge has been good on a 2:1
     
  27. May I ask how you found that pair? Seems very interesting?

    Thanks Royce
     
  28. Hi Mark,

    Thanks for the reference to the books! I'm looking forward to reading yours. (It's on its way from Amazon.com). I enjoyed your TASC article as well.

    I'll post some follow up responses and plan to add to this thread with some thoughts, ideas, or discussion about your book.
     
  29. Hi Al,

    As you can see, I'm just catching up on this thread. Sorry for the delayed reply. While Tradestation has its liimitations, as Mark Conway pointed out in the TASC article, I have done most of my systems work in EasyLanguage. I think that Wealth Lab, Neoticker, and AmiBroker might be helpful in your systems, as most will interface directly with eSignal or Quote.com and will manage two data streams or more. I don't use these packages, so I'm not clear on the exact benefits of each.

    What you ultimately select will depend on your programming propensity. I am very interested in comparing EOD correlations with 60 minute, 30 minute, etc. to determine whether different timeframes can help me be more selective based on the types of pairs arbitrage systems I want to run. i.e. a shorter, faster trading system with lower average trades is probably better selected using a shorter time compression for your correlation runs.

    If you know Visual Basic, or C, or some other programming language, it's possible to run studies on nxn spaces in fairly short periods of time. That is, to run the correlation of each of, say, 600 stocks against each other. I plan to do this in Visual Basic for Applications using Excel. I'll build my system directly in Excel, and not only run the correlations, but also record the system results for each pairing. I'll be looking at that resulting data set in many ways. This may not be the direction for everyone though.

    I'll take a look at that pair that you suggest. Thanks for the info!
     
  30. Hi Royce,
    I found this pair by calculating 2 year correlation and adding some strict parameters on 40DMA over 6 months. The book that I bought at Pairstrading.com helped a lot as well.
    I understand your concern about FO / ITT . One is a furniture business and the other one is a waste manager. I was surprised myself. But think about it, if these two companies have been doing the same thing for 2 years, there is about 90% probability they would do it again (based on my S&P500 testing so far). Plus they all are within S&P500 which imposes certain correlation as well.
    Now, if you want a pair within the same industry, how about Fannie Mae and Freddie Mac (FNM/FRE). They look so good to me that I am opening a position tomorrow at the open.
    FO/ITT is making money so far. Will see what happens tomorrow.
    Al
     
  31. Al,
    Thanks for the info.
    Royce
     
  32. Hi Dave, thanks for the kind words, and it's great to be on board here.

    Mark
     
  33. looks interesting.
     

  34. Are you covering some of your position this morning (FO/ITT)? Either way ->Nice call

    Later Royce
     
  35. Hi Al,

    How useful has the book from pairstrading.com been for your trading? Is it a good intro, or more of a detailed approach to stat arbitrage? Any thoughts, positive and negative would be appreciated.

    Publicly or PM me if you'd rather. Thanks!
     
  36. i trade pairs on intraday basics. i was right in about 80-85%. before-i may left position for overnight. usually-greater risk-greater reward. dont keep for more that two days-if treid did not going in you way, exit or it will kill you later.
    since everything so great, i don't have stop or proper exit point)) sometimes-i just killing me. working on it.
    usually-i just take 1% and run away)))) can tell- you easy can make 2-3% intraday on nasdaq stocks. problems-damm stocks so cheap now, you have to spend big good chunk of you profits on commisions. my point-can you make 10% in intraday in directional trading? yes....15% ? yes.. but-if you can make so much in one day-it cannot be easy. can you make 100% in one day-yes (in options for example)-go try make it....it may happends once in 10 years. but in pair-you create kind of limited risk-limited profit position and from statistical point of view-you have much better chance to make money.
    sorry for bad english
     
  37. whats up with your english...lazy or foreign?
     
  38. Don't make it more difficult than it is!! It's great hedging strategy. Here are some ways I incorporate it into my trading:

    Late afternoon leader/laggard play:

    Take a look at the days leaders and laggards. As an example - let's say AMGN is looking strong at midday (up 3 %). Look for other biotechs that have poor charts and look like they may break down further. Use one of the laggards as your other half of the pair (let's say GENZ for this example).

    Take one leg of the pair at a time. In other words, enter AMGN long at what you would consider a good entry point. If it starts gaining immediately, put off getting into the second leg of the pair. Don't use it until you have to!

    There are 2 ways to get into "leg 2" of the pair:

    1) If you see the GENZ setting up as a great short play anway -- pull the trigger. You're now in your pair. Figure out what your "spread" is and set up a chart w/ your 2 stocks and the realtime spread.

    -OR-

    2. If your long (AMGN) starts declining, put on the second leg at that time.

    This afternoon startegy is a "high probability" play. Intraday leaders generally tack on additional gains as the day progresses. By the same token, intraday gainers tend to "decline less" than losers if the market heads south.

    Other notes:

    Some people prefer to put on the short side of the pair first (due to the uptick rule).

    Exit the pair when the current spread meets your profit goal (or exceeds your loss tolerance).

    Don't waste a lot of time worrying about "historical correlations", etc. Why pay some website $200/month for them to tell you that KLAC and AMAT would make good pairs!

    I typed this very fast so I may have left out a few details. It's Friday -- time for Happy Hour!

    GD
     
  39. Hey Royce,

    Exited FO/ITT today at the close R=0.786 with 5.87% profit.
    Not bad. Now will wait for reverse.

    Al
     
  40. Hey Dave,

    I think the book is pretty good. It is written by professional HF managers. The perspective of a pro on pairs trading is hard to get these days. The models that hedge funds use are ether proprietary or cost a fortune. These guys kind of open up on a subject providing conceptual and technical info on how to trade pairs. They even talk about trading as a business and market/trader psychology. I find that entertaining and helpful. I actually built the correlation analysis portion of my trading system-to-be based on their formulas and concepts.
    I am still straggling with entry/exit points however.

    Al
     
  41. what about if amgn start move immidiately against you? (what usually happends))))you will need couple more sec to buy sec. leg. at this time you will be in about 0.5% loss. what you going to do in this situation? how big the difference between two stocks can be considered big enough? just trying to clarify few details..
    thank you!

    p.s. i agree about hist. correlations))) some of my pairs been good for couple years, but in last few month,as i mentioned somewhere before-trends start show up. trends in totally different directions for 20-30, even 50%....good stocks,same products....but...that why i stop hold them for more than one trading session.
     
  42. did you ever get into the FNM/FRE position- i have been holding it, taking some heat, but i will NOT let the market and the sucker moves shake me out- in fact i have been making money trading the good moves (50-70 cents!!) to pay for holding the position (well, some of the cost) Its down to a 2 year low and the pair usually has a spread price of around 11 dollars. so, historically it is a very attractive buy with the very realistic probability of it reverting to the 11 average spread price. the only thing that concerns me is the worries about FNM and interest rate risk being higher than that for FRE- (the other worry is that MER is telling their clients to sell their FNM and buy into FRE- you can obviously see stronger buyers in FRE intraday than FNM- but you can make good money when FNM gaps up and prints it.) so FRE isn't getting whacked as hard as FNM- FNM makes crazy prints all day and has been hard to hold in a 1:1 ratio last week- i have switched from trading it 8 to 10 to 1:1 because if the market rebounds next week, which i feel we will get some upside retracement from the previous lows, then FNM will rebound more significantly to the upside than FRE. both stocks are in the EXACT same business and i don't think that FNM has significant risk over FRE- so it is stupid that people are selling FNM more heavily- i feel the pair will return to fair value and i've been holding it (3000 x 3000) when it reverts back to the historical mean, i will make a nice profit, but unitl then, i feel confident in holding it (3000 both) and also scalping the good moves on 1000 shares to make some decent gains. what are your thoughts (or anyone else) on this pair???
     
  43. I have been trading FNM FRE on and off for a while. I always trade it in a ratio of 4 FNM for every 5 FRE, or 6:7. I always start with 400 and 500 shares respectively, and either take a profit intraday if it goes my way, or add to the position (in the same ratio) at key intervals if it doesn't go my way.

    I too noticed FRE becoming way out of line with FNM and couldn't resist taking a position in FNM/FRE late on Wed the 18th. It didn't go my way initially as I expected, so I decided to add near the close and take it home. The next day, FNM gapped down and FRE was slightly up - I was down $1200 on the open, which for me is a little more then I am willing to risk on a pair trade. I didn't panic and managed the position the best I could to try to contain the damage, as I will do sometimes, if at the point of extreme pain, I step back and realize that if I were looking at this trade afresh, I may take it in the same direction as the direction that I am currently in. Well, I managed to recoup about $400 on the trade, but then I saw that when FNM started to go up, FRE would be right there with it. I got out because I do not like to be in pair trade where I lose money on one side of the trade when both stocks are going down, and at best break even or lose money when both stocks are going up [or vice versa] - to me and my time frame on these things, that is the definitive answer - you are in a bad pair.

    The next day there was a terrific article in the WSJ about FNM and FRE. If you haven't already, I suggest you read it.

    I commend you for the balls and conviction that you have in staying with this spread with the size that you did on it. In fact, as I look at FNM/FRE now, I would be looking to get in again real soon - just not here.

    One last thing - I do not know what the ultimate answer to holding a pair is in relation to risk. For example, everything in my being says that this spread has to come back in line, but I feel that the risk of it staying away from par is just too great at this point. On the other hand, I know of an example where Bob Bright held MRK/LLY as it went against him, continued to add to it with the tenacity of a Bullmastiff until he was down $2,000,000 on the pair, only to make $4,000,000 in one day when LLY cratered - but only month(s) later. I do not have this kind of staying power or this much risk tolerance... :(

    nitro
     
  44. Hi all,

    I did open a position on 9/20/02 at the open R=1.104. So far it is a loosing trade. I decided to hold it over the weekend. The question is what do you do next?
    a) Do you hold it till it becomes a winner? If you do, do you add to your position? If you do, when? And how do you know if it becomes a winner at all?
    b) What is the risk of holding this position? Does it become greater or smaller as the pair moves against you? Is there a way to calculate risk/reward ratio for a pair of stocks?
    c) How do you decide when to exit the position? Is is a time parameter, max allowed loss, technical indicators?

    Please let me know your thoughts. These issues are the most difficult ones. If we could solve them we would be on our way to a pretty good trading system.

    Al
     
  45. hi!
    from my expirience-the longer you hold-greater risk you take-it basically obvious. most of the time-it not worth it. i would say if it did not go in you direction in monday-close it @ close. i know, there is a great correlation on those stocks and it very unusual to see such difference but......BW-why YOU don't have plan, when you enter into this trade? you MUST have it. you must know-how much risk you willing to take, how much profit you are looking for and so on..or you just enter, because somebody else enter?

    Good luck!
     
  46. These are extremely difficult questions and have no answer that is too particular to statistical arbritrage. If you are thinking that pairs trading violates many of the traditional cliches of trading maxims, I agree. I often feel, after having done hundreds of pair trades [if not thousands now] that the ONLY holding period for a pair like FNM/FRE is until it becomes profitable or forever, whichever comes first, while adding to the position as the the pair goes against you. This is so contrary to intuition, but it is probably the only way to trade pairs on the larger time frame [I have not thought about the mathematics of the R/R aspects of trading this way - it is just an observation that I have made countless of times when I got out of a pair at extreme pain, only to have been vindicated days later - and watching traders like Bob Bright do these things...]

    The problem is that you need balls from Chicago to New York to stay in this slaughter...It is for this reason that the best traders seem to blow their accounts once or twice in their trading lives...

    THERE IS NO TECHNICAL INDICATOR THAT WILL GIVE YOU BALLS AND CONVICTION - THE MARK OF THE GRANDMASTER...


    nitro
     
  47. soon it will be 4 fre to 5 fnm lol
     
  48. yeah- lol FNM is getting whacked.... but this ugly bitch of a pair will make me money- but hard to hold.......

    anyways, what do you all think of trying to set up a forum for pairs trading or if we all were to do a pairs trading chat sometime after the close and we could all talk strategy???
     
  49. I've never done any pairs trading but it sounds interesting, particularly in this market. I am struck by one thing however. Pairs trades seem to be the reverse of the normal hedge strategy of long/short, because a long/short trader would be long the stronger stock and short the weaker. A pairs trader would seemingly be in the opposite position, looking for a reversion. Am I missing something?

    A good example might be JPM/WFC. One crap bank, one solid bank. I haven't looked at the spread but it must be wider than normal by a long shot. Certainly, I would be nervous about this trade, so nervous that I might want to be short JPM, long WFC. So how do you factor this sort of analysis in? Or is it irrelevant?
     
  50. Hi AAA,
    Pairs trading and long/short strategy are both hedging strategies. If you trade/invest primarily based on FA, then you have a long/short spread (long a stock with good fundamentals, short the stock with poor fundamentals). Pairs trading, on the other hand, is a form of statistical arbitrage where a trader capitalizes on price action inefficiencies in a pair of stocks which are manifested by divergence from a historical mean. It is based on statistics, math, and TA, less so (if at all) on FA.

    Historical correlation and s/w predictable price action gives me as much comfort as being long a good stock and short the bad one gives comfort to you.

    I just finished testing several combinations for entry points. The best results are produced by 3 SD from the 6mo mean with 30>RSI>70 for 2 sessions. Ex: D/ED. This pair has 86.475% probability to converge towards the mean within 2 months.

    By the way, I found FNM/FRE recommended for today at the open in my Sunday update from Pairstrading.com. Gee, I guess I am doing something right. The question still remains, however: How come I opened a position last week and have been loosing for 2 days, and these guys recommended it today? I guess their model told them to wait on Thursday and Friday.
    It's all about entry and exit points.

    Does anyone have any experience with a combination of indicators for entry/exit points with good results?
    Any thought is appreciated. I'll run statistics on all of them.

    Al
     
  51. i am an experienced pair trader and in my opinion the entry/exit arent nearly as important as trading the pair. the first time you put the pair on isnt when u make your money, its usaully the 2nd or 3rd time youve added to it. by trading the pair it can be 5 points against your first position and you could be up a few thousand in it. by trading it you improve your price. the market starts to sell off dump your long and buy it back later/cheaper and ride your short. i dont know how the rest of you pair trade but when i put on a pair i am always long the weak stock and short the strong one. you fight the trend. the pair wouldnt be a good price to put on if this werent so. look at fnm fre why do u think thats a good price right now? because fnm has been getting pounded. anyway gl all
     
  52. What about trading the trend of a spread? From my limited pair trading experience and watching, it seems that you can get some great trends to follow. HD/LOW has been in a beutiful trend for weeks now. This would be the opposite of looking for convergence, but would seem to work too.
     
  53. Hi trendsetter,

    You're describing what the majority of pairs traders do: you find a correlated pair of stocks and trade it increasing the position with increasing deviation and decreasing it with reversion. This could and does work, but what happens if you continue to double down against the trend? Look at AMR/DAL. This was a nicely correlated pair until 2mo ago. If you doubled your position anywhere within 2 months you'd be flipping burgers now.

    I am trying to develop a different approach where you treat a pair as a unique occurrence, not a trading vehicle. You choose from hundreds of pairs today the ones that have the best odds to become profitable in the near term. Having well-defined entry and exit points will prevent you from blowing your capital on something like AMR/DAL.
     
  54. This is a variation of a pair trade - you are simply playing the pair to "diverge."

    This is a perfectly good way to trade - I have noticed that some traders prefer convegence, others divergence. They may even both make money on the same trade being in it at the same time - they may even have bought and sold to each other! The exit is the difference for each, and the time frame...

    nitro
     
  55. This is the essence of the way I trade pairs.

    nitro
     
  56. Pair trading works until it does'nt . As you gain success and experience in pair trading, you tend to double up , take more size as it strays from the mean and at the point that you think the rubber band is about to rebound back to the mean is usually your point of capitulation plus one more tick. If if reverts you make a killing BUT if the band breaks (witnes fnm/fre and HD/LOW ) you lose months worth of profits. The trick is the discipline to not be at that point of capitulation - to trade the pair and not be at the praying stage when you are +5000 / -5000 or whatever size you are comfy with. One day I will get to that point. Read up on Talebs book called "fooled by randomness' great point about strategies that work until they don't and secret is in knowing that it is not working and get out. Best of luck to the pairtraders out there.
     
  57. i agree gat trader, that is the trick make enough to cover your hits and still have a profit. find pairs with a small range most of the pairs i play have less than a 10 point range over 2 years. and i dont keep doubling up well i do but if i am getting crushed and i have my position up to say 10000/ - 10000 intraday i wont hold that amout over night unless i have a good reason. i may hold a few thousand and take the hit on the rest. i will make it back tommorow.
     
  58. At what point do you cut your losses if it keeps moving against you? Anyone who pair trades care to comment?
     
  59. When you get a call from the LLC main partner which usually means you gotta come up with more $ in your cap account. That is why it is so important to know your capitulation point and get rid of it before the band breaks OR use options as insurance.
     
  60. i think you have to decide what you are willing to risk. how much pain can u take. if there is news on either of your stocks that makes a difference. you should evaluate your position to the best of your ability. news is really the only thing that effects your pairs drastically. i remember i was short cof and long pvn i put the spread on at 18 then over a few days it went to 24. a few days later it came back to 18.5 and i took it all off. good thing
     
  61. I analyze it from a risk/reward perspective, i.e., if I expect the spread to revert to the mean 67% of the time, then I'll implement a stop loss that is equal to the reward of the spread. So, size of loss equals size of win, but if you win 2 out of 3 times, then you're golden.

    MC
     

  62. Hi Mark,

    How do you find the spreads you use and how do you determine the percentage number?

    Al
     
  63. Al you do a pretty good job yourself of finding spreads (e.g. FO/ITT).

    How do you find the spreads? Last time you told me correlation...is there something more?

    Royce
     
  64. market topology used to be a good place to find pairs but they started charging for the site. a good way to find pairs is get a list of all the stocks in the s&p 500 that do over a million shares note the sector for each symbol. then overlay the stocks in the same sector on a chart to see how they track. go back a few years and check the range for the spread.
     
  65. are you using a website for charting?

    I've been using Yahoo, and also

    http://stockcharts.com/charts/performance/


    thanks :cool:
     
  66. Hi Royce,

    Currently I am testing several systems. All of them are based on correlation analysis. However, some of the constrains are different. For example, FO/ITT and FDC/GWW belong to the "flat" 40dma system. Something like DLX/ODP is more trendy but has better risk parameters on a longer term. The reason I ask people questions is because in order to come up with a great system I need as much help as I can get. Any idea is welcome...

    Al
     
  67. Thanks for sharing the link, this is value added.

    Royce
     
  68. Closed FNM/FRE today at R=1.137 with 3% gain. Could not wait any longer. It is still quite far from 40dma. Working on my timing...

    Al
     
  69. WTG!!

    I got in earlier than you I think - I don't have the stomach for that ride...

    Congrats,

    nitro
     
  70. how long you been in this position?
    i look @ pairtrading.com performance-they far away from my results....

    http://www.angelfire.com/de3/options/TRADES.HTM

    looks like they put at least 20000 on each side to get 500-1000 profit. man..........you can make it easy on nasdaq in intraday.
    imagine, where i would be, if unstead 2500-5000 on each side i put 20000)))))))))))heh....
    one big drop on news will kick you out for month, if not forever.
    i been trying to hold some positions overnight-it's nightmare.
     
  71. Hey Bob,

    I opened the position on 9/20. I looked at your record... Do you have a system or do you scalp general correlation 100 shares at a time? What is your total relative P&L?
    As far as pairstrading.com performance, making 100,000 on 300,000 account in severe bear market over the past 9 months is pretty good.
    They think like hedge fund managers, not like daytraders. So, for them being market neutral and outperforming the S&P by more than 50% is an achievement in my opinion.

    Al
     
  72. yeah, looks like they are funds managers, you right about it, i just go over this site once again)))))). but, they make same money from hold position for days, which create more risk. why i should sit in FRE-FNM for ten days, if i can make same(or more) money between ALTR-XLNX today in just couple hours? i do have system,i also wrote software for pairs,still work on it + prices got so low,commissions killing me. i make about 9000 on 10000 capital for 3-4 month, then lost about 3000 in just few days.as you can see i have few big losses @ end)))))) work on it now.
    basically-all my losses come from me)))))there is nothing wrong with system-it very hard for me to take a loss(((((
    i still not prepared mentally for it.
    Thank you!
     
  73. I've read all the posts and thought i'd put in my two cents..i use metastock to correlate my pairs , searching each night thru various sectors..i tend to trade only same sector pairs but i keep cross-sector pairs also..I've been watching fdx/rtn as a possible trade at r+1.75 but didn't pull the trigger ..today it moves ..for entries i have criteria that has to be met..first it has to be in a defined area of FLAt support(i don't like trendlines too much i.e.r/cnf had a good trendline until now).next the 21 day stoch. has to be way overbot/oversold along with either a 6 day rsi or a 5 day stoch which also has to be extreme..i look at fundamentals also but they aren't always reliable(didn't do much for s/low today)..right now i just started selling ip/tin r=.92..just my thoughts..i saw someone mention a chat room?..i'd be interested ..good luck all
     
  74. Hi galtg,

    Welcome. I am going to plug in your 21 day stoch with 6 day rsi into my correlation. Sounds interesting. Let me test it. If you like IP/TIN, you should love WMT/KSS and ED/TXU. A bit to scary for me. Quite a rare event. Not enough occurrences to produce good statistics. But definitely tempting...

    Al
     
  75. yeah ip/tin loks like something u'd want to keep away from.but i've traded this pair alot in the past with good results ..this was an extrordiary move ..but i figured with tin near a support area and oversold and ip just going nowhere it was time..and besides my stop is placed a point above..both firms pre-announced..i'll risk a point
     
  76. I was looking into trading pairs, and I wrote a simple spreadsheet which was linked to TC2000, and basically followed some spreads that way.

    Here is a link to show what my SS basically looks like. This hasn't been updated lately.

    http://www.geocities.com/joebhernandez/spreads.htm
     

  77. i dont get it
    if you interested-you welcome to see my trades somewhere here

    what your criteria to find a pair? or enter into trade?
    Thank you!
     
  78. bigj,

    that is sweet.

    nitro
     
  79. Here's an updated version Nitro.
     
  80. Hi !
    can you explain what is this?
    SPRD C1 MAX 5 MIN 5 AVG 5 MAX 20 MIN 20 AVG 20 5 SP DIF 20 SP DIF

    also, did you try to build a chart from % difference?
    i did one for you-just pick one of your pairs.
    i hope you (and all other pair traders) will understand, that find a pair, based on correcation can be bad idea. like in example .

    from correlation point of view-great pair!, but, if trade did not go your way-you never get out. unless-you have a stop loss.

    just my word of caution.......
    Thank you!
     
  81. bob..i agree with you about correlation, but in regards to the pair whose chart u posted i wouldn't go into it..i've read the pairs trading manual and in it there is a very good point made..keep away from trending pairs..in pairs trading the trend is not ur friend..a good pairs chart has flat boundaries , where the pairs bounces off..in other words the pairs is constantly reverting back to its mean..
     
  82. i trying to understand, big_jdez picks from his file OR what all this stuff(that i mentioned in my post above) means. also, so far, i did not understand a logic or criteria for entry point in pair trading.
    as i understand it, trader enter into position, because he believe, that difference between two stocks based on his analysis big enough to enter? what time frame most of pair traders use?
    why they use dollar ratio instead of plain % change difference?
    why they always talking about historical correlation?
    from my point of view-best picks will be always those stocks, which make crossovers every day.(i already mentioned it) but they keep talking about correlation. what about DIA-SPY? correlated enough? can you lose money on it? o yeah!!!!! yes you can... i'm asking for pure and simple explanation of pair trading, enter /exit point and i will be more than happy to see picks or actual trades(like i did) and then-we can work together on something.
    also, i'm looking for someone, who may help me finish my own intraday pair trading strategy. i will post examples of it later.
    i need a person, who will perform simple calculations and smart enough to analyze things. i been work on software for pair trading and simply don't have enough time to do all this stuff.
    thank you!
     
  83. Correlations are tricky because stocks that have been historicall correlated can suddenly become uncorrelated. The realization of this unhinging lags due to the nature of statistics and real money can be lost.

    At the risk of giving up some secrets I find its great to pursue pairs where there is both a fundamental as well as statistical correlation. The fundamental correlation allows you to confirm that a spread, indeed, has gone out of whack along with the correlations.

    Example: BKS and GME.

    -- BKS has a $1.4B mkt cap and GME has a $1B mkt cap. BKS happens to own 63% of GME, or about $630M of GME.

    -- BKS is currently at $20, give or take, and they are on target for earning about $1.75 this year. GME contributes about $0.37 of that, leaving $1.28 in earnings from the book biz.

    -- without getting into all the gory details, if you subtract the GME stake from BKS/market cap and subtract the contribution to earnings BKS is grossly undervalued compared to other book companies (BGP), other mass retailers, and the S&P 500.

    -- GME also is overvalued compared to other game retailers (ELBO), other mass retailers (BBY, CC) and the S&P 500.

    Due to the enormous stake BKS has in GME there should be some correlation between the two stocks. That correlation has gotten out of whack (BKS's stake in GME is being valued at 0 by the market).

    What's the play? Long BKS/ short GME. Lets take some extremes:

    A. GME doubles and BKS stays the same. Not likely since now GME is worth more than BKS. A great past example of this situation working itself out is COMS / PALM.

    B. GME stays the same and BKS goes down 20%. Again, not likely since now the public is valuing the book business at zero. The book business has either grown or been flat every year since 1920.

    C. GME goes up 10% and BKS goes down 10%. Possible given the enthusiasm for video games and the reluctance of people to buy other retailers' stocks. In that situation I would buy more of the pair (but more BKS and short more GME) simply because either:
    a. the book business is being valued at close to 0 which is unrealistic when you look at comparables
    b. the stake in GME is valued at negative. BKS then simply dividends out the stake in GME and the stock goes up accordingly.

    I've had enormous success with pairs like this throughout this bear market and even during the bull rallies. I've also written a lot of code figuring out correlations between related stocks, etc. and that stuff just never seems to work out if you base it solely on statistics and completely ignroe fundamentals. Just my two cents.
     
  84. bob..let me tell u what i just tried with limited success..like you i needed to find entry/exit points on daytrading pairs..so i took the daily net change ,irregardless of +/- and did a 5,10 and 30 moving average..as an entry point if the spread moved twice the moving average it was an entry point..mixed results..then i did a 5 ,10 and 30 day standard deviation which i multiplied by 2.5 figuring i was fading the spread towards an extreme..if there was news then i give them more room..i did this on highly correlated pairs mostly in the .90s again mixed results..plenty of good trades but also many that didn't turn..i wonder if this extreme market volatility is affecting results.
     

  85. Hi galtg!

    man.......you been in to it deep enough.......
    but, you missing something in your.......(system?)
    MA or StdDev did not work there...(just my opinion)i think, i getting close to it, and
    if you want, we can do something together.
    and, if lok at my trades-you wil find, that i have about 90-95% positive trades. people just dont want believe in this)))))
    problem-last few trades. if i have stop loss-it will be no problem, but......i did not.

    sorry for maybe bad english

    thank you!
     
  86. do think it's a secret? are you sure?
    i just want to give you some examples of your secret weapon:
    LLTC-MXIM
    and you will see, how deep the losses can be, even if you have perfect picture....
    those companies-damm twins fundamentally and by ANY other criteria, you can check it.
    now-look @ daily % change chart.
    if one stock % change>than other-i add it to previous close % change(opposite-just opposite)

    there is no secret, and there is nothing new..........
    but , i do believe-together-we can win.
    Thank you!
     
  87. That pair is completely different from my example. Yeah, that pair is correlated. And trading that correlation might work. For awhile. But then one day it won't. And the spread will get wider and wider. And all the king's men won't be able to put it together again.
     
  88. exacly. that why i work on intraday. it muck easier to predict price movement. and if you wrong-there is enough pairs to trade.
    FNM-FRE been as perfect example for while here.........but...guys! common! yo ubeen sitting on it for two weeks for 3%?
    huh......such waste of time.........look at this-
    I been riding all day on it, but i want more.......i want whole nine yards....got more than 5% today just on ALTR-XLNX
    and there is.still some room for all of us........
     
  89. vishnu..i've traded those pairs in the past..ltd-ibi is a perfect example only ltd was the stock that was the one being grossly undervalued..rd/sc also a classic example ..im always on the lookout for these but they arent easy money either.rd/sc went to a freefall when rd came out of the s&p..the one u quoted sounds risky.. its been trending lower..where does the trend end?
     
  90. that mean you have to know direction of the market?
    at main opinion-good pair strategy work well in any market.
    Thank you!
     
  91. Hey!
    where is all pair traders?
    working on stops?
     
  92. Does anyone know of a place where I can find lists of correlated stocks?

    (leaving aside impactopia.com; I already used my free trial 3 weeks ago)

    Thanks Royce
     
  93. try this-http://www.bullsector.com/

    but, read my posts about correlation, if you want trade pairs, or try to build chart like i did.
     
  94. Thanks for the link Bob. I especially like the combo quote page where you can see the thumbnail charts of the individual sector components. Thanks again, Royce
     
  95. Can anyone tell me the formula to calculate the correlation coefficient?

    How are you calculating the correlations mathematically?

    What is the formuls used by Impactopia to calculate the correlation coefficient?

    How can I calculate the correlations between various stocks?
     
  96. If you don't know the formula for correlation, you should be investing in a statistics book immediately. Otherwise, a simple web search will bring up a couple thousand examples of correlation coefficients and how to calculate the same. A good, basic site for statistics information is as follows: http://davidmlane.com/hyperstat/index.html
    and specifically on correlation: http://vassun.vassar.edu/~lowry/webtext.html

    Impactopia uses returns in their calculations; (someone correct me if I'm wrong on that) they also use one year of back data to do the calculation.

    And to calculate the correlations you will need data for each stock data series (you can download data for free from yahoo finance) for the lookback period over which you wish to capture the correlation coefficient.

    Royce III
     
  97. You can download data in to excel and calculate correlation in it. The help file in excel explains how to do it pretty well.

    Corey
     
  98. anyone want to help me with long term pairs?
    1-5 weeks?
    as you may see from my posts, i have some expirience in this field)))) comissions killing me in intraday pair trading + i want to try NYSE stocks. also lack of time for all this stuff. also think about website about it.
    BTW-is anyone try to add activity on options as some sort of indicator?
    Thank you!
     
  99. Pairstrading.com
     
  100. Hi, everyone.
    I just got back from vacation. Did not post for a while.
    Opened a new position today: DRI/APPB @ R= 0.7589.
    High correlation, same industry, better news for DRI today. Any thoughts?

    Al
     
  101. Good thing you were on vacation, otherwise you might have been scooping up those .90's and blown out or doubling down at this point. It's all about timing isn't it. And it looks like you may have timed this perfectly.

    I would add, on a fundamental note, that APPB started preaching that their carry-out service was going to be successful and they recently delivered increased same store sales as they foretold. I would expect APPB to remain strong but there's no reason why DRI can't pick up some ground.


    Royce
     
  102. CNBC to have segment on pairs trading in a few minutes.
     
  103. they gave one of the most logical examples that certainly makes sense

    MU - Micron Tech to
    SMH - Semi Cond Holders

    the correlation was very tight, by visual reference, I estimate over 85% (just saw the chart on the TV)

    anyone have other pairs?
     
  104. if you find link to my trades somewhere here, you will find plenty of them:D :D :D
     
  105. well, today was a crappy day for pairs trading....

    LLYMRK, ADITXN, KOPEP, MWDMER were complete garbage
    why the hell is ADI up 3.50 today while TXN, NSM up under $1??? MER's price target is raised, less brazil exposure- yet MWD was WAY stronger than MER.......it seems illogical. and everyone has had a hard on for LLY which was strong while MRK was the weak piece of junk......hmmmm....well, i guess its just the 'ol squeeze....i went a little past my stop losses today when these ones started trending.......:mad: :mad:

    but the real problem today was CVSWAG- i need to find something else to pair with WAG cause CVS is just such a dead stock- it sat in a 15 cent range for most of the afternoon while WAG went all over the place- those kinds of pairs are very hard to trade unless you have the one moving stock really figured out. well, i was just pissed off because WAG, even with 14% better same store sales was downgraded intraday by AGED, and my crappy briefing.com reported it 50 minutes late and by that time, the stock was already getting hit and it was all over. I stopped myself out for a $1500 loss.
    so, all in all, a really tough day for us pair traders, or at least the ones in my office.i hate trading days like this where there are trends with little retracement. but, i'm holding ADITXN and CVSWAG short overnight and with those and the opening plays i always do, making my total day's loss of $1800 back should be relatively easy to do tommorrow- lets hope :D
     
  106. Totally agree with you with regards to CVS/WAG. Still holding 2 legs and will hopefully close them out tomorrow at a smaller loss.:(
     
  107. Al strikes again. Nice call. How far are you going to ride it?
    Royce
     
  108. Closed DRI/APPB today at R=0.8081. Nice profit.
    Opened FO/ITT as PairsTrading.com recommended today.


    Al
     
  109. hallo to everbody,
    I am new to ET and this is my first post. I find the discussion on pairstrading pretty interesting, since I've been doing some research in that field for several months.

    Le me give you a brief overview of what I did. First thing was that I tried to find a pure mathematical/statistical approach to the whole subject. so i took the s&p 500, formed all possible pairs (125.000) and tested a simple startegy using bollinger bands. i did that to have statistical material to play with. so i was able to sort the results by the different GICS-levels. it was interesting then to see that there were sectors that performed quite well, like banks with a SharpeRatio for the whole sector in the area above 1.2. it was no surprise to find that boring industries with rather homogenous companies like oil and financials did best in a test with all possible pairs trading. unfortunately it became obvious that the edge was diminishing at the end of the nineties.
    so we entered next step which was to select the pairs instead of taking all of them.
    we used two hand full fundamental data like pe ratio, priceToBook, sales to debt ratio and stuff of that kind and tried to improve the strategy by using stops, some filtering and take profits. we came to portfolios that seemed not loose their edge. but we still were in low sharpe ratios of about 0.9. and we had several weaknesses in our model.
    first we assumed that the s&p members were the same in the mid nineties as they are now. this is not the case. since 1993 there were about 300 changes within the structure of the s&p.
    second we had a natural survivor bias. bankrupt and unlisted stocks are not included in the analysis.
    third we did our fundamental analysis at the end of the data. so we were kind of prophets for the similarity of two companies.

    now to get rid of these weaknesses we went one step back and gave up the fundamental comparison and said we test weekly which pairs out of the universe of all pairs which are in the same subindustry had the best sharpe ratio in recent time and traded this for the next couple of weeks. doing this we got rid of our fundamental comparison problem. we could test a sharpe of about 1 with this dynamic retraining. still we have survivorship bias in the fundamentals. we try to get rid of that by building a database that contains all changes and all data about all stocks in the sp1500, since we feel it is better to have more stocks for each sector than the sp500 offers.

    the actual reason for me posting here is that i am still unsatisfied with our strategy. i tested quite different things from correlationDependentBands to modifiedBollingers with filters in volaDifferences, shortTermCorrelationFilters, RSI-DifferenceFilters and other stuff. i find it very strange not to be able to efficiently trade highly similar and correlated pairs consistently. it seems that correlation and tradeability do not go hand in hand. at least with my current tools. it sems to me that it is hard to find one strategy - no matter how flexible you build it - that works consistently for longer on a daily basis.

    we are still considering whether we should trade the pairsstrategy i mentioned, since the results seem rather low for the effort put in.

    sorry for this long inital mail. i hope i am right placing this in this thread.

    peace
     
  110. Stocks have a mind of their own intraday for sure. The problem for me has been that the "intermediate" term has now also loosened more than it used to have, causing the pain threshold to increase [greater time to convergence/divergence] and therefore the risk.

    Therefore, IMHO, the "only" way to trade pairs in the "current" environment is:

    1) Trade companies that are not going to zero bar nuclear war
    2) Expand the time horizon in which you are looking for them to converge/diverge
    3) The most difficult thing of all, ADD WHEN THEY GO AGAINST YOU. This is the key and knowing how to do it is tough at best.

    I have also been thinking about modifying the position sizing as it relates to pairs.

    Finally, _TRADING_ one side and using the other as a crutch can be helpful if you notice that they are moving together that day.

    nitro
     
  111. Hi everyone,

    Still holding FO/ITT. Opened a new position today BK/MEL @ R=0.9190. I am planning to sit in it for a while. Despite some fundamental problems that BK had in a recent past, I still feel that the ratio will correct to 1.1 or so. Technically it looks just beautiful IMHO. Tight stop loss though at R=0.8900...

    Al
     
  112. Al are you doing any original research or are you merely taking the pair trading service's trades? Royce
     
  113. Hi Royce,

    I am doing both. I only trade some of the Pairstrading.com's pairs that make sence to me such as MXIM/LLTC, GM/DCX, TGT/WMT, SBC/BLS, XLNX/ALTR, FO/ITT, AGN/MRK.
    I am also testing 2 year correlation with ratio at 2SD from the mean or greater where the graph crosses 10DEMA in the direction of the 50DEMA. A good example would be KLAC/NVLS. Once I am done with that I am going to add some more Tech Indicators to confirm oversold/overbought conditions such as 14D RSI and Stochastics. I'll let you know the results.

    Al
     
  114. hi nitro,
    thanks for your reply. i appreciate your thoughts and feel that you are right in extending the timeframe and add to loosing position, which has to some extent the same effect by smoothing your average entry price.
    what i find so interesting about the whole thing is, that it is not quite obvious why the situation is as it is. i think the market place is an integrated object that cannot violate certain rules for too long without creating extreme inefficiencies. basically there might be two major explanations for the current situation: the bear market distorts the level of market integrity subtsantially, which i doubt since the correlations between stocks do not seem to deteriorate. or there is much capital (not necessarily many people) playing the same game in the same way, thus eliminating the "obvious" inefficiencies. i think the existence of a service like pairstrading.com is an indication in that direction. who did pairstrading fifteen years ago?
    i think that there is much money betting in this area so it might be necessary to explore completely different tools for analyzing and trading. as a person interested in systematic trading i find that extremely challenging since it inherently requires some kind of selflearning systems.

    peace
     
  115. Hi all:

    I am new to this forum, and have found the postings interesting. I have been researching/trading pairs for a while now, and have found Ordinary least square regressions to be an ineffective tool. I do agree with Man, that new systems need to be developed, for us traders to maintain an edge over mass marketers like 'pairstrading.com'. My suggestion is for all of you to think about other tools to build your models, a possible answer could be orthogonal regression. Has anyone worked on this before? I welcome other suggestions & comments.
     
  116. u have got to trade them. market rallies cover your short, market tanks you have a free bullet. put em on take em off all day. everyone in here is talking correlation and looking at 1 or 2 year charts. i day trade and i prefer to be flat, if i have that option. so when u look for pairs look to see if they go way out and come way in during the day not if they come back within a year. spread goes up a dollar back down a dollar back up a dollar this is ideal. some of these pairs i seen here used as examples are horrible like sbc and bls or adi txn i traded those before and they suc because they trend. a few months ago i was putting on txn adi started at $4.50 to go up (long adi). i like to add every .30-.40 cents. sure enough the spread went negative stayed there and wouldnt go i up i held it and held it traded the best i could, the txn specialist is criminal. anyways i was ready to take a loser then it finally bounced and its been trending ever since. these stocks are pigs. i put it on if adi is up over 2.50 and txn is up less than a dollar and first thing in the morning it comes in enough to take a quick profit. if u play this pair this is how u should play it. sbc and bls is worse. i would rather take a nap in my chair right at the open then trade these slow stocks. good luck all
     
  117. We are working on a similiar concept of a portfolio of pairs, pulling the shorts on an up run, pulling the longs on a down-run, putting them back on as the run stalls.

    Are you making money?

    How many pairs are you doing at a time?

    Are you doing it manually?
     
  118. has anyone been watching bollinger bands on pairs and finding it a useful/ useless indicator??? thanks.
     
  119. Yes. I've been doing some trend following pairs using 60 minute bars and bollinger bands. It's usually best to go with the trend when the spread starts to ride on of the bands, basically going parabolic in it's movement. If you are trading convergence, this would be a bad sign and you probably have a lot of pain coming while you wait for things to revert to a mean, which is probably starting to trend now too.
     
  120. metooxx

    yes, i am making money
    i watch about 30, but mainly focus on 3 or 4 that can be traded
    yes, all manually
     
  121. veeru,
    we have been playing with least square analysis as well and I feel about it as you do. actually we are a team of three people and I will pass on the orthogonal analysis to our statistic guy.
    right now it seems that we have profitable tests with sharpe ratios at about 1 or slightly above [over a five year horizon]. we are using automatic pattern recognition for that. it has the advantage that you do not necessarily "see" a trade on the chart. I think this is important because the only reason I can think of why standard tools suck is that it is too obvious. In a to some extent efficient market there is necessarly pairs trading going on.

    When I look at pairstrading.com and their recommendations I wonder about these pairs trading in completely different sectors. I always try to understand my trading and I have not got the rationale behind these trades - so far. Let's assume they are working - why are they working? Statistically there is at any point of time a number of accidental outliers. This means that there is necessarily a "stupid" pair, that cannot be traded at all but shows significant correlation in recent history. This is actually the reason why I tried so far to keep my hands away from these trades. But if pairstrading.com find it usefull to trade them, I assume they will have reason to do so - meaning it will work. But why. Does it mean that for whatever reason a big trading book has the two stocks in one basket and buys and sells this basket therefore creating the correlation + possibility to trade? Or is it a game that rather refers to a kind of "overall integrity" of the market than to a pure pairs concept?

    One other thing: if you look at two stocks and build correlations on an intraday basis it seems as if there was a constant decline of correlation throughout the day. I wonder why that is so and if it can be of any help for pairstrading. More in detail: take two stocks and use the 9:45 prices of each day and calculate the correlation. Then take the 10:00 prices and do the same and so forth. I found a significant and constant decline in the computer sector in 2002. This in itself would not mean to much but it contrasts with the fact that each stock has its most volatile period in the morning session. There is biggest volume, even bigger than in the closing range, and there is the period where the market makes more range per time than at any oterh point of time. But it seems as if the market was not only volatile but more aware of market integrity. I find this rather interesting.
    One argument might be that the market is highly correlated because the biggest nonTradingTime where stocks could not react is just over and now stocks are effected by news in the very same way. Nevertheless, it could mean fo intraday pairs that convergence trades are better placed late during the session (???).

    Does any one trade on a purely automatic basis intraday?

    peace
     
  122. IMO trendsetter is perfectly right in his approach. Unfortunately I am too lazy and love to delegate work to the computer [plus you can hardly handle thirty or fourty positions at one time].
    There are pairs trending and others mean reverting. Problem is to find out when [its always the same thing]. I think baskets of stocks will rather trend - I looked at some and had that impression.

    Another thing. I am always trying to find out how the market is developing. It is obvious that pairstrading has become a standard tool of trading. And this changes the market place. But how can I measure this? Maybe by calculating the average all possible correlations?
     
  123. Hi,

    You are absolutely right about pairstrading-related changes in the market place. That is the reason you are seeing companies like FO and ITT or FO and LMT having correlation of greater than 90%. IMHO, correlation is a key to some underwater processes in the market not visible otherwise. For all you know, FO and ITT could be traded by a big hedge fund on a daily basis from support to resistance and back... Knowing the correlation and identifying the oscillatory pattern of price action just bought you a nice ride which is predictable and profitable...
    BTY, I closed FO/ITT Friday hoping for overbought ratio today or tomorrow. FO tanked without reaching my entry point. Will wait. Still holding BK/MEL. Moderate profit so far.

    Al
     
  124. Man:

    I can't make out much of the methodology used by pairstrade.com. I do agree that sector pairs work the best. Although a variety of factors need to be used as filters even within sector pairs.

    While unrelated stocks can be highly correlated, do we trade them just on this factor? My answer is no! (We can show high correlations between butterflies flapping in Malaysia & a storm brewing in the Atlantic:D -- statistics is a tricky subject, anyone read 'Fooled by Randomness' by Taleb, it is an excellent book)

    One final comment, I've found position trades to turn in better results than just intra-day trades. As an experiment, I traded the same pairs on both styles, and came to this conclusion.
     
  125. Veeru,
    have you heard about a hedge fund "Whitebox"? they are doing pairs trading. they choose their pairs fundamentally and then optimise the strategy on the specific pair. this contrasts the view of those, who try to find one strategy for all pairs.
    isn't your butterfly example a classic finding of chaos theory?

    peace
     
  126. Neutral AI
    especially this FO-ITT pair puzzles me. I have never seen a better chart for pairs trading. It is simply amazingly regular.
    Maybe your argument with that hedge fund is correct, but why are they doing this? I doubt that one hedge fund could "produce" such a chart without loosing significant amounts of money. There must be something more to it, though i have no idea what that could be. Maybe there is a connection in terms of a potential merger or something like that.

    A great thing for pairs trading would be to forecast merger activity. Imagine such a trade! Right now it might be interesting to look at all potential candidates for a takeover by microsoft.

    nevertheless I am not sure about non-connected pairs, though we are currently papertrading such a strategy, using pattern-recognition.

    another thing. do you have an idea in which circumstances pairstrading works best? bull-market? medium volatility???

    peace
     
  127. I've searched the site a few times...can't find what I'm looking for.

    a site that offers FREE daily spread charts. Not the typical 2 stocks on one chart that plot their percent changes.

    I'm looking for stockA - stockb= value. With the value plotted.

    Thanks in advance.

    Friday is here! And we have possible overnight snow forecasted up here in the North East!:D
     
  128. what is the problem to do it by yourself in excel for example?
     
  129. Bob-
    wow you're fast...i grabbed a coffee re-fill and bam! Thats why I love this site.

    Ok, how do I get the data? If I have a DDE software link..maybe with that? I'm not that great with excel....but I'm a quick learner!
     
  130. pick data from here by using a webquery
    http://table.finance.yahoo.com/k?s=qqq&g=d
    then-do the math and then-click on chart wisard and build a chart
    it may take a while to build a chart, but after you get familiar with it-you wil be fast enough)))))))))
    also-you can write a macro and it will do it for you automatically
    Good luck!
     
  131. Bob-
    Thanks for your help.
    I'll give it a shot.
     
  132. Also, go to www.quote.com and use their live charts for free. You can put in A - B or A / B or even (A * X.Y) - B. Just make sure to leave spaces in between any operator.

    Royce
     
  133. thanks guys
    I actually built a chart with excel and it was surprisingly easy.
    Thanks again.
     
  134. Hi man,

    I find that despite being market neutral, pairstrading is still dependant on the direction of the market when you trade pairs based only on correlation. For example, the pair TXN/MWD is highly correlaed over the past 2 years. However, TXN will always underperform MWD in the downtrend and outperform it in the uptrend.
    If you trade pairs in the same sector like LLTC/MXIM, XLNX/ALTR, etc., market direction matters less.
    Some pairs like FO/ITT are affected neither by directional changes nor by volatility.
    I closed BK/MEL today with a nice profit. It is getting too close to 40DMA, and it looks like we are turning south, so I expect BK to underperform.
    I also finished my model. It provides me with correlated pairs, extreme RSI, 10DMA cross towards 50DMA. It also calculates desired SD from the mean. Different samples test from 75% to 87.8% profit.
    Now I am ready for REAL TIME.
    Does anyone know a data provider that can stream S&P500 reliably and affordably?

    Al
     
  135. Al I know you were using the S&P 500 stocks only, but I was wondering if your test was on intra- or inter-day data.

    Thanks Royce (and cheers on finishing your model)
     
  136. Neutral AI, Royce,
    I find this rather exciting. we finished our model as well and will start trading next week. we use a pattern recognition tool for the analysis of the spread.

    data
    We use CSI&Bloomberg for data history. Both seem to be pretty reliable, delivering dividend adjusted data, although CSI struggled during last days to adjust T for the 1:5 split. But usually they are doing quite well. If you need data split- and dividendadjusted for longer periods than about ten years, you get a problem since CSI cuts numbers to two digits after the comma. that means that if you look at adjusted data back into the sixties, that some stocks could only move by 5% - everything else falls out of the tow digit-borderline. Nevertheless, they offer excellent price/value ratio for their service. the database is stable. we have been using them for over year now.
    for realtime data we found eSignal quite sufficient.
    Tickhistory is nice from quote.com.

    I would like to stay in contact with you guys, if you do not mind. Have you heard about a hedge fund "whitebox". they are doing pairstrading using technicals plus fundamentals.

    peace
     
  137. Hi man,

    Could you please explain what is the pattern recognition tool that you are using.

    Al
     
  138. man/ veeru
    I have read Talebs book and recomend it.
    Most sites/ people fall into either trader or quant gategories but elite trader strikes a good balance. contingencyanalysis.com is very quant biased.That is you need to spend seven years doing a phd to prove two indices are negatively correlated before we can make any scientific statements - by which time we have forgotten about making money -as we have been distracted into other interesting branches of mathematical analysis. Cynical , but there is some truth.
    We must remember we are playing the percentage game like the gambler to know enough about maths but noy too much!
    After all- you dont need to know about the physical laws determining movement of objects through space to be a great snooker player- infact it is more likely the less an individual knows about physics the better the snooker player !!!
    Perugino
     
  139. Everyone please advise me when you enter a pair,the stocks, the size of your trade, and your equity balance. I will then proceed to deliver as much pain to you as is necessary to force you out of your position.

    Pair trading combines three of my favorite pasttimes:

    1) Trading

    2) Gambling

    3) Smoking crack rock

    I've seen the gobbledygook study purporting to support this strategy, and for my money---it only works well around the opening.

    I know plenty of people who make big dough using this strategy, but ALL of them are at risk of blowing up( the non-terrorist kind) at many points in time.

    God bless 'em. They need it.
     
  140. trader,

    as usual, you are right on the money. I have stopped pair trading altogether until I see another way...

    It was a gold mine for me for a while, and then it just went COLD.

    nitro
     
  141. Probably for reasons, that I mentioned above, but nobody pay attention to it till you get it…:D :D :D
     
  142. thetraderprofit,
    why don't you choose your words more carefully? It is possible to have an opinion without being aggressive.

    Since I am a person starting to trade a non linear pairstrading strategy tomorrow, I feel a little offended.

    I understand quite well that there is a special (cool?) way to talk about things, but this is a free forum and the right of free speech works best, if people try to choose their words carefully. Abusing it means diminshing it.

    Sorry, if I am now offending.


    peace
     
  143. Perugino,
    I agree with you. What you are actually claiming is, to use the very best computing facility on earth: the human brain. Now this brain is the perfect tool for learning. Imagine where humans live, from coldest ice to hottest desert, and it is the adaptive capability of the brain which makes this possible. Two children of about four years were found in india who were raised by wolves. It was impossible to make them behave like human beeings - they learned for years to behave like wolves; moreover they became wolves.
    If you train your inner system to the markets in the right way, you will inevitably beat the best machines in the best funds out there - no doubt about that (another thing is how to organise the learning process, since frustration is a hard hindrance for the learning ability).

    There is one thing where the human brain is really bad and beaten each and every time by any standard PC: processing mass of uniform data. This is not a good field for our structure to work in.

    Now being aware of that, there is several things one can do about it. First is to ignore computers at all: will work with proper training on few markets - no doubt. Second use small computable edges on a mass of instruments and profit from combining what a human brain could never combine. It works because the inefficiencies are so small and vulnerable on the single market, that they are not tradeable for the first category of traders.

    Now, is there anything special about this? no. Use computers to the amount where they really support your brain. Your example with the theorists is right. The same accounts for many traders who neither use their brain extensively enough nore rely on computer power.


    peace
     
  144. Man,
    The proof of the pudding is in the eating- If Ihad made a million dollars from tecnical analysis I would be praising it and having a discussion on its forum- I havent so I am here instead - what attracts me ? The notion that if an indices(r) is oscilating around a mean I stand a good chance of predicting its path.
    My model isnt complex!
    Its not really scientific.
    - but that isnt to say computers cant assist- lets face it if it wasnt for the computer we wouldnt be here !!
    You know- thereis alot against us succeeding at this - I want to digress on your point on the human mind- when we win our bodies change we become lightheaded and optimistic - what do we do - convinced of our success we invest more in our "proven" technique only to see failure. Its about balance - account for the fundamentals - the blow out (5-10% ??)
    We must all feel we can succeed or we wouldnt be here !!!
    LETS GET UPBEAT -


    As Abe said
    "Good things come to those who wait - but only what is left by those that have taken the initiative"
     
  145. right.

    peace
     
  146. Man,

    I wasn't trying to be aggressive, only humorous.

    But....

    You seem to be saying, "I'm going to start doing this tomorrow
    anyway, so please don't bum me out ."

    Well, sorry, but here are the facts:

    I know quite a few people who make a living at pair (or pairs)
    trading. It must work to some degree, and I use it around the opening, or at the opening, particularly when there are upgrades/downgrades/news.

    When I say, "a living", I mean a very good living --low to mid six figures. One guy uses it on bond futures and is up $1.5-$2 million this year between bonds and stocks. Of course, the position sizes necessary to generate these profits are such that great damage can be done to your net worth......just as with any arb strategy.
    For example, the guy who is up $1.5-$2 mill, had, at one point, $1.2 BILLION in long/ short bond futures positions. Yes, that's a "B", not an "M." I'm not sure what the haircut is on that position, but his firm was obviously at risk--not just him.

    To a man, these pair traders do NO fundamental analysis, and almost no technical analysis. Their selections are based almost entirely on past price relationships.

    Now, having encouraged you with the above, I know of many,many more ex-traders who lost all of their capital with this strategy. I say I know "of" them, because they remind me of the poorly-trained new recruits sent to the front during WWII, never to be seen again. The experienced soldiers didn't want to know their names. Unfortunately, some of the new recruits were friends of mine.

    Here's an example of the risk of any arb strategy gone awry. In August, I lost $10,000 in an arb involving 2,000 shares of a $2 stock. The short side rallied and the long side dropped 80%. So, the point is that all arb strategies are risky. Usually, a great deal more risky than you first assumed.

    I do pair trade, primarily with low volatility stocks like REIT's.
    I do some fundamental research, but not much. But even pairs of VERY similar stocks like FNM/FRE have blown out many a trader.

    Man, I don't pretend to have some of the answers, just all of them!

    (My private pair(s) trading course is available for the small fee of $100,000)
     

  147. Hi Nitro,

    What was the reason you stopped pairs trading? Was it the pairs that blew out because of earnings? Or was it because you just got bored with small profits and those positions that went nowhere for weeks?
    I feel the same way sometimes. But look at the reality. The most successful hedge funds are the market-neutral ones with profits on the order of 20%+ per year. Look at Pairstrading.com... Even with most positions going against them at first, they still manage to show profit of 1 - 3% per mo (9mo out of 11 that I saw). Look at FO/ ITT... I traded it successfully 6 times in the past 2 months with total profit of 30%+. I went long ITT/short FO today again.
    Making money trading pairs lacks glamour and thrill. It is a slow process of adding 1, 2, 3 percentage points to your capital at a time. But that is how MONEY is made.

    Good luck,
    Al
     
  148. thetraderprofit,
    I see your point. In the end it is all about results. Thus we do not have to argue the ifs and whens. We do our first thirty pairs today and will see how things turn out.

    I find your approach quite interesting, since we could prove that although the morning session is the most volatile, it is the time were pairs are most integrated. I do not want to get into detail, since I assume this would bore you, but I think trading at the beginning of the day offers opportunity.

    You mentioned FNM-FRE. It is very funny that with standard technology this pair can hardly be traded, although it has a 0.65 correlation even on five minute data. this is astonishing since this correlation must come from somewhere. meaning someone is trading these stocks parallel otherwise this correlation could not sustain.

    My understanding of trading is to have a wide portfolio with a small edge in each trade. This is not at all satisfying for a daytrader who looks at a dozen stocks, where he has a rather big edge.

    Sorry that I did not get your humour immediately ...

    peace
     
  149. I disagree that it lacks glamor and thrill. I guess it just depends which stocks you're pair trading. If you're feeling really sexy, try UNH/WLP, or some of the brokers; thus, my references to gambling, and smoking crack.
     
  150. There are dozens of guys/gals trading this pair..... many more who USED to trade it and either lost all their money or someone else's.

    I know one guy who blew out his account at Bright on GE/HON (and quite a bit of the firm's money), then went to Generic and, I've heard, he blew out on FNM/FRE there. I also heard he was the Risk Manager at the local Generic office until they decided he needed to go.

    I think the best strategy is to find a firm willing to let you gamble a lot of their money. If you blow out at the first firm, keep hopping around. Eventually, you are bound to make a mil by sheer luck! And....I'm not kidding about this.
     
  151. I think that is like saying "trading stocks" got cold. There are
    an infinite amount of ways/styles/strategies withing pairs
    trading to say that it went "cold" is impossible.

    However, my style went cold this fall. 23 months up in a row,
    two months down. I attribute it to a temporary interday
    correlation breakdown across more names than usual
    (in particular the names I trade (liquid listed universe).

    I absolutely believe with the comment on the other thread,
    in general pairs trading can be a lower risk (albeit) lower
    reward way to trade. But a slow steady rising p&L curve
    is what most traders should try to achieve.
     
  152. However, my style went cold this fall. 23 months up in a row,
    two months down.

    Markets change. Your universe of stocks change. decimalization has really hurt my trading. Program trading has hurt my trading. A lot of different factors have affected my performance trading pairs (including my own psyche). What I have had to do to stay in this business is to adapt to different styles of trading. I have forced myself to lower my expectations and to look to hit singles and not give up the home runs instead of swinging for the fences like I have the last 3 years.
    good luck
     
  153. Due to a request from the dude who made $2 mil pair trading, I will no longer be posting in this forum. OK, GG??
     
  154. I just found this site a few days ago, and have enjoyed reading the post and glad to see that ideas are being exchanged. I have a model that I am working on that I would like to share, and hopefully get some input with...

    It starts with a matrix written by Bloomberg (those w/ access to bloomberg can email me for directions on how to get the matrix). The matrix allows you to input a list of stocks (max of 256--# of excel columns), or you can enter an index like SOX or NBI (Nasdaq biotech index), and the days back and it will retrieve the data for you...it does not do intraday. The spreadsheet produces 2 tabs, one with the matrix of each comparison/correlation...the second with a list of the two securies and thier correlation.

    This just gives you the stocks and thier correlation...nothing more. I wrote a coniditional format for the matrix to paint stocks >95% correlation BLUE and <-50% RED. The second tab listing the pairs and correlations can be 'filtered' in excel to show the same data. Once the data is listed, you can then fill in a number of columns on current market data (% Change, Open, High, Low, etc) various fundamentals, relative strength, technical info (put/call ratios) etc.

    Once you fill in the blanks, you can write conditional formats to show which one is 'better' than the comparison. IE-all you are looking to do is identify the stronger stock and the weaker stock, NOT forcasting direction!! For example....
    With Bloomberg, I can pull in 1,2,5-day % change. The both may be up or down randomly over the period, but when I apply the conditional formats, and see that the 1,2,5 day % change is green (three greens in a row) I can see that it is outperforming its counterpart with 3 reds in a row. This is not rocket science, and it is not the ideal entry...but it has lead to fairly decent returns (-5% - my stop loss tollerance to 15% in a few days) ...and I never even looked at a chart or any fancy indicators. Granted, timing can likely be improved by looking at multiple regression analysis, normal distribution curves, Z-Scores, etc.

    There are other excel 'add-ins' out there that will calculate the matrix for you. Check out www.factset.com or www.portfolioscience.com Esignal and others may have the same info now, I'm not sure. TradeStation or MetaStock may do them as well...if anyone knows, please let me know! The matrix is a quick and easy way to calculate correlations. Many people traded BRCD/QLGC for a long time, because they were very highly correlated, but due to QLGC's recent run-up, it may not behave the same way as BRCD, and their correlations have likely changed.

    I have read about a lot here of people loosing a ton of money trading pairs, but the odds dramatically favor pair trading, as opposed to directional trading. Consider the following:

    2 stocks, 2 outcomes (up or down/win or loss)

    Pair Outcomes:
    Stock A - Win
    Stock A - Loss
    Stock B- Win
    Stock B- Loss
    (win/loss NOT EQUAL to Long/short or Up/down)


    Trade Outcomes:
    Stock A- Win & B- Win (winning trade)
    Stock A - Win & B - Loss (wash)
    Stock A - Loss & B - Win (wash)
    Stock A - Loss & B - Loss (loss)

    The final results mean there are a 1/4 chance of winning, 1/4 chance of loosing, and 2/4 chances of the trade washing out...compared to the 50/50% chance of winning on a directional trade.

    Just because you are 'hedged' through a pair, doesnt mean you can disregard common rules of trading....CUT LOSSES, use proven position sizing/pyrimiding money management techniques, etc.

    The point is...I've demonstrated earlier how the risk/reward favors pairs over directional trading...When you're right, the profits can be componded because both long and short can work in your favor at the same time, but they can also work against you.

    Good luck....
    PearTrader
     
  155. like one dude said before-it may work for some time.
    it been work for some time for me (not the way you choose)
    but, one day, after i left for 30 min to pick my son from school pair got from 3% to 20% in different directions. regular stocks, no news, just like finger snap..BUM! and since then-my winning system start been unprofitable at all. more than 300 winning trades over about 10-15 losses became totally useless.))))))
    why do i say all this-because i do believe, there is not much difference between directional trading and pair trading.
    sorry for may bad english. just my word of caution)))))))
     
  156. I spoke above about cutting losses....

    For those who use excel for their calculations, you can program a simple 'stop loss' that basically says

    IF %ReturnSinceEntry is LessThanOrEqualTo - 5 %, THEN ......
    ....then make that cell RED (conditional format)
    ....then play a wave file that shouts "SELL, SELL, SELL"
    etc (requires VisualBasic code)

    Bloomberg's Tradebook ECN and other programs (possibly RealTick and TradeStation) can launch orders when certain excel conditions are met...

    For example....
    If ABC/XYZ < -5%, then Sell ABC @ Mkt AND Buy XYZ @ Mkt

    There are probably other ways. The good thing about pairs, is that RARELY does a pair move so fast that you cant get out with more than 0.25% slippage, unless one gaps up or down....which brings up another point....

    **EARNINGS DATES**...
    Some programs allow a DDE link to excel to import the expected earnings date....
    If you have todays date in a cell =today() and the earnigns date in another cell, then write a conditional format that says

    IF (EarningsDate cell) is LessThan 5 WeekDays from today, THEN....
    ...then make that cell YELLOW....
    If (EarningsDate cell) is LessThan 2 WeekDays from today, THEN....
    ....make that cell RED....

    This will ensure you dont hold a stock through its earnings date, unless that is part of your strategy.
     
  157. Good idea. Does anyone know a way to import an earnings date into excel automatically through DDE? Perhaps you can download the earnings date through Yahoo finance the way some download EOD quotes?

    Dividend dates might be useful as well.

    :cool:
     
  158. PearTrader,

    All this stuff and your previous post is "arithmetic" to a professional pair trader, i.e., it is elementary.

    I am not trying to dissuade you from doing pairs - I did it successfully for a while. But I can tell you in no uncertain terms that something about the market changed that made this kind of trading very difficult [disclaimer - as I usually say, this is more a statement about _me_ than anything else.] If you like, the risk holding period changed in such a way as to make it leave the bounds of my risk tolerance...If I reduced the holding period, the chances of stopping out were much greater [than they used to be]

    I have not given up on it. I have some new key insights that I am still working on, but I am not holding my breath...

    nitro
     

  159. nitro-dig into volume... one of the keys-there...

    mrDinky-how many pairs you planning to trade, if you need automate alert on them? a thousands? i trade few stocks, they make about 15-20 pairs. you must know everything about them. from products to behavior in different market conditions.
    i don't know what you mean, on export earnings thru DDE))))but, you can create web query in excel and it will raise alert. need a little VBA knowledge
     
  160. I know I said I was finished on this forum, but I just had to get this in:

    Bob 111, your problem is NOT pair trading. It's children. Stop picking up your kid during trading. Has he ever heard of a school bus?
    Taxi? Bike? Rickshaw?


    Here's the advice I gave all new traders when I ran a Bright office:
    Ditch the kids. They represent the biggest drawdown you'll ever have in your life. Join Big Brothers. At least they become someone else's problem when you are trading.:D :D :D
     
  161. :D :D :D :D :D
     
  162. Exactly my findings, too. There is very little left to trade on that I've been able to identify.
     
  163.  
  164. I got an e-mail from www.pairstrading.com over the weekend stating that they lounched a new pairs trading software product on their site. I subscribed today for the first time. The software gave me 33 pairs for the Russel 1000. Some of them look pretty interesting. Some are really strange.
    Has anyone else seen this new product?
    Let me know your thoughts...

    Al
     
  165. What is the holding period for these pairs?? hours, days, weeks, months?? I have a feeling that the russell spreads have a long holding period which exposes you to a LOT of risk imo. I have traded many longer term pairs but I prefer the shorter term quick profits (quick & nimble in this market). I have also signed on for a free trial to the website and have been very unimpressed with the pairs that they have recommmended
     

  166. Hi Homer,

    Apparently what they did was they took certain correlation and divergence parameters and calculated all the pairs for the Russell for the past 10 years. According to them, any pair has a probability of 80% to revert back to mean within 2 months. I guess if you trade all the pairs that they give you and hold them for 2 months, then you should win 80% of the time. The problem is twofold: a) you cannot trade all the pairs, and b) they don't give you exit points (profit or stop loss). So you have to do a lot of research yourself. I would not blindly trade these pairs...
    On the other hand, this software could be helpful as a research tool. They do give you correlated pairs and they do calculate the most highly correlated stocks with the stock of your choice. That is pretty cool. All you have to do now is some FA and TA to determine the most tradeable pair.

    Al
     
  167.  
  168.  
  169. Be carefull. Any correlation between such difference stocks could be (and probably is) a fluke that will not continue in the future. secondly, where did you get that correlation measure? i show that there r-squared is .3 using prices and .05 (noise) using percent changes on the day. bloomberg screen attached.
     
  170. Al,
    lmt/nly correlated at 86%??!! Does that site give you return correlation or price correlation??? There is a big difference. The CECO/APOL is a better example though. I have run across this pair before but have never traded it. What time frame are you using to determine your entry/exit/stop prices?? I'm assuming your technical levels were for the ceco/apol spread, but i'm not seeing what you are seeing. I like a - 1.6 entry point (short) or buy the spread around - 2.6. Of course I am not looking to hold onto the position for two weeks though. Thanks for the info on the web site, I think I'll take another look at it.
    Good Luck
    Homer
     
  171. Does anyone know if TradeStation is capable of naming a pair of stocks as a single stock? In other words, if I have a universe of highly correlated (or negatively correlated) pairs, I want to scan the PAIRS for certain criteria....the same way anyone else would scan single securities for certain criteria.

    If anyone can offer any suggestions, I would greatly appreciate it.

    Thanks
    PT
     
  172. First, Tradestation 6 doesn't have scanning capabilites. The pair would have to be in an open workspace, with an alert set. Practical for 50 pairs or less. TS 2000i has a workspace scanner and can scan across a large list of symbols. TS promises this scanning capability will return in V7 (don't bet the farm on it).

    Secondly, if you setup both stocks on the same chart you can then program most any sort of indicator off them, using data1 and data2 as input. You can then apply most types of TA to the pair indicator, however, you cannot "buy and sell" the indicator. You can't buy and sell the pair either, just the first dataset. This is a big "gotcha" for pair trading.

    I've gotten around this limitation by programming an indicator to simulate the buying and selling of the pair and reporting the results to an output window or file.

    Also, you can't import data into TS (yet).

    Hope this helps.
     
  173. I don't quite understand the point you're trying to make. Please elaborate. I leg into and out of pairs many times per week based upon technical indicators.
     
  174. I agree with homer that correlation on price means something very different than correlation on returns (=logarithmic yields).

    LMT - NLY highly correlate on price but do not on return.
    Looking at a simple chart of prices, they indeed look as if they walked together after early/mid 2000. Nevertheless, current 250 day correlation is 0.21 on yields and 0.71 on price.

    Normal analysis would not take that pair into account, since usually correlations are not calculated on price.
     
  175. I apologize for not reading all postings of this thread, so I might have informed you already. For those long threads it would be good to have a “search the thread” function. Anyway, here are a couple of scripts (systems) that do what you might be looking for, I am not posting the code, since you can take a look at it when clicking the link but I add the description of the publishers so that you know which might be of interest for you. You can also test some of them after clicking the link::

    1. From the article "Pair Trade II System" from the Trading System Lab section of the February 2002 issue of Active Trader. This script always uses Microsoft as the second symbol of the Pair Trade. You can modify the script to use a different symbol. http://www.wealth-lab.com/cgi-bin/WealthLab.DLL/editsystem?id=14030

    2. This is from S&C March 2001 issue.
    Pairs Trading is a very interesting strategy because it is "market neutral". That means it should perform equally well in bull or bear markets. Two closely correlated stocks are chosen. Whenever there is a strong deviation from their "should-be" ratio, one is bought and the other sold short, hoping that they will soon come back to their "normal" ratio. Additional information and a complete explanation of the strategy can be found in the S&C article by Stéphane Reverre.
    To use this script, you must create a watchlist which contains the symbols SC and RD.
    The results are certainly not spectacular, but this strategy produces a steady stream of income and shows exceptionally low drawdowns. Example $imulator result: RD/SC pair, 2000 bars, 50 % of capital for each trade (i.e. 50 % SC AND 50 % RD), one way commissions 2 $.
    The 103 % profit in 8 years is about 9.5 % a year, but the exposure is only 20 %, so we could try to find some additional pairs to approach 100 % exposure and obtain a reasonable income from this strategy.
    To get this strategy to produce money, it is essential to chose a broker with low commissions. We should pay no more than 1 or 2 $ commissions for a 5000 $ position. Example: 5000 $ is about 123 shares of RD (as of 02/08/2002, closing price = 40.54 $), commissions 1 Cent/share = 1.23 $. http://www.wealth-lab.com/cgi-bin/WealthLab.DLL/editsystem?id=6655

    3. This is an enhanced version of the original "Pairs Trading" script (the one above). It avoids unnecessary trades by providing exit thresholds. It now uses market orders to make it convenient to trade.
    Thanks to jrbyrd's contribution, you can now click on any symbol in a watchlist, and it choses the best correlated symbol automatically. The correlation coefficient is displayed in a small text pane below the price pane, so that you can see if the chosen pair is appropriate for pairs trading.
    This script is vulnerable to bad data quality. There should be no missing/double bars in your data, otherwise you will see a shift between the two trades of a pair trade (different entry/exit dates for the two symbols). If you're not sure about the quality of your data, you can disable "Strict Synchronization Check" in Tools/Options/System to avoid error messages. "Reverse Synchronization" must be enabled.
    If you want to further explore the correlations between the stocks in your watchlists, see discussion thread "Correlation Analyser" (on the WL site).PS: The script doesn't work on the web site, for WLD owners only! http://www.wealth-lab.com/cgi-bin/WealthLab.DLL/editsystem?id=6990

    I hope you find some valuable information in the various systems. Programming pairs systems is surely aimed more for the advanced user then for beginners, but the description might give you some idea. All example above are not my work. The credit belongs to the nice people who were kind enough to share their programming work and ideas with the trading community. Most of all what you will find is the possibility to add empirical evidence on some of the assumptions or a base for furhter studies.

    Volker
     
  176. man,

    What's wrong with calculating correlation on price? How calculating correlation on yields gives you better prediction?

    Al
     
  177. Neutral AI
    in most cases it will not be too bad. but when there are extraordinary large moves the picture and the interpretation will change.
    I include a graph of FNM FRE, correlated at about 0.9. So roughly speaking in the same category of correlation claimed for LMT-NLY. I feel that even looking at the chart only one can see significant more similiarity. 0.8 or 0.9 are very high corrleations, thus the stocks move almost always parallel.

    Actually I do not care too much about discussions on methodology. if correlation on price works for you, fine, don't change.

    peace
     
  178. a. Volker, those scripts are great and people should definitely check them out. I use modifications of them for all my pair trades.

    b. don't get so obsessed with correlations between similar stocks. Those correlations are old news by definition. If you really want to make "nickels" trading pairs than look for equities that are LEGALLY linked to correlate and then observe the correlation between their historical volatilities.
     
  179. Vishnu
    what do you mean by legally linked? I find the idea of correlating volatilities a very good approach. Sounds very reasonable to me.

    peace
     
  180. check out the correlation of the volatilities between DUK and DUR when DUK had a downswing abt 1.5 months ago. the correlation between the 10 day historical volatility went from 0.97 to 0.80. Then look what happened to both stocks in the week following that. Factor in the beta of both securities, plus the yield on DUR, and it gets interesting.
     
  181. I see your point. I never looked at stocks of differnt type of the same company. I am a little scared that there might happen legal or accounting changes and I jump on a trade which is none.

    From the chart I would think that I would have seen the trades, you are referring to. It looks like you can trade the spikes of the spread.


    peace
     
  182. God, I just finished reading this extremely long thread. All this info to take in. Let me take a breather. WHEW.
    Thanks a lot to everyone that posted, lots of good info. Now I can start trading pairs. I am so excited.

    Gennady :)
     
  183. This morning Salomon Smith Barney lowers ratings on oilfield equipment and services. I have noticed that many times SBSH will recommend 'rotation' from one stock to another, which suggest the two may have similar fundamentals, products or services, (taking market share away from a competitor, etc) for good pair trades.

    This morning, they DOWNGRADE WFT (Weatherford Int) to In Line from Outperform and UPGRADE VRC (Varco Int) to Outperform from In Line.

    In recent weeks, similar upgrade/downgrade news between APA/BR yielded about 4% in a few days, and more than 15% between NSM/MU.

    This kind of news often serves as a fundamental-based catalyst, but still needs to check out to your technical/statistical rules before entry. Remember to keep tight stops in case they don't trade as expected.
     
  184. This is an interesting thread, but has anybody tried to adjust their parameters with the day of the week? I have a swing trade system and i've noticed that there is a historical diference for each day of the week.
    Just something one might look into a bit.
    jj
     
  185. just my 2 cents-as i mentioned before-it all about volume. not correlation. difference in volume create % difference. from my expirience-friday is good day for quick profits. i mean-really quick(1-1.5%) in 1-2min. because usual low volume by end of the week create greater volatility.
     
  186. Al,

    I was busy with another project for some time, sorry. I have tested your ideas with different stocks as well and I find whole concept sound and interesting.
    I`ve been through whole this thread since then, so few thoughts come to my mind:
    1. No such correlation is perfecty reliable, but as you said once the correlation is broken you may just accept losses. Risk/reward ratio is just simple statictics.
    2. The key as usuall, is diversification. Trading few different correlations should make equity line smoother.
    3. Considering all above I believe that trading pairs needs to be well sized to be consistently profitable. Intraday trading may lower requirements, but this is another story.

    Anyway, I find myself whole idea very interesting. Thanks for bringing new light into trading area.
     
  187. Hi everyone,

    I haven't looked at the thread for quit some time, was out for Christmas and New Year. No trading for the past two weeks, just hot sun and warm ocean...
    Now I am back. A lot of my pairs that exhibited extreme oversold/ overbought conditions in Nov. and Dec. are now converging. Makes me think that we are approaching an important top. I ran some stats: the highest number of extreme pairs occurs in the middle of a trend. They tend to converge the most at the tops and bottoms.
    The earning season is approaching... gotta be careful. On the other hand, could not help today to open a new position MU/ALTR R=0.6950. It looks extremely oversold(14 D RSI below 30 since mid Dec.), should correct.

    Any thoughts?

    Happy New Year To All

    Al
     

  188. yeah.....never mix stocks from NYSE and NASDAQ
     
  189. Currently on INTC/ALTR (higher correlation + high fundamental skew- earnings expectations and profit margins). Considering TXN/XLNX (analogs). On GLG/MDG as of close. Earnings can be the best time to enter these spreads- IF volaitlity breaks your way. My biggest wins have been during earnings.
     
  190. Hi Avaturk,

    INTC/ALTR looks good. INTC/XLNX was recommended by Pairstrading.com a couple of days ago. Kind of same thing. Earnings on Intel would be beneficial if they were bad. XLNX and Altera have a higher Beta so they would fall harder than INTC. If earnings are good you will see the same yield on both stocks IMHO. GLG/MDG looks strong. Although RSI14 is 86.89, I would wait for 10DMA cross from above. What is your stop loss/profit objective on that?

    Al
     
  191. wow after all these posts I am still confused. You buy 2 stocks in the same industry:confused:
    All that school and all those seminars for nothing.
    On a serious note all I can say is that you have to possess pain tolerance to pain trade-opps I mean pear trade.

    inmate451

    which I don't have:mad:
     
  192. I haven`t been here for a while. So, how long you`ve been doing this and what is your experience now ?
    We recently also test kinda mixed pairs/spreads strategies. What do you think ?
     
  193. haven't been here recently too. our strategy started early december live trading and is up 3.65% by yesterday evening. is much better than tests indicated but it seems that this year end was good for all mean reversion strategies.

    we are moving towards fundamental analysis.
    I am thinking of something like pairing a small cap and a big cap on fundamental and technical data. the idea is that small caps should offer more upside. thus I actually try to capture the different risk profile between the two. I have not hear about that in pairs trading so far - anybody ever tried such thing.

    at the same time we are starting with intraday testing. we prepared our test software and should be able to start tests today or tomorrow. in pairs trading context i am most curious about very similar pairs like fnm-fre. i think there must be a way to profit from this obvious pair if you are not too big and well equipped.


    peace
     
  194. Hi ChrisM,

    I have been trading for over 12 years now. I am doing pairs trading exclusively since September 2002. I use my own technical system based on price correlation analysis. I looked at the yield model as well. The price model is easier to compute and it gives you better stats ( at least in my case).
    Believe it or not, mixed pairs produced better stats and actual results. Last 4mo record: 86 trades/ 72 gains/ 14losses. cumulative gain 8.86%. Average per month +3.48%.
    Same sector pairs: 32 trades/ 18 gains/ 16 losses. cumulative gain 2.01%. Average per mo +0.64%. The problem is that the pairs approach the stop loss before they revert. So the assumption that the same sector pairs perform better is wrong, I guess.
    What are your strategies? Are they purely technical or fundamentals-based?

    Al
     
  195.  
  196.  
  197. Neutral Al,

    I also switched to pairs trading about a year ago. Overall positive results. However, I've generally not had stop loss points. I've either added to my position or waited it out.

    For your trades in which you have gotten stopped out, have you ever gone back to see how the pair performed longer term? I've had a few loosers that ultimately would have been winners had I hung on and plenty of times I've recovered from a loss and sold at breakeven only to see shortly afterwards that I should have hung on.
     
  198. Careful with those. I know some people who are STILL hanging on to the FNM/FRE pair someone mentioned. "It's gotta come back..." More traders turned investors.

    :cool:
     
  199. Actually I've had much success with FNM/FRE. When I started pairs trading them the spread as I recall was about 14 points. Now it's in the 5 to 7 range. Somehow I managed to turn profits. Looking back at my trading for 2002, I pairs traded FNM/FRE 84 times without a single loss! OK, sometimes I cheated and sold my long position while continuing to hold my short position or vice versa so I wasn't true to pairs trading.
     
  200. Sure it is..."crutch pairs" trading is less comission intensive and a perfectly good way to trade a "pair."

    nitro
     
  201. Hi Al,

    glad to see your post - found it through a google search. I've become fascinated by the concept of hedge trading correlated pairs. I've actually made only one trade of this kind - put $2,000 on the line and profited $150 in 10 days. I used options on indices - not much risk, but not much gain - worked, but maybe I was just lucky?

    Anyway, maybe I'm over simplifying, but you can download stocks' price histories at: http://moneycentral.msn.com/investor/charts/chartdl.asp,
    import them into excel and use excel to get the price ratios. My intuition is that you use excel functions to find average and standard deviation of ratios...

    For put/call ratio history I use:
    http://www.schaeffersresearch.com/streetools/equity_OIPCRatio.asp
    It gives only 3 months of history, but that's enough for me to see direction.

    Hope this helps, and would like to hear how you're progressing.

    Jerry
    jerryjuracich@hotmail.com
     
  202. Hi,

    I'm new to this group and didn't realize there was more than the first page of this thread when I posted my earlier response. Sorry for not reading more before posting - It's clear that you guys are way ahead of me.

    Wow - after reading the next 20 pages of the thread, my mind is really reeling - It's great to see that there are so many traders into hedge trading correlated pairs and sharing their experiences - I'll try to absorb some more later.

    Jerry
     
  203. Jerry,

    there is no "maybe" in trading. If you had good, backtested trading plan the profits come from statistics. If you had none, it could be only luck.
     
  204. is anyone doing pairs only within the same sector, maybe using fundamental data as well?
    anybody doing pairstrading cross-country?
    japanese stocks?


    'always looking for diversification ...


    peace
     
  205. All of my pairs trading is from correlated pairs within the same sector. I like gold, telecom, utilities, and oil. I've had limited experience and bad luck with financials and retail. Also, I'm always trading with expected convergence rather than divergence. Often I'm setting up the trades based on the ratios of the stock (e.g. buy long 200 @ 10 and sell short 100 @ 20) but more often I buy an equal number of shares both long and short and play the spread. This works best when the share prices aren't dramatically different and makes ongoing analysis of the trade a lot easier.
     
  206. jrweiner
    what's your average holding period?


    peace
     
  207. "profits come from statictics" - and luck.
    "if you had none, it can be only luck" - not true, you can trade without backtested system and your profits won't be connected with luck only.
     
  208. Man,

    you haven`t changed at all :D
     
  209. ... arguing

    some have the backtest implemented within the fastest computer on earth ...
     
  210. Man,

    this is not really arguing. We are just old friends.
     

  211. Ooops,

    I can see the error. I answered to DT saying "Man" (i.e. Yo` Man !) and here is misunderstanding. Sorry Man,
     
  212. I've missed the key word "could" in your statement, sorry man :D
     
  213. Ups I think it's another misunderstanding. You should say "there is no maybe in pairs trading".
    This is arbitrage, where profits are more related to statistics. Directional trading is less related. Correct me if I'm wrong - I've never made any pairs-trading.
     
  214. Chris
    I saw you did not mean to talk to me. I just wanted to prevent what I was afraid could turn out as the old backtest-or-not discussion. Sorry for being impolite, I was not aware you knew each other.


    peace
     
  215. Actually my belief is that whole trading is statistics related (more or less). But therm "statistics" covers large area.
     
  216. Just misunderstanding. And you were OK. No appologies necessary.
     
  217. It may look good on paper. In real trading two things are crucial in such strategies, IMO:

    1) Correlation changes over time
    2) Commission & slippage costs x2 higher vs directional trading (can be a problem mainly in intraday pairs trading)

    Does pairs trading have any advantages (in terms of return/risk ratio) over directional trading? I doubt it.

    It would be a good idea to include pairs-trading systems for diversification purposes.

    The whole idea is to profit from market's over-reaction (or inefficiency). Traditional trading systems try to profit from exact same thing, by using profit targets for example.

    Nothing new. I'm boring... :D
     
  218. ... market neutrality. that's something in itself.


    peace
     
  219. My average holding period for a pairs trade last year was 3.1 days. Lots of day trades and some I got stuck in for a long period of time. I would probably do better if I was more patient but it wouldn't be as exciting.
     

  220. Hi jrweiner,

    I do go back to look at the pairs after I got stepped out. Some of them improve, some loose even more. Trying to analyze what would have happened if I stuck with a loosing position for a little longer is a pointless and psychologically damaging job. You have to embrace your loses if you have a statistically favorable model. Each loss means that your have more gains ahead.
    You also have to have discipline. If you do not have a set stop loss, you are a trader no more. You become an investor, most likely an unsuccessful one.
    Professional traders say that you have to love your losses more than you love your gains. It is liberating. It dissolves your ego. You realize that the Market is not about being right, it is about making money.
    The bottom line: Any system should have a preset exit points (profit objective and stop loss). Otherwise it is not a system, it's a collection of waisted broker's fees.

    Al
     
  221. has anybody been doing pairs trading on other markets? Europe? Japan? or even cross country?


    peace
     
  222. I wrote this message on Jan 14, 2002. The S&P 500 closed at 931.66 that day. That was the highest close since Dec.02. Today we finished at 861.40.
    Now we can safely say that Jan14, 2002 was an important top.
    The reason I am pointing this out today is not to brag, but to suggest that pairs behavior could be predictive of the market direction. Not that we care about it as pairs traders...
    I closed MU/ ALTR today at R=0.7500 into MU weakness (opened R=0.6950). 7.9% gain in 10 days.
    I am eyeballing AMR/CAL and AMR/DAL now. They look unbelievably oversold. Technically it is a great trade. However, I am not sure about AMR's fundamentals. Are they going bust? Any thoughts?

    Al
     
  223. I have a very small AMR/DAL trade going now. Before this latest decline, DAL had a nice runup while AMR didn't budge much so I'm down a bit right now but holding on. Prior to UAL's bankrupcy and for a time after, UAL was still trading close to AMR's current level so I'm hopeful that the worst is over for the short term. DAL is fundamentally better off than a few of the other majors but I wouldn't call their outlook rosy. So rather than thinking AMR will have a near term recovery, I'm thinking the gap will close with DAL declining.
     
  224. man,


    About Europe, - Continental Europe has much greater volatility - and fewer stocks-so harder to find correlations.
    In London Banks hire Phd's to find the pairs- they dont share info- you just see the adverts for Phd's in the press. Remember we are 5 hours forward and 5 years behind. I am more attracted to US stocks to find tight channel trading for pairs. I dont have the resources to data mine in Europe.
    One point - does anyone else increase stake as R moves away from the mean ? I do on low beta trades. But am I guilty of averaging down ( chasing my losses).
    I suppose the proof is in the pudding and so if it works do it- but I would never do it on large movements - any thoughts ?
    Perugino
     
  225. what software platform do you all use for pairs trading (execution) ?? we currently use rediplus, which is very user friendly and allows you to customize many features. We keep many montages up to quickly execute trades in many different stocks that we use in our pair trading model. We are thinking of switching over to the hammer, what do you guys think of that platform as pair traders??? any other platforms that you are using that work well for pair trading????
     
  226. Does anyone use charting software that has alerts for pair trades??? I currently use Qcharts and it's alerts only work for individual stocks. I am looking for a better alternative than to conditionally format my excel spreadsheet, which really sucks up the juice on my computer. Any input would be appreciated
    Homer
     
  227. Try linnsoft.com. , aspenresearch, maybe neoticker and first alert. With linsoft you can create a custom instrument which then looks, acts just like a regular stock. i.e. You can set alarms to it, apply indicators to it,etc
    good luck
     
  228. Thanks GAT
    I'll give it a try
    I have used some more sophisticated software that will easily handle pair alerts, but it slows my box down to much to use it
     
  229. Could someone please direct me to that article about pairs trading. What month? Thanks
     
  230. The article that introduced me to the topic is "Pairs Trading" by Stephane Reverre. It was in the March 2001 issue of Technical Analysis of Stocks & Commodities magazine.
     
  231. Al,

    I have found the pairs stategy to be very successful but was also having a hard time finding trades, never mind developing a model. I have found a really good source for information is a site called pairstrading.com. They have a software product that helps you to find trades. I would recommend the advanced version though, particularly for developing a model. I have started working on my own model, but the trades I've gotten from these guys have been great.

    Your bud Bud....
     
  232. It is my understanding that there is a discussion of pairstrading in last weeks issue of Newsweek in an article on daytrading.

    Any thoughts....
     
  233. In Neewsweek? Really?


    peace
     
  234. That's what I've been told...

    Haven't checked yet...
     
  235. it's either good ar bad if many people enter. but it will get more attention and that is not so great.


    peace
     
  236. I hate to play dumb, but why is that bad?
     
  237. too much interest of too many people rushing into a narrow room. if it was dumb people - very good. might be that only experienced people enter and that makes the game tougher - very bad.

    crowd is coming - gentlemen, fasten seat belts.


    peace
     
  238. Hey Bud,

    I use the advanced version myself. I am only able to do the S&P500 with my hardware. They have the S&P, the NAS100 and the Russell 1000. That is a lot of pairs to choose from. They also calculate RSIs and Standard Deviations, so you can look for shorter-term trades with signif. divergence.
    For those who interested, here is the link:
    http://www.pairstrading.com/pairs-generator-adv_demo.asp
    The only problem they have is the charts.
    The Yahoo charts they show are not enough. They need to be able to display ratios. If you use Qcharts, however, you should be fine.

    Al
     
  239. Hey Al,

    Thanks for the tip. I have been trying a variety of charting services and trying to pick one. I will try QCharts. As for the service, you are right, it is very easy to use and I find that I can easily customize it to my needs. I usually make more on each trade than I pay for the service, so the lack of charts doesn't bother me that much.
     
  240. has anyone tried to trade buckets on an interest rate curven by means of pairs trading? I mean actually trading the moves within the curve, either by futures, swaps or whatsover.


    peace
     
  241. Man I am asking my question here so that the thread may remain here.
    <BR>
    I was wondering if you would please take the time to post (if available) the links to any papers on the cointegration of stock data series that you spoke of on the other thread.

    Thanks, Royce III
     
  242. Could anyone suggest a broker with a pairs trading platform that I could use ?

    Al
     
  243. Hi Al,

    I first became aware of hedge trading stock pairs when taking some trading training form Mike Engmann, the principal at Preferred Trade and SectorBase and a hedge fund manager. I then did a web search on the topic and found this group. Mike and SectorBase (www.sectorbase.com) have now added a module, EquityScore, that includes hedge trading pairs. It is an extra cost module and I haven't looked at it. I have asked to sample other modules and they have been willing to give a month's access for free. My guess is that you want to ask at the very beginning of the month to get the maximum number of free days. To be completely honest, I don't know if the module is a good trading platform for hedged pairs; perhaps coupled with SectorEngine and Preferred Trade it is - in my novice opinion SectorEngine is the Cadillac of trading platforms - but then, it's only the second one I've used - the other was a real clunker. If my life ever slows down I'm going to look at EquityScore. EquityScore's portfolio returns are fairly impressive, especially when the reduced risk is factored in.

    Hope this helps.

    Jerry
     
  244. we started our pairs system by december 3rd, saw about ten good weeks and got hammered during the last ten trading days, bringing us back to flat for the year. an unusual streak of seven loosing days in a row costed us 3%.

    how did other pairs traders do during that time?


    peace
     
  245. Man,
    I don't know about your system or timeframes but the past few weeks of low volume has made it extremely difficult to generate any significant net profits for me trading pairs. High liquidity & high volatility is essential for shorter term pairtrades in my opinion. I've adjusted by cutting back my initial size and be more patient on entry points. As soon as the volume picks up I think you'll start having more success
    Good Luck
    Homer
     
  246. Man,

    I agree with Homer with respect to the need for liquidity and volatility for short-term pairs trading. I trade pairs from both a short-term and an intermediate-term perspective. The last three weeks of February were very bad. Out of fourteen trading days, I think I was profitable on two or three. I don't have a solution to the problem --- yet.

    SamW
     
  247. what kind of risk parameters do you guys use for pairs? how much heat do you take, and when do you add on to your pairs?
     
  248. Steel

    Steel,
    I totally depends on YOUR risk tolerance. How much are you willing to risk? I make my living trading, so I can't afford ANY sort of extended draw down. I trade each pair using different parameters. I'll add to a position if I am comfortable with the trade. Any other ?'s you can pm me.
    good luck
    Homer
     
  249. we are holding positions ten days - rather long I guess. we are doing this on sixty pairs at any one time. I see your point on volume, yet I do not see too much difference compared to jan or feb.
    we experienced a seven day streak of loss - spells real pain.

    I am considering to add short term intraday trading soon. did very short frames suffer the same?


    peace
     
  250. man,
    I trade mostly on an intraday basis. The last few months have been the most difficult months that I have seen but I have still been profitable. unfortunately decimalization, lack of volume, lack of volatility, and the loss of the daytraders have really hurt my margins. It has been difficult to cover my costs but it has forced me to be more patient and nimble.
    Homer
     
  251. Be careful with adding to a position and holding it for days (unless it's a small position). I've seen many people losing months profit when something abnormal happens overnight (and over several days). Don't confuse good trading with luck and random events.

    I'm trading pairs myself and adding intraday in spreads I'm cofortable with, but never holding it overnight. You'll never know what's gonna happen. Many lost big in fre/fnm last september and pvn/cof in 2001.
     
  252. Homer
    how many pairs do you hold at any point of time?


    peace
     
  253. Man,
    I look at about 10-20 pairs currently. At any one time i'll either be flat or have up to 5 to 10 trades on. These numbers vary based on how things are going. Like I said before, it's rough out there so I am trading extremely defensively right now. Quick and nimble. How have you been doing??

    Homer
     
  254. we started that strategy in december, made 2% in dec, another 2% in jan, then started good in feb, but lost all we made this year within the last week of feb. since then basically flat, leaving us somehow flat for the year.
    we think it is due to a lack of dispersion, thus stocks are volatile but all very similar. this seems to be a bad environment for us.

    we consider developing something on intraday pairs. first i would like to test out some extremly similar pairs like fre-fnm or some banks or oil companies.


    peace
     
  255. I am not a big fan of trading listed pairs. The specialists have way too much latitude on handling your orders. You will get filled on on one side AFTER you try canceling your order because the other side has moved. I think the specialists are playing the game too, except they are the house. Execution isn't a big deal if you have a longer time frame, but it's essential for intraday trades. It can be profitable but I have found the profits aren't worth the hassle of dealing with the specialists in my opinion (in this market).
    But to each their own
    Good Luck,
    Homer
     
  256. Do you trade trending spreads (go with the trend)? It's my impression that nasdaq spreads trend a lot more than listed pairs. Just an impression, i haven't conducted any research on nasdaq spreads, i only trade listed spreads.
     
  257. Large Capital Listed pairs are the only way to survive and make a living trading pairs. I've learned the hard, expensive way!

    Use #1-#4 (by capitalization) within an industry. Plot the 'spread' carefully, go with the 'side' of the spread to return to the center/average (short-term).

    And so on. It works, I've made a reasonable living for almost a year once I 'blew-out' my account several times trying.
     
  258. Such is the nature of reversion trading strategies, you make $ 8/10 but the other 2 times you lose most if not more of what you make. Commodity pits are littered with stories of traders doing these spreads. I personaly clerked for an oil trader who went through the cycle at least three times doing reversion strategies in the crack spreads. Not for the faint of heart and definitely not less risky than straight up trading, as some pro firms might want you to believe. Good luck
     
  259.  
  260. Hi skiabox,

    Yes, Wealth-Lab supports developing Pairs Trading strategies.
     
  261. after our hit in feb we made it back in march.

    2002 03.12.02 100.00
    Dez 101.98 1.98%
    2003 Jän 104.08 2.06%
    Feb 102.14 -1.86%
    Mär 104.10 1.91%

    working on our understanding regarding performance drivers. high dispersion and high vola are best for us. sharpe (3%) of 2.0 so far in real trading is satisfaying and in line with expectations, yet the timeframe is still too short to tell about the quality of the approach.

    one of the next things will be selected pairs intraday.


    peace
     
  262. <subscribing>
     
  263. jester
    thank you for sharing this with us.


    peace
     
  264. sorry ...was subscribing to the thread, is their a better way to do that>

    J-
     
  265. on the bottom read "subscribing".


    peace
     
  266. lost all profits made in march. down 0.3% for the year. no big excuse. tough time. we do not beat the indices.


    peace
     
  267. man,

    How do you pick your pairs? Just curious.

    I've been pairtrading intraday for some years, but implemented my own swingtrading system to get a better risk/reward. I've picked 35 pairs based on certain criterias. So far, it's been going pretty well.

     
  268. For those of you who got stuck on the wrong side of a FNM/FRE pairs trade last year but decided to become an investor and stick it out, are you smiling yet? After the price difference between these stocks fell to about $5 a share earlier in the year, it's now over $15 a share, up near it's highest level of the past year.
     
  269. holding a pair and becoming an investor is stupid- you can make much more and avoid the pain by trading the moves in it....
     
  270. a recent and historical correlation between

    QQQ + SMH ?

    thanks ...
     
  271. Here you go!
     
  272. Here's 250 trading days.
     
  273. Here's 60 trading days.
     
  274. for the charts guys
     
  275. Hi All,

    I have put together an informational web site on pairs trading. On the site, daily, is posted one or more equity pairs that is currently in an identifiable trading channel. Todays post is a pair of bank stocks.

    However, the real gem of the site is the "Pair Finder" tool. Sign up for the free trail and and tell me what you think. Goto: http://www.pairfinder.com/ to check it out.

    Thanks Royce:)
     
  276. Looks like a nice site. Similar to www.pairstrading.com. I like that "The Pair Finder" allows you to select the analysis period and filter out stocks with low volume, market cap, and price.

    Incidentally, there's another site, www.pairsfinder.com, that I accidentally went to. It's under construction. Will be interesting to see if it has anything to do with investments.
     
  277. You might also look at www.pairtrader.com. Good info and products, etc.

    Let's keep up the good places for research/education for pairs.

    Thanks.
     
  278. Pairtrading is something I'm going to be doing shortly. I checked out pairstrading.com it looked ok.
     
  279. I see many people discuss correlations and some choose to use price while others focus on returns.

    I use returns due to an argument that I have referred to as the "law of one price" which simply suggests that comparable assets with comparable risk/reward characteristics (2 companies in the same industry, for example) should return the same reward. As a result, a $10,000 investment in 2 oil drillers cannot have hugely divergent returns else there would be an arbitrage.

    Those that focus on prices would say that a $10 and a $50 stock that both went up and down by $1 over a time series would be perfectly correlated despite the fact that the returns our grossly different.

    Anyway what do you other pairs/stat arb guys do?

    Also, how do you set up your hedge / position ratios? I like beta neutral positions where I compute the beta of my long versus a single unit of my short.

    Looking forward to your thoughts.
     
  280. I think any references to price in this thread are just for simplicity sake. Most pairs traders are actually looking at the ratio of stock prices. So with a 10% increase in price on stock A, all other things being equal, we would expect a 10% in stock B. The www.pairfinder.com site does base their analytics on offsetting betas as you indicated. My question is, how much of a difference does this make? If your pair is two major banks, for example, is their difference in betas that significant? If so, I might question whether they are good pairs to start with.
     
  281. I disagree that most traders trade ratios (C divided by JPM) versus differentials (C minus JPM -- with a multiplier factored in to the differential -- Like (C*.7) - (JPM *1)).

    Differentials by far better reflect the real risk of a spread in true price changes versus percentage changes.

    I'm a professional pair trader doing this for a living. Know your risks.

    Surely you can trade pairs via ratios but for new traders you'll find differentials to be easier to understand and define risk. Ratio's may be for more long term PROFESSIONAL pair swingers or investors, but even then you can get lost in the apparent lask of risk in a ratio chart. Look at the losses the pairstrading.com guy has been taking lately. He trades ratios and for the life of me I don't see where he defines his risk.

    Seems to me some ratio traders assign the same capital size to each pair via ratio analysis, but when you look at the differentials you can clearly see that some pairs are far more risky than others and therefore you'd trade smaller size on these pairs. Varied bet size is critical.

    Just something to think about. More info found at www.pairtrader.com and others.
     
  282. sslaven,

    Since you obviously have experience in this area, let me pose a few questions to you. I've been learning on the fly so to speak but would like to compare notes.

    1) What sort of holding period do you typically experience for a trade?

    2) When your trade is going against you, do you typically take your losses at a set point, hang on and wait it out, or increase your position size? This assumes there isn't a fundamental reason for the pairs relationship to change.

    3) Do you normally trade based on convergence or divergence?

    Thanks.
     
  283. Sure...I'll give it my best shot.

    1 &2) I'm a prop trader with Bright Trading. My comish is less than .75c/sh. What I do many can't simply because their costs are higher. Comish is definitely an issue in any strategy, but especially pairs since you have the long and short to put on and off.

    I scalp all day long. I have defined biases, defined ranges that I want to play in, and defined risks on every trade. I do not suggest the "building into a position route till your account is on the brink of destruction - method". I do add 2 legs to almost everything. Lately, it's hard to make money without having 2 legs when you're scalping for 30-50c all day long.

    Some of my account is dedicated for swings. Those last 2-3 days. More people are going this route lately since intraday ranges are so poor or trendy.

    I hold positions when I know there isn't news on the pair. If it's not business as usual, I bail when trouble sets in.

    I trade 5-10 pairs at a time. This helps too. Always something going on. Lately, I've been cruching more -- leaning on one side and only putting on the other if I have to. But each day is different.

    3) I normally fade moves. Trading trends in pairs is tough if all you've done is fade for years (which I have). It's easier with anticorrelated pairs to trade the trend. I fade into support and resistance. I fade on quick retracements into a bias. I do my best to space out the distances between legs when they trend. You have to know the bigger picture -- as you can get blurry eyed in a shorter 2-3 minute time frame.

    Hope that helps a bit.
     
  284. Do you trade in the direction of convergence or divergence of the pairs?
     
  285. I think it just depends on your time frame you trade in. Let's say you do your research and conclude on a 5 year chart of a pair that the two stocks difference is "out of whack." Say...at an extreme relative to their past. Yet the fundamentals suggest a convergence back to parity over time, and the chart suggests you can enter a trade at this logical resistance...then yeah...I fade the trend (the trend which has taken it to the "out of whack point"). Am I losing you... :eek: But there's nothing saying you can't just wait for the trend to reverse first before guessing the trend is over -- right?

    You form your fundamental and technical bias...then you wait for the trend to go in this bias and you strap it on! There's lots of ways to do this as you can see.

    So, if you come to this same conclusion on your monthly or 60 minute charts the same game can be played. Fade or trend? It depends. The 60 min chart says the trend is starting, but the 2 minute chart says this is reeeeealy extreme here to fade IN THE DIRECTION of the 60 min chart...then I play the divergence.

    I think there's a big problem with defining your bias time frame (say daily) but playing 2 min charts. They don't mesh. You should have 5 x the time frame. Find bias in a 10 min, yet trade the signals in a 2. Find bias in a weekly, yet trade the signals in the daily, etc.

    Rambling...sorry.

    In general, you'll find the normal pair trader FADES plays more than trades trends. We expect correlated stocks to come back to parity. But there is NO rule that says you can't or shouldn't play the divergence when your analysis suggests. As a matter of fact, there's tremendous money to be made doing just that...but you have to be open to the idea. Most pair traders aren't.

    Hope that helps.
     
  286. Looked at the www.pairsfinder.com site and its under construction. Does anyone know how long its been offline?
    My free trial to www.pairfinder.com is almost up and I'm going to subscribe. I've used the "Pair Finder" section of their site to determine return correlation relationships between several stocks in the same industry. Return correlation seems to be a more accurate measure because it doesn't treat the moves of a $5 stock and a $20 stock the same, like price correlation.
    The site also provides a market neutral hedge ratio that has helped me even out the risk of my shorts and longs.
     
  287. I believe the earlier post said that "pairsfinder" and "pairfinder" are registered by the same person.
     
  288. If there are 2 functions (in our case 2 stocks) then the best way to find out if the functions related to each other is to find correlation between them (that’s mathematics). You of course may try to find correlation between secondary functions and in some cases the result will be close.
    For example, if there are 2 stocks (A and B) and during 3 days period they move like this:
    (A): 3,6,9 and (B): 2,4,6 then both correlations (price and return) would give you 100%.
    However you don’t trade secondary functions (in your case you trying to find correlations between 2 differentials divided by functions itself). You trade the stocks – primary functions. Even if the secondary functions will be highly correlated, stocks themselves might behave quite different.
    Here are some examples:
    Let say stock A goes like this: 10,1,2; And stock B goes like this: 10,9,18
    I would not say these are highly correlated stocks. Return Correlation however would give you 100%; but Price correlation would give negative (!) 10%
    Some say that the most important that Difference between stocks should be in some range. If it goes on edge, then, based on historical data, we might go into trade. Yes, we might, and in some cases we could make money. However, good thing about correlated stocks that you are protected (kinda) against market moves. If you trade correlated stocks you are market neutral (and then - range may go into account). Let’s take another example:
    (A): 10,11,10,9,10…; (B): 10,9,10,11,10…
    Difference would be between –2 and 2. However, these stocks correlated with negative 100%. Based on historic data we may trade this pair. But, in that case we as well might trade these stocks just by themselves. They don’t need a pair.
    Can we make money here? Yes. Are we going to be protected against market moves – not in this “pair”.
     
  289. Please help me with my definitions of Price and Return Correlations. I understood Price Correlation to be a $ for $ comparison of two individual stocks: (to use your explaination method) If stock (a) 3,6,9 and stock (b) 2,4,6 then stock (a) moves 3 then 3 and stock (b) has moves of 2 then 2. So in my understanding of Price Correlation, (b)/(a) would be 67% correlated. Your defintion seems to be a direct ratio comparison of (a) and (b) over a common time period: (b1)/(a1), (b2)/(a2), (b3)/(a3) then since the ratio result is constant throughout the time period, they are 100% Price Correlated.

    Now for Return Correlation, in my understanding, is the change of (a) divided by the price of (a) compared to the change of (b) divided by the price of (b) over common time periods. If (a) is a $10 stock and moves 10% and (b) is $5 and also moves 10% then my definition of Return Correlation would result in 100% correlation and 50% price correlation.

    I appreciate your reply and welcome your tutelage. I may e-mail the guys at www.pairfinder.com and see what their definitions are since I've been using their site for correlated pairs.
     
  290. Quote from Fraze14:

    Return correlation seems to be a more accurate measure because it doesn't treat the moves of a $5 stock and a $20 stock the same, like price correlation.

    --------------------------------------------------------------------------------
    In reply to Fraze14 -- again, price correlation can accurately measure movement IF you trade them with a factor. For stock A at $5.00 you buy 4x more than stock B at $20. Of course, these can move over and over again out of their ranges and re-capitalizing would make sense. But the same can happen with return correlation -- in that they can easily get unbalanced, and therefore not market neutral. As a professional daytrader, I could care less about % return. I measure things in cents and points just because it's easier to compute in my head. Each person may approach this differently. For investors, you may be more used to % return.

    Like any other type of investment, two pairs are not created equal. One can be very tame and the other a wild child. % return to me seems to disguise some risk just because the relative movements seem so small.

    I'll have to look at a few intraday in %return format to see how this adjusts my thinking, so that I can more clearly respond to it. I've been doing this for years, and frankly $ return works for me.

    You might look for more info at: pairtrader.com, pairstrading.com, pairfinder.com, etc. I'm sure each of these have a different approach to it -- and like all trading strategies you'll have to grab on to one of them and just see how you like it.

    sslavin
     
  291. Ok -- I've looked at both a price difference chart and a % chart (or ratio chart) of MWD:MER and MWD:GS. I use a 1:1 differential for MWDMER and a 1:1.5 differential for MWDGS. This was on a 10 min chart over 5 days.

    First off -- what would visually confuse me is the small numbers in % than float by. I'd have to really get used to this. We're talking hundreths of a %, yet for the differential this is points moving by. I watch 30 pairs a day and frequently have 10 or more on at one time. These similar numbers would likely confuse me. I'd have to set alerts for everything (sup/resistance) and not rely on my memory as to what number are important.

    Second -- a $3.00 point move in MWDGS150 is the same as .02% move. For MWDMER100, this same .02% move is $1.00. So, in % terms you might be fooled into thinking that MWDGS is the same type of risk as MWDMER and let me assure you -- it's not. MWDGS is a much wilder pair to trade.

    So, how do you address this issue again, and why are ratio charts your visual choice for pairs? Thanks.

    Do you all trade stocks or the S&P's in % terms or % charts? I have yet to see a professional trader do this and I have a pretty vast library of books on traders. I see a reference from time to time when one compares returns of sectors against eachother, or when a fund says we're running 2% above the S&P's for the year. But intraday trading in % terms is a new one I'd have to look at.

    Thanks. Good posts going.

    sslavin
     
  292. There is a formula for correlation for any 2 functions. It’s pure mathematics. This formula is not very complicated (no integrals or differentials involved) but not the easy one either. You can search for it on the Internet.
    That’s not the point. The point is –what functions you use for this formula. In Price Correlation you use 2 functions – 2,4,6 and 3,6,9. Correlation for these 2 functions would be 100%.
    In Return correlation you also use 2 functions – but the different ones. Because the stocks go 100% up on the thirst day and another 50% up on the third day in both cases – Return correlation would give you also 100% here.
    But consider the previous example:
    10,9,18 and 10,1,2 – if you looking for return correlation - first stock goes down 10% and then up 100% and the second stock goes down 90% and then up 100%. Believe it or not for these 2 functions Correlation (in this case Return Correlation) would be 100%.
    I’m a programmer myself and trade pairs for fun. I found www.pairstrading.com back in December and since then I back tested their results since 1998. It’s seems like the long weekend already started, so early next week I’ll give some results.
     
  293. Hi, all.
    It looks like pairs trading model might do much better then combined with ADX indicator.
    I’d like to back test it. Formula for ADX is pretty simple for a stock. However, what are High and Low for a pair? Obviously, they could not be derived from the Highs and Lows of 2 stocks that are in a pair. Do I really need all intraday data for all the stocks (I’m interested only in SP500)? Or I need it per minute / 10 minute / an hour? And if so, where can I get it? Any thoughts?
    Thanks
     
  294. A study of FNM/FRE over the past year, a highly correlated pair, should be a must for anyone considering pairs trading. It highlights one of the biggest problems I have had with pairs trading. A very high success rate (90-95%) with some killer losses mixed in.

    I'm long a mere 100 shares of FRE at this time, but that was just good luck. Could have been many more shares as a spread of $15 seemed to favor holding FRE.
     
  295. Yeah,

    The study would tell me that I WOULD BE INSANE TO TRADE IT.

    nitro
     
  296. This is a great example of the pitfalls of ANY speculation. I don't care how great you think the trade is, in the market you have to work based on probabilities with no guarantees.

    Pairs can do weird things -- things you never thought possible. Each pair has a history, and a range, and only you know what pair may fit your personality or style. I've been trading FNMFRE for years. It requires near perfect order execution if you're a scalper. It can pay you very well, and cause you lots of pain. It's not for everyone.

    What's especially important is that you do your homework on each pair and decide whether your pair is business as usual or not. If it's not, than it may be best to simply not trade that pair till it calms down. Or you could specifically look to trade that pair as an anti-corrlelated one and trade only the trends. If you had done this on Monday and Today on FNMFRE you'd have done very well. But again, the risks were high -- so it may not have been right for you.

    Best Regards -- sslavin
     
  297. What is your profit objective?
    FNM/FRE could give you a few percents here and there. I’m not in a trade if I don’t expect in return at least 10%. And it pays off. On 5400$ account I’m up 16 percents for the year (including commissions - 40$/pair on datek). My average position is about 1200$.
     
  298. Have you considered a journal describing pairs trading strateges.
     
  299. It's coming :)
     
  300. I used to make a living doing this stuff. I have made thousands of pair trades. I cannot agree more with the poster above that said that to trade this (FRE/FNM) spread INTRA day, you have to get perfect executions. When I tried to trade it longer term, I got killed.

    What is funny is that today I would have a better chance at getting right, since I have the use of software for executing in tight spots.

    All trading has gotten tougher, but pair trading, given the returns recently, seems to be a diminishing returns business to me.

    nitro
     
  301. I think nitro is right in claiming that pairs trading shows diminishing returns. this gaim is really tough.

    we started a pairs strategy on daily data with an average holding period of ten days and sixty pairs traded at any point of time in december of last year. tests would indicate a sharpe ratio of above two. so far we traded about 1.25.

    03.12.02 100.00
    Dez 101.98 1.98%
    Jän 104.08 2.06%
    Feb 102.14 -1.86%
    Mär 101.66 -0.47%
    Apr 102.15 0.48%
    Mai 103.98 1.79%
    Jun 105.29 1.26%


    we currently work on intraday pairs strategies, but we struggle to make it profitable for the period of the last six to eight months.

    this is a tough, tough game.


    peace
     
  302. ... where is future returns.

    nitro, where do you see increasing returns? which strategies do you think will outperform with the next let's say two years?


    peace
     
  303. Overall market is in a trading range. Whenever we break out of the trading range maybe pairs trading will again be worthwhile.
     
  304. one big problem with trading pairs (especially intraday) is that you have double the volume for each trade. With decimalization and the historically low volatility it has gotten tougher & tougher to cover expenses. I am encouraged about the recent push in the market because I think higher stock prices will translate into bigger moves in the pairs. I also think that as we top out (I think we are close, but who knows :confused: ) we will see the volatility levels take off as well which I think is the key to making money (intraday) in pairs. I do think that the migration of the majority of day traders out of stocks and into futures has hurt as well. just my 2 cents
    Good Luck,
    Homer
     
  305. any decent books on pairs trading.
     
  306. Intraday pair scalping has been unreal lately. It's very tough to make anything without having multiple legs...it's possible...but it's tough. It requires a lot of patience and multi-tasking. I'm trading at least 5 at a time -- going after 25-75c moves. Tough stuff. Overnights haven't been too bad. Swing trades are consistent as long as you avoid the story stocks (or trade the trends on them).

    Good books to read? I think the guys from www.pairstrading.com has a manual. I know the guys from www.pairtrader.com have a huge manual -- specifically geared for daytraders and active traders. I know of NO other published works that specifically focus on equity pairs.

    Good trading.

    sslavin
     
  307. Each of these books has a chapter or two on equity pairs trading:

    http://www.amazon.com/exec/obidos/tg/detail/-/0071359958

    http://www.amazon.com/exec/obidos/tg/detail/-/0971853649
     
  308. thanks
     
  309. IMHO,

    In the current environment, the "opportunities" are in what the market is keying on:

    1) Dollar/Euro
    2) Bond "Bubble"
    3) Natural Gas/Oil

    Trade idea for 1) Take a look at Adidas Solomon AG/ NKE Pair
    Trade idea for 2) much harder, but look at the coorporate bond spreads vs the NOB spread
    Trade idea for 3) Tons of companies to choose from.

    nitro
     
  310. Does anyone know what is going on with the pairfinder site. It appears to have shut down today.
     
  311. nitro
    I like the idea to follow the action in order to profit from inefficiencies happening then and I especially like the idea of pairs trading Euroland against US, expecting negative surprises once these companies realise the rate change in their books.

    I think the only way to overcome the problem of high hit ratio with low profit factor in pairs trading, the latter due to big losses now and then, is bigger portfolios. This, in most cases, will require a systematic approach.


    peace
     
  312. nitro
    I think after so much turmoil in the corporate credit market in the last two years there are quite many distortions, yet the spreads for corporate credit have come back substantially already. I think it is quite possible to play the US curve, yet I have found no systematic approach so far, that makes me happy.

    What are you trading? Index futures? You indicated that you are not so much in stock pairs any more. Are you playing futures spreads?

    peace
     
  313. Here's a pairs trading recommendation that came out from an analyst this morning. Naturally, I'm on the wrong side of this one. A good example of where playing divergence based on the fundamentals would have been better than expecting convergence based on technical analysis.

    7:46am 06/13/03
    UBS suggests 'pairs trade': buy Federated and sell May (FD, MAY) By Tomi Kilgore
    Analyst Linda Kristiansen at UBS Investment Research is recommending a "pairs trade" -- buying Federated Department Stores and selling May Department Stores -- given that the companies' strategies and execution "are diverging to a greater extent that we can ever recall." Kristiansen upgraded Federated (FD) to "buy" from "neutral" and lifted her price target to $44 from $30, citing increasing confidence in the company's relative growth rate, and belief that Federated will continue to gain market share at the expense of May and others. The stock closed Thursday up 40 cents at $36.58. She also downgraded May (MAY) to "reduce" from "neutral" and cut her price target to $18 from $20 on expected sales underperformance and margin pressures. The stock is sliding $1.08, or 4.7 percent, to $22 in Instinet pre-open trading. She noted that May was implementing a more "clearance-driven" strategy, while Federated was "dialing down" promotional activity.
     
  314. I've tried trading pairs both as mean reversion and trending. My research and experience tell me that mean reversion is the way to go (I only trade highly correlated stocks, not anti-correlated). I put on one unit, add one unit, take off one unit etc, trading only the same pairs every day and keep at least one unit 70% of the time. I only trade NYSE except for one Nasdaq pair. In my opninion Nasdaq pairs are not suited for pair, at least intraday.
     
  315. Does anyone know of a site which gives pair recommendations?
    Pairfinder.com used to be one until it shut down.
     
  316. learn to think for yourself- sheep that follow the herd get slaughtered
     
  317. Pasternak,

    FD and MAY have a correlation of 93%, so presumably this is one that could have (might have) met your requirements. Would you handle a sudden change in spread/ratio as an opportunity to put on another leg or just wait it out and see what happens?
     
  318. No way I would trade FD MAY at all, in my opinion it's too volatile. Several dollars a week is more than my stomach can handle..... I prefer low volatile spreads, FD MAY is a wild beast.

    I've never looked at it before now, but there's not enough "ebb and flow" in the pair for me to trade it when looking at a 5 mins chart. Actually the pair seems to trend a little bit, but i guess still difficult to trade that way. I don't trade retail at all.

    So, I don't know what to do in that situation. I usually scale in every 10 cents, wait 5-50 mins, take one unit off etc. Of course, sometimes I end up in a trending spread that day, but then try to adjust by offloading long or short position according to the market (S&P).

    I'm still learning (done this for two years - never had a down month), the market changes from month to month. The NYSE prints are less frequent now, that's why i scale in and out.
     
  319. Pig, er Hog,

    Thanks for your deep insight, but it really was of no value to me.
     

  320. Reg, er the sheep

    not only can you not think for yourself since you look for the easy way out- i.e. someone else doing all the work and telling you what trades to take on your 100-lots , you are a smart ass too- you're going to go real far in this business, buddy.
     
  321. You've got something working then.

    What wrong with trading FD/May with a small amount of shares if it's more volatile than what you're used to?
     
  322. Great insight again Boss Pig. Very profound words indeed.
    Don't bother to reply since you're now in my ignore list.
    Have a great life and don't get slaughtered.
     
  323. I have a question, gentleman. the correlation between FNM FRE is quite extraordinary. I once looked at five minute data and that was still correlated at 0.65 or something. Now my question is: where does this correlation come from? it must be some kind program trading, but how do these people handle the spread divergence from time to time?


    peace
     
  324. Nothing wrong with trading less shares in volatile spreads, I do that, but FD/MAY is not my cup of tea. Just by looking at the 5 mins chart over the last month I know instantly this is not what i want to trade. It doesn't "regress to the mean" very often. That's why I stopped trading FRE/FNM last september. Too volatile as well, but also less mean reversion.

    Some of my best performing pairs correlate less than 0,8 on a daily basis. There's a lot of good pairs hardly anyone trades. Spend time looking at volatility in sectors instead of just correlation. Very high correlation is no guarantee for mean reversion (my best pairs are around 0,8).
     
  325. Hi everybody,

    I am glad that pairs trading is of interest to so many people. It is impressive that this thread received more than 300 posts and over 40,000 views.

    My experience with pairs trading has been tough but quite profitable so far. It took me a while to find appropriate pairs to trade and pinpoint favorable entry and exit points.

    Recently I have implemented an options on pairs strategy that work well with low underlying market volatility.

    I tried intraday pairs trading and found it to be a waste of time and money (too much noise, stops are too tight, high commissions, and small profit objectives). Positional pairs trading with profit objectives 10-15% works the best for me.

    I have attached some of my favorite pairs that I have successfully traded over the past 9 months.

    Good luck to all. Keep it going...

    Neutral Al
     
  326. Another pair...
     
  327. And another...
     
  328. One more...
     
  329. And the last...
     
  330. I don't understand what these companies have in common that makes it appropriate to trade them as pairs. I mean, I see the charts, but I'm wondering how long the strategy has worked for you, whether you use leverage,etc.
     
  331. Hi thetradeprofit,

    What these pairs have in common is high correlation over several years. This high correlation allows you to stay truly market neutral as these stocks react to the market moves the same way over a significant period of time. The correlation is much less over 30-60 days, which tells you that the stocks are diverging at this time.
    The stocks that are highly correlated over short and long periods of time are nearly impossible to trade because the divergence (the spread) is minimal. The fact that the stocks are in the same industry or sector does not mean much because the stocks could be highly correlated and stock A would consistently outperform stock B creating a trend. My statistical model shows that oscillation is more important than any other parameter to create a high probability trade.
    What I look for in pairs are high correlation over longer periods of time ( ex: 2 years) and predictable oscillation around the mean (moving average) over the same period of time.
    One could come up with a million reasons why the stocks oscillate around each other. The point is that if you assess statistical probability of each trade based on the broader markets, then neither underlying fundamentals nor the behavior of individual stocks matter.
    This strategy worked well for me over 9 months now. I use no leverage in my trades.
    I hope I answered some of your questions.

    Regards,
    Neutral Al
     
  332. I'll be interested in seeing how it plays out for you over a much longer period, say five years. I guess we will find out.
     
  333. Hi, thetraderprofit
    It’s interesting, you mentioned 5 years period.
    It also seems that Neutural Al uses www.pairstrading.com for his analysis.
    In fact I back tested exact the same strategy www.pairstrading.com seem to be suggesting for 5 years back – since 1998, so I can say, you would be in the positive territory.
    Using very simple trading model – enter the trade with extreme RSI and hold 7 days for example.
    This strategy might be good for some hedge fund. With a lot of trades they would be able to make some (good) money on a fraction of a percent per trade.
    That would not work for retail however.
    As an individual I’m going for trades with potential high profits and I’m using some additional indicators – ADX for one.
    Because I created my own model I do not use pairstrading.com, but here is some thoughts about the site:
    They have 2 services: PairsGenerator and PairsGenerator Advanced.
    I find PairsGenerator useless and even suspect one would be in the red trading ALL the pairs it finds.
    PairsGenerator Advanced however has some useful indicators – like search for pairs with extreme RSI or RSI reverting from extreme.
    Last time I checked they gave PairsGenerator Advanced for free for one week.
     
  334. 15:06 ET Pair trade call in the brokers
    Hearing a pair trade call to short GS and go long LEH. According to this research call, GS is trading at a 47% premium to LEH on price to tangible book basis (2.8x vs. 1.9x) and a 32% premium to forward 04 earnings at 17.5x vs. 13.2x, respectively. The gap is significant given LEH's increasing market share in both Fixed Income and Equity trading. In addition, the 4th round of Goldman IPO shares unlocking on Thursday; the last time this occurred (24 mln shares on 5/08/03) the stock was off on significant volume.

    nitro
     
  335. can anyone do backtesting on a pair? if so i will share a strategy i return for the backtesting work... PM me
     
  336. I have done a lot of backtesting of pairs trading and as for all trading strategies there are periods that it works perfectly and there are periods that the strategy generates losses. And you don't know when losing period comes for a specific pair. I was not able to find a magic pair that is constantly profitable over long period. Pairs trading is not significantly different from simultaneously trading long and short system and selecting log and short stock based on some other criteria. And also the correlation is not the last measure for the profitability.
     
  337. Correlation just give you some kind of protection against market moves - Market Neutural
     
  338. I can do that, but probably not before the weekend. My current approach is very short term moves, you might even call it scalping pairs...
     
  339. Test
     
  340. thanks for the replies, i resolved the issue however...
    did anyone take the gs/leh pair... well have to see if the lock up sinks GS tommorrow, i wa more interested in that aspect of teh trade vs. the valuation call, that to me is more pairs investing and i am not an investor...
     
  341. This thread seems to keep returning to questions regarding the nature of markets (general), the process by which future moves can be forecasted and the best way to exploit those moves.
    To look for pairs which " consistently make money " is in my opinion not the question. Why? because of the dynamics of the markets. If an opportunity is seen it will be exploited - fast. The market is not a static being and therefore analysed as such, seeking fixed laws governing future behaviour. It must be seen as dynamic - reacting to many pieces of information at any one time.
    So where does that leave us ?
    If our hypothesis is that R regresses to the mean - sure we need to test that hypothesis by "back testing"
    We can test SCIENTIFICALLY correlation, standard deviation etc. etc.
    and what does that tell us ?- the probability of future events happening based on historical events.
    So armed with this information we can enter the market with the hope of exploiting it.
    Always remember you may be wrong.
    At the end of the day it is only a bet.
     
  342. Sure - everything's a bet in the markets. But pairs trading is a higher probability bet than trading flat price. I scalp pairs and butterflys all day long. It works if you can trade size and your execution costs are low.
     
  343. bone
    how low are your costs?


    peace
     
  344. Institutional.
     
  345. aha. and what does "institutional" mean - you do not know, since someone in the backoffice is doing that, or you do not want to tell, because you are doing trades with brokers that let you trade some issues for free if you do enough volume?
     
  346. "Institutional" means I clear a large bank directly, it's cheaper than IB, and I am under an agreement not to name names or rates.

    But the real point is that intraday pairs trading is profitable on a consistent basis - providing your execution costs are reasonable. If you do this on an intraday basis, you're going to generate alot of volume for modest gains. But that's the whole point. To consistently generate income. Slow and steady wins the race. Not glamorous, but effective. I mean, that's what we all want. To generate enough income to trade for a living and support ourselves (and for me a family). I can't sit there and suffer through huge flat-price P&L fluctuations during the course of a business day. Don't have the guts for it.
     
  347. bone
    see your point on rates and respect your point - though you are on an anonymous internet board.

    we are currently testing an intraday pairs strategy and start feeling good with the test results. and, as you say, execution cost is key in high frequency trading. i always wondered about the correlation of fnm-fre on five minute data. must come from somehwere. styles like yours will answer that.


    peace
     
  348. There's only one way to find out - to trade it. You can pencil-whip it to death - I'm good at it, an engineer by training. But you have to get your nose dirty to really see if it pans out.

    Just trade them and try to build up some capital. Once you establish a volume pattern, you should be able to shop rates.
     
  349. My experience (somewhat limited) is that slippage is far more significant in pairs trading than transaction costs, particularly when you exit.
     
  350. bone
    agreed. we have to trade it. good luck with your trading.


    peace
     
  351. To jrweiner and other,
    As you pointed out, slippage can be significant. I have two questions:
    1. Under what size of the trade can the slippage be insignificant?

    2. How to deal with the slippage? I am think about using limit order and choose a most liquid time.

    Thanks.
     
  352. I have read articles and talk to traders about the proper way to set up the hedge on a pair trade. When it comes to stocks that are part of a takeover it is simple but how about correlated pairs such as KO/PEP Home depot /Lowes or even SPY / QQQ. I know some traders use $ equivalency method so if SPyis $100 and QQQ is 40 they would buy 100 QQQ and sell 40 SPY since both are $ equal i.e. $4000. Does that make sense?

    should'nt the primary consideration be the volatility or true range of the 2 instruments? In my example above if one were to set up SPY/QQQ pair at 40% and if both have $1 range for 30 minutes, then it would be an incorrect hedge-exposing trader to a larger PnL due to the fact that QQQ might be 40% the price of SPy but have shown $1 range for the past x minutes?

    Can anyone suggest a way to hedge such intruments>? Thanks
     
  353. GATrader,

    The first method that you suggest is probably the most common, but I feel more strongly about making a volatility match. Of course, the volatility sample and history window should match your trading timeframe.

    I trade pairs very short term. Clearly, any dollar balancing or volatility matching is based on some recent historical average or spot calculation. Will the volatilities from the last two weeks hold up in this next hour? I doubt it will hold up exactly, especially in longer timeframes, as volatilities tend to cycle. In fact, if I get these things perfectly hedged for that given time period, I'll go broke paying commissions on flat trades.

    So, it's my opinion, that some uncertainty or slop is necessary in sizing these pairs. For my purposes, I get my answers by backtesting the combinations to verify that it has actually worked in the past. That helps me arrive at my final sizing. Sometimes other factors come into consideration, such as....I don't do odd lots, so using your QQQ:SPY example above, I'd be limited to a 500:200 ratio as a minimum. Based on risk evaluation of this pair, let's say that I'd want to max out at 4 units (buy or sell this thing 4 times if it moves against me). I may consider a 2000:800 share position unacceptable, so I'd look at alternatives. I'd test a 200:100 ratio, 300:200, etc. Oftentimes, these unlikely sizes will perform more favorably.

    If you don't have access to backtesting software like Tradestation or WealthLab, I'd consider getting it and learning how to use it. Yes, some may find that a lot of work, but the rewards are well worth the effort.

    For anyone that's considering pair trading, this world of statistical arbitrage is full of very sophisticated traders. Trading in this niche is not going to become simpler, and going in without the proper tools will be like using a hammer and chisel when your competition is using laser cutters.
     
  354. I use the Bloomberg to calculate currency and volatility-adjusted hedge ratios. That's what made it so popular for bond traders in the 1980's.
     
  355. Thanks for the reply bone and Dave. Does the bloomberg help in the ratios of stocks as well? In addition, can someone suggest another crude method of estimating the hedges since I can't afford a Bloomberg Term.
     
  356. Try iVolatility
    http://www.ivolatility.com/options.j?ticker=spy&R=0&top_lookup__is__sent=1

    For example, using SPY above, Historical Volatility 10 days is 17.97%.

    So, 0.1797 / Sqrt(250 days in a year) = 1.137% in a day

    Multiply this times the price of SPY, $100.24, and you get a 1 standard deviation move for SPY of $1.14 per day.

    Do the same for QQQ, and you'll see the ratio is approximately 2:1, that is, SPY is about twice as volatile as the QQQ, on a dollar balanced basis.

    If you are looking at a longer timeframe for trading, consider using Historical Volatility for a 30 day period.

    Also, HV's will change over time, so be sure to update your ratios periodically.
     
  357. Let me see if I get this straight, you'd want to establish a ratio so that a 'typical' move by one instrument approximates the other leg. So if abc moves on avg 30 cents per 10 minutes, you'd want to extablish an opposite hegde approx equal to that range on a single unit. Example : pair abc with xyz given xyz has 10 cents avg true range, you' want to hedge abc with 300 abc thereby equalizing their true range. This is entirely diff. that $ equalizing, more like vol equalizing. Bone posted this a while ago -use dv01 to establish nob, bund bobl, fob ratios not $ equivalent. Am I correct?

    Thanks again guys. Last question, given the prevalence of auto
    spreader software such as TT and other prop systems, I feel that someone equpped with a mouse point and click Jtrader would stand little chance of putting on spreads on a scalp basis since it would be no match for those systems. I've tried to leg into these spreads and always get bad prices, even though I think I am reasonbly fast, probably no match against the auto spreaders,etc. I feel that my only chance is to lenghten the spread trade fractal i.e. > 10-40 minutes -play the tide not the ripples so to speak. Does that assumption make sense? Thanks again
     
  358. The descision/calculation of what ratio to use is intimitaly related to whether the pair is out of whack or not in the given timeframe.

    nitro

     
  359. If you are (intraday) spread trading something that does not have the uptick rule, e.g., SPY/QQQ, then your disadvantage is not great. If you are, e.g., equities, your best bet is to concentrate on 2 to 5 pairs at most and get a one month conversion on each side and pray they move enough to justify the added cost of the conversions. This, IMHO, is the achilles heal of this strategy for equities...

    nitro
     
  360. Correct me if I am wrong - and I'm sure that somebody will - but who is in back of this fairly recent emphasis on pairs trading?

    Isn't it firms that earn a large part of their revenue from commissions? The very same folks who brought you the strategy of scalping for fractions circa 1998-2000, whereby a zillion poor suckers enjoyed all the benefits of carpel tunnel syndrome while spending 20K a month in fees for their "employers"?

    But hey, you got to be a "professional" with a license(!) to trade there, so they MUST know something you don't, right?

    Anyone get the picture?
     
  361. Essentially yes, you'd like to equalize these to 30 minute, 60 minute, or whatever movements are in line with your timeframe. The profitability of this method comes with the timing of each ones move. That is to say, statistically, these two tickers have demonstrated historically a greater probability of moving the same *dollar* amount in a 30, 60, or X minute timeframe, if you have balanced the pairs as we are discussing. And, yes, I agree that this is really a volatility balancing, rather than a dollar balancing.

    Going to nitro's statement, when the spread gets out of whack, or in other words, when stock A has moved, and you are statistically confident that stock B will follow (or stock A will return to it's original state), you could adjust your ratios. However, I'm less sure about whether it'll be stock B or stock A that will move to bring it back in line. In some cases, if you are trading pairs correlated with the SPoos, then say, stock A has moved to follow a selloff in the SPoos. It's highly likely that stock B will follow rather than stock A returning, so it may make sense to either sell more stock B or buy less stock A.

    As to your statement about automation, I scalp pairs both manually and using automation software. As with any automated program, it's not magic, and you should be able to do the exact same trade by hand. The automation just let's you do it faster and trade more of them at once.

    Where I used to get caught up, especially in the shorter timeframes, is not with slippage, but just with figuring my entry price and exit price. If you figure that you will cross the market, i.e. hit the bid and lift the offer to get into the spread, you should look at that price instead of the last print of the two stocks. This can be significantly different. Trade your spread based on that, and you'll find you have much better entries and exits.
     
  362. TruthTeller,

    I'm sorry if these recent posts are coming across wrong.

    This thread's emphasis was based more on swingtrading pairs like FRE:FNM and the like. This clearly does not fit the prop model.

    I am the one who has suggested that I personally trade pairs on a shorter term. I'm just relating my experience, which hopefully translates to others on a longer term. I wouldn't advocate that anyone else trade this way. Furthermore, I hope it doesn't sound like I'm pushing an agenda, because I am not. It's my hope that discussing things like entry pricing and the like will help traders doing pairs on any timeframe.

    But hey, let's say that someone figures out a way to net 3 cents a trade after deducting 1 cent per share for commission. Trading four pairs 200:200 making 35 cents per day can easily be done in an Interactive Brokers account. That only amounts to $70 per day net to the trader. Make that 400:400 and it's $140 per day.

    Realistically, with a $25K account to satisfy pattern daytrader requirements, margin will give you $100K to work with. Conceivably, that trader could do 1000:1000 and average about $350 per day net (actually more because of IB's reduced pricing for larger orders). Not huge, but probably better than most traders are doing. There's no prop trading agenda there, as you probably don't want to get much bigger in some less liquid pairs. So, do this from an Interactive Brokers account.

    But, I agree, like scalping, this is a very busy way to make a living. And the costs are higher than many other trading methods. Each person should evaluate this, or any other approach for that matter, based on time, costs, risks, and trading style preference. It's clearly not for everyone, and clearly doesn't require a Series 7 or prop backing.
     
  363. I used to net about $500/day intraday scalping equity pairs until I went cold with it.

    nitro
     
  364. Well,

    I do think that firms do this on "purpose" and for good reason. You need to read any introductory text about markets. THE ABSOLUTE axiomatic principle from which all other trading stems is the principle of a RISKLESS ARBITRAGE. At the CORE of understanding all trading is some form of arbitrage, all the way up the food chain to the speculators. Arbitrage profits (heat death) by it's very nature is the culmination of many markets.

    Saying that a pro firm encourages this kind of trading is like saying that a young turtle after hatching has an urge to go into the sea.

    nitro
     
  365. Not my intent to slam you, Moderator. But I see and hear of a lot of prop firms pushing this as the "new thing" when in fact it serves their purposes of generating fees ofttimes more than generating income for traders. Sorry. A sore subject of mine.
     
  366. Quoted from: "Searching for Alpha," Ben Warwick - page 149.


    Andrew Lo and A. Craig Mackinlay put a unique spin on this issue in thier book "A Non-Random Walk Down Wall Street." When they began examining stock prices in 1985, they were shocked to find a substantial degree of auto-correlative behavior - evidence that previous price changes could have been used to forecast changes in the next period. Their findings were sufficiently overwhelming as to refute the random walk hypothesis, which states that asset prices changes are totally unpredictable.

    The most important insight from their work occured when they repeated the study 11 years later, using prices from 1986 to 1996. This newest data conformed more closely with the random walk model than the original sample period. Upon futher investigation, they learned that over the past decade several investment firms - most notably Morgan Stanley and D. E. Shaw - were engaged in a type of stock trading specifically designed to take advantage of the kinds of patterns uncovered in their earlier study. Known at the time as "pairs trading" - and now referred to as statistical arbitrage - these strategies faired quite well until recently, but are now regarded as a very competitive and thin-margin business because of the proliferation of hedge funds engaged in this type of market activity.
     
  367. Just sounds like sector trading to me....Cmon hasnt everyone done pairs trading in advertently. Watch the sectors and make the calls. Its still just risk management....

    Ok I am doing TRIPLETS NOW. My methodology will soon be published.:D

    Seems like someone is always coming up with the secret potion.
     
  368. You are missing the point. What's important is not how much the firm makes, but how much you (the trader) makes.

    The firms push this strategy, yes, maybe because it generates more in commissions, but that is irrelevant. The fact is, it is arb and arb works, if you have costs low enough and knowledge enough to exploit it.

    If you can take home twice as much in profits by paying your firm/broker six times as much in comms, wouldn't you still do it?

    Riskless is not quite true, however, anything truly riskless would have no profit potential. I have found that the good arb traders are usually also good straight up directional traders, so arb may not work for those who are crappy traders to begin with..
     
  369. Riskless Arbitrage DOES exist. However, you and I (well, at least I) do not have the capacity to do it. AFAIK, the person that comes closest to being able to execute a Riskless Arb on the entire ET forum is metooxx.

    Statistical Arbitrage does not equal Riskless Arbitrage, but it is a great education to understand many of the Riskless Arbitrage trades anyway and who and how they get executed - in fact, it is the basis of PREM/program trading.

    nitro
     
  370. Intraday statistical arbitrage aka pair trading..Not hard.

    Its done all the time. You dont need sophisticated model ...

    All you need is a sector leading and sector laggard.

    Pull up the futures (ex. BTK,ES) and trade. That's it. KISS

    It aint brain surgery. :)
     
  371. It is not rocket science or brain surgery, however, first you state that one should find a lagging sector and a leading sector - that is not statistical arbitrage, at least not in the classical sense, as there is no STATISTICAL CORRELATION. That is just buying strength/selling weakness, or sector trading.

    You then say to pull up the futures, and then give the example BTK which is an index?

    nitro :confused:
     
  372. Oops ... didnt proof read. Indices (DJ Titans) and futures (ES).

    I know quite a number of intraday pair trading setup. Using autocorrelation, cointegration, CAR, and NONE has outperform a simple sector laggard/leading combo.

    If you run the cointegration of a pair based on correlation, the number would not be as good if you ran one using simple sector LL.

    Thus there's no point trading based on intraday pairs... just use sector LL


     
  373. Ok.

    Now this is making sense. When I pair traded, I had gotten a "brilliant" idea to combine pair trading with sector strength and to play pairs to diverge (but not accross sectors.) I was never quite able to get it going, however, if I ever get a chance to go back to the style, I will pursue it further.

    I have never quite gotten the hang of the sector thing. To me, a stock that follows the spoos may or may not lift/fall with the spoos dependig on it's own technicality at the time and less so with it's sector. I am talking about the smallest of time frames btw. I have watched INTC and AMAT VERY carefully in this regard for example, or C, JPM and countless others.

    Don Bright and Silk are actually the people I credit with getting me thinking about this. RS7 I think also had a long thread about somthing similar. I have never quite pieced it together though...

    nitro
     
  374. I'm wondering why this thread seemed to die out in July.
    Most of the pair traders I know lost lots of money in the past few months.
     

  375. How is that possible. I thought pairs trading was a no brainer. Just read this thread and you'll see where there must be some mistake. Maybe they were trading pear futures by mistake.
     
  376. ROFL, hoho
     
  377. This thread has been resurrected more often than a cat with nine lives.
     
  378. Because whole idea brings hope for easier trading.
     
  379. Hope? Which system is that?
     
  380. Whole idea of pairs trading.
     
  381. Hardly.

    A property of many observed time series, call the series Y, appear to be non-stationary. For example, a simple form of non-stationarity arises when the first differences of a series are white noise. In this case we say the series is integrated of order one, I(1), and the differences: delta(Y) ~ I(0).

    An important step in the statistical analysis of integrated series was the realization that is also possible for a linear combination of I(1) variables to be I(0). The variables are then said to cointegrate.

    it is not that pairs trading does not "work anymore," it is that too many smart people are doing it now.

    nitro
     
  382. ...which is the case of all "not working anymore" strategies e.g. trading systems.
    Trading is a business like any other, once competition in particular strategy is too high, the margins do not cover expenses.
     
  383. I find that it does still "work" around the opening of NYSE stocks, but in general, I consider the strategy akin to smoking crack. The highs and lows are probably worse with stocks, and you may need more and better drugs to get cured from it than crack.

    At least, you can't go to jail for doing it.
     
  384. It's been a while sense I heard from anyone. You're right...pairs have been VERY hard in the last quarter. Unreal trending going on. Of course, much of pairs is fading price action, so for those who do that it's been tough. I've never seen anything like it really.

    Let's try to be more vocal.

    -- ss
     
  385. hi sslavin.
    Could you please elaborate on what you mean by unreal trending lately. I interpret it as "pair trading is harder these days since the trends don't revert to mean " and since pair traders generally fade the trends, the "unreal trending " would tend to hurt the strategy.

    Thanks.
     
  386. Where's Don Bright stand on all this. I believe pairs is one of the firms core strats.
     
  387. Guys,

    Well...I'm not sure where Don is on this situation. I really don't think Bright has ever said it's a core strategy. Maybe that was true for one quarter here or there, but I seldom hear anything from anyone from the Bright camp on this issue.

    Trading volume in general has been up for July compared to many months past. I know Bright vol was up too, but since they're a private company I don't think you'll ever get acurate data on this from them. Nor the numbers that pair trade.

    You're right on the assessment of how hard or "unreal" the trading's been. Everything is trending after the first hour more than it chops. And if it chops it's in 30c ranges and you usually can't get the other side on to make it easy to take this range all the time. Very frustrating. Ones like AIGHIG100, CATDE100, APCAPA120, MWDMER100 and others are tough to predict to say the least. Of course, you look back on 5-10 days worth of data and say "duh" but it's seldom that obvious in the seat. Not too long ago (last year) you could take these pairs for 30 trades and 30-50 cents all day long, now you're lucky to get 5 shots where you feel like you have an edge intraday.

    Now...from a longer time frame, this work might be easier. I'd definitely say stick to the trend, form a bias, add 1 extra leg from time to time, but be very careful out there. More pairs the better.

    I do think this next quarter will provide our best shot this year at more normal behavior as traders come back from vacation, the war is essentially behind us, and the economy is trying to pick up. We'll see.

    Are you experiencing the same thing?

    -- ss
     
  388. Man, how did your pairs strategy perform over the last month or two?

    thx
     
  389. NYSE
    boring sidestep.


    peace
     
  390. Fortunately, I trade big enough that my rebates made me some money. But I'm only making a few thousand a month after all is said and done. I'm up 4 out of 5 days, then some freak takes out the week's gains. Been doing that for months. It's a far cry from me making $30 and $40 grand a month over the last half of last year...litterally.

    Clearly, part of my issue is discipline. What worked last year doesn't work this year. Moves are slower, it's trendy as hell, carryover's are mixed, fills are worse, but still...discipline and preparation should smooth out the performance. I'm getting better, but it's still not what it was.

    I've got to isolate my playing to the pairs that I know are more range bound. My universe is about 28 pairs. I trade at least 15 of those a day. 5 of those I likely shouldn't. That's my next focus.

    Have a good one.

    -- ss
     
  391. ss
    28 pairs require some technical strategy i guess. right?


    peace
     
  392. Yeah... I guess. I have 10 monitors in front of me. 4 for pairs. 4 for one sided analysis, and two for order entry. 4 computers total. You don't need to have all this, I've just morphed into it over the years.

    My in and outs are all technical based. Looking for logical sup/resist, and scanning for gaps on my pairs -- visually mostly.

    It works for me, but not for others. I can usually have at least 10 pairs going at one time when the market's rockin. I usually have $1MM to $2MM on at one time (at least I used to -- haven't lately). I've tried to train others to do the same, but many can't multi-task that much. You gotta start small and work up, but you'd be amazed at how quickly to you can trade a lot of pairs at one time on an intraday basis.

    Got to do my prep work for the day...trade or die!

    -- ss
     
  393. Ha-ha-ha.
    If there is a working strategy (any one) - sure it is easy. The most difficult part is to follow the strategy.
     
  394. Ain't that the truth!

    The biggest issue for me lately is justifying risk for the reward, and frankly there are many hours in the day where the risks are just not worth it.

    I've heard someone say that on any one pair they'd be willing to risk 25% of their account. I heard this recently and it blew me away. This is the dumbest thinking I've ever heard.

    I really think there are some people who are doing well in this environment, but I'd have to say the large majority of them are risking A LOT to make a little. If you have to support a wife, babies, huge mortgage, big city expenses, etc. then risking a lot to make a little isn't an option.

    A trader friend called me today and said he spoke with Bob Bright about getting back into trading. His reply was that those that are making money are trading huge and trading for little gains. He said 10,000 share lots. I'd like to see how they're getting filled on that 10,000 shares!

    I also hear thru the grapevines that mutual funds are programming the fills of 50,000 share lots and whatnot. They're taking out the human doing the filling. It's subsequently helping the funds, but it means tape reading's going to get worse not better. That's all I do, so it's clear there has to be some adjustments on my part. It also means that gaps on NYSE stocks aren't going to happen much, and that ranges are likely to be reduced. So, the way I see it...we're going to have to adapt because I don't see anything changing to help the professional daytrader anytime soon.

    As far as the economy goes, Greenspan is setting up the market to chop for at least another 4 quarters. Unless there's some world event I don't see the uphoria for speculation coming back either.

    So, that's my 3 cents for tonight.

    -- ss
     
  395. Big orders are being sliced and diced everywhere, Nasdaq included. They'd rather print 100 shares 500 times than 50k once.
     
  396. That's very true. I can't believe lately how many fills it takes to get 500 and 1000 share orders done. Can you imagine a 50,000 share block. It could take the computer a while...and meanwhile...we're the ones getting sliced and diced.

    Has anyone else seen the programs that the funds use to place the orders. It'd be interesting if we could disect how they try to fill orders...then we could step in front of them as the little guy.

    You know, something I've used in the past is basket software. Maybe I need to dust it off and try to play the fund game. I play those as relative strength trades. Make a basket of 4-6 stocks in every sector. Chart the basket. In a quote page watch for %change after the first 15 minutes. Buy the strongest, sell short the weakest. Typically the trade lasts a hour. You're out of the basket when say a 20 period moving average on a 2 minute is broken. Taking both off independently.

    You could chart the difference between the strongest and weakest and then trade just the differential. Investor/RT software lets you do this quickly...if you're willing to spend sometime learning it.

    -- ss

    (maybe I too need some secret coded name! SS or sslavin's a bit boring, eh?):D
     
  397. not to forget that many institutions are trading automised in order to get VWAP for whatever period.


    peace
     
  398. I have been considering this a lot lately. I just finished closing a position for about $6,000 profit, but I was in it for weeks, and at one point I was down about $30,000. I don't think this is wise, so I am going to wait longer to put on positions and add to the position much less frequently. I really don't know if this will increase my profit much, but I believe it will decrease the size of the draw-downs. Most pairs eventually revert - at least to a degree, but this is the biggest problem with trading pairs I think.
     
  399. Vwap trading is the key to successfull trading . My win/loss ratio improved hugely after trading using VWAP ...


    iraj
     
  400. Is intra-day pairs trading worthwhile? Or is pairs more suited to swing and position trading?

    I'd like to try implementing intra-day pairs for listed stocks but I want to make sure the potential profit offsets commissions. Doesn't pairs trading rely on efficient market theory, which was most notably debunked by the spectacular collapse of Long-Term Capital? Since pairs trading is inherently contra-trading (going against the trend by shorting an "overvalued" stock and buying the "undervalued" stock for each pair) it seems that the chances of both the long and short position going against you are high. The Pairs Trader assumes that both stocks will revert back to their mean based on number crunching historical data. I'm a newbie to pairs trading but this is my limited understanding.

    Would like your thoughts on this.
     
  401. I don't think pair trading neccesarily involves playing mean reversion, could be a trend play where one trades breakouts from "bands" since a lot of people are trapped.
     
  402. ltcm did no equity pairstrading as far as i know.
     
  403. ------------------------------------------------------

    I am exclusively tading pairs intraday and I think it is a great strategy in the current markets which lack the momentum we had before.

    There is a lot of ways to play pairs. You can trade the correlations using standard deviations and reverse to the mean. You can trade the pair like a stock applying technical analysis to it. That works great. Remember all currency trading is basically pairs trading. You can trade the trends or breakouts. Something I am developing a model for now is news pair trading. Play the news in a stock and hedge with a correlated stock without news. That has worked really well so far.

    Thers is as many ways to play pairs as there is traders. Being creative helps...

    Hope it helps.
     
  404. LTCM did not fail because their assumptions/model was flawed.
    They had on too much leverage to be able to withstand the positions going way too far in the wrong direction. The correlations eventually revert back, as the model says that it will, but before that happens, it goes so far the other way you get ruined in the process.

    I think this is a good lesson for pairs traders. It doesn't matter about your assumptions being right. You can be "right", and still lose everything. You cannot get in too deep with leverage. BTW, basing your risk entirely on historical Std devs is dangerous as well (the infamous "fat tail problem")

    Better to figure your risk based on a) time (ie. the longer you hold, the more risk you have) and b) a preset stop where you will exit the positions immediately. Historical factors too, but downplayed somewhat. (No one believes that -Past performance are no indication of future results-, anymore! :) )
     
  405. esc

    i sign all of that


    peace
     
  406. I back tested S&P500 stocks for the last 5 years and found out that some pairs tend to diverse and some to reverse to mean. And for any given pair with high enough correlation that already diverse enough from mean you can say if it will tend to diverse or reverse. Win/loose ratio may differ and some months you will be negative but usually it’s from 3/1 to 4/1.
    Here are some pairs I’m currently in.
    I use www.pairstrading.com for charting. Are they the only ones who have free charts for pairs? Though I like RSI and MACD, I also use ADX indicator in my analysis but have to go to QCharts for it.
     
  407.  
  408. No, but their whole bond arbitrage strategy was based on deviations from the mean between historically related bond-instruments. Like equity pairs, they sought to take advantage of large deviations by shorting overvalued instruments and buying undervalued ones.
     
  409. is ANYONE making any serious money trading pairs in this current market? seems like the edge has diminished significantly with this type of trading.
     
  410. golden
    right, but if you throw all relative value trading into one basket you are way to general to make trading-relevant statements.


    peace
     
  411. To be more specific, is relative value intra-day pairs trading (comprised of being long one stock and short the other in a pair) a consistently profitable technique for listed equities? And is there really a "hedge" in this type of transaction? Given that there are many pairs traders who are looking at the same pairs, it would seem that good opportunities are becoming more scarce and the potential profits have narrowed too much to cover commissions. I'm not trying to dismiss this technique since I'm always looking for new opportunities, but I would like feedback from traders who are consistently successful with pairs trading (if they exist).
     
  412. you won't find many this year i guess. our strategy on daily pairs is up only 4% and our intraday strategy is just tested, not traded so far.

    we found the pairs game has become more sophisticated - well no big surprise. that's the business.


    peace
     
  413. Well ... all processes revert to some mean after enough time. Not everybody can wait for 10^30 years to observe "average" behaviour.

    LTCM failed IMHO because they were arrogant people who started believing their own press. They clearly failed in what should be the very first item any firm in this busuiness invests in: a compliance system that does not allow anyone to violate the risk model and reports to senior management each and every day - or even hourly or more if the risk parameters go outside of the comfort zone - with the total risk exposure of the firm.
     
  414. Cal

    i agree. LTCM was human failure not so much a modelling weakness. nevertheless most people who quote on LTCM do that without proper understanding. the leverage is one issue of that kind. yes, they used way too much leverage, but still you can use more leverage in curve trades on us governments than you should use on eminies. so when people use the LTCM case in order to tell that fixed income arbitrage as such uses too much leverage - that is hard to swallow. nobody mentioned that here, but i heard that on many discussions since 1998 and i sometimes hear it before it is said.

    the portfolio as such was sound. it was too big, but it was sound and suffered a four sigma event or even more. LTCM should have survived this with a 20% draw down if they had not been the kind of personalities they had been and pushed the engine too far.

    we looked at LTCM in early 1998 beause it was a very popular fund at that time but we decided we could not invest, because they would not tell anything about what they were doing. thus one could not check if they had any kind of real risk process.


    peace
     
  415. LTCM failed because they began to make more directional bets, abandoning the cautious hedging strategy that had been its trademark. With spreads shrinking in bond arbitrage due to other copycat firms, LTCM was desperate to park its capital someplace else. The fund started dabbling in equity derivatives and other fields where it didn't have expertise or an edge. Combined with huge leverage and the hubris of its partners, LTCM was poised for disaster. The reason LTCM had phenomenal growth during its first few years was because bond arbitrage was still a fairly new field, and the partners in the fund were the founding fathers in this arena while at Salomon.
     
  416. They were, and if I recall correctly from Lowenstien's book, it was the only part of the fund that was still profitable when they finally broke it up. The book said they were big in RD-SC, among others.

     
  417. Don't forget Merton and Scholes also started telling their investors the "actual" odds of LTCM losing 5%, 10%, etc... as if they were rolling a pair of dice. If that doesn't send up a four-alarm signal, I don't know what does.

    Lescor's right, according to Lowenstein, LTCM did pair trade. They had about 10x what Goldman had in RD/SC. They got involved in M&A deals too.

    :cool:
     
  418. time to get out of the markets. when trading i get so chopped up that i feel like a cat in a chinese neighborhood.
    the specialists are all scum just look at FNM FRE. the fnm guy will spread the stock 30 cents wide keeps on dropping it and then none of your bids get filled when trying to buy and then when you try to buy the offer stock goes up 30-50 cents
    protest by not trading anf giving these scumbags your money
     
  419. With an attitude and remarks like that, you should expect to get chopped up anywhere.
     
  420. Well , Statistically all PRICES SHOULD REVERT TO MEAN but this is a long term strategy ( as all statistical strategies work most effective in long term ) not suitable for intra day traders.. In fact this assumptions is dangerous..
    Lets say STOCK X is trading @ its mean . i.e. Intra day VWAP ... Now DOW rallies pulling stock X up more than 2 SD away from its VWAP .. So STOCKX is a good candidate for shorting as statistically there is 95 % chance of STOCK X reversion to it's intra day VWAP
    The problem is if you take a short position you are going against the general market trend TRYING TO CALL THE TOP (TOO RISKY ) .. Alternatively you can have a strategy to Long those stocks that deviate from their mean fast and furious.. The logic behind this strategy is :
    IF STOCK IS STRONG ENOUGH TO DEVIATE 2SD FROM ITS VWAP FAST ENOUGH THEN I WILL BACK THE STRENGTH ON THE FIRST PULL BACK..

    I have been back testing both strategies i.e shorting or Longing those stocks which deviate 2sd from their mean and this is my conclusion :--

    Most statistical approaches donot work intra _day and a trader is far better off to back the strength ( Go long when deviation exceeds 2sd) than short and go against the trend.

    Hope this helps
     

  421. I also back tested both strategies (not intraday though).
    Here is my conclusion: Pairs obviously do not always reverse to mean.
    I also do not consider 2SD to be enough to decide if one stock is overbought (oversold) compared to another. On my opinion, RSI is much better in this case. So, for the pair to be a candidate, it should have AT LEAST 2SD plus RSI.
    Even these pairs do reverse approximately 3 or 4 times out of 5.
    Those times when they do not reverse, eventually, of course, they will, but at that time MEAN would be obviously slightly different.
    More interesting question – is it possible to predict in advance, will the pair reverse or diverse if it has 2SD and RSI is at extreme levels? My back testing shows that it is possible. So, I trade both strategies. Even more interesting that winners/losers ratio about the same in both cases (3-4 to 1). And I also had 2 months out of 8 this year when result was negative
     
  422. quote from m&m&m:
    More interesting question – is it possible to predict in advance, will the pair reverse or diverse if it has 2SD and RSI is at extreme levels? My back testing shows that it is possible. So, I trade both strategies. Even more interesting that winners/losers ratio about the same in both cases (3-4 to 1). And I also had 2 months out of 8 this year when result was negative
    ------------------------------------------------------------

    m&m&m,
    How many pairs do you follow/trade? Care to provide any insight on what characteristics you look for to determine prediction of divergence vs. reversion for a particular pair? Your win/loss ratio is excellent. :cool:
     
  423. Hi,
    For back testing, I used only S&P500 stocks for 1998-2002. Actually, to get correlation for January 1998 I used data since 1996.
    I do not trade many pairs, but follow all the combinations available for Russell 1000, S&P500 and Nasdaq100, though I like stocks with a volume. I hope I’m close to finish a system that will trade automatically. For now I trade just 4 –5 pairs at a time. I know, some guys were laughing here about my volume, and specially for pairs trading you do need size (I for example give up about 4 percent per position just on commissions), but I think you can risk some serious money when all experiments are done.
    Regarding divergence vs. reversion – well, I like trends. If a pair is in a trend – I follow it (meaning I bet pair will diverse even more). If it’s not in a trend (yet) – I bet it will reverse. My experience shows that some pairs (appx. 1 in 3 or 4) that I like (when I say “I like” I mean – the program likes, I may not like it at all if I look on a chart) – some of them instead of reversing will diverse even more, and at some point they will fall into trend. At that point we (me and the program) also like them and I follow the trend. There are some indicators to decide if a pair is in a trend or not – ADX for one. Unfortunately, I was not able yet to back test the system with ADX – I need intra-day data for pairs to do that. In real life I go to Qcharts to get ADX for the pairs, since this is the only one charting program I know that have ADX for pairs – www.pairtrading.com do not have ADX, though they recently implemented MACD (which could be used for entry-exit points). During back testing I used another function (some function of standard deviation) that also can show if a pair is in a trend.
    Here is one more of my current positions (I had it the previous time I uploaded few positions, but back then it did not move yet). Here you can see why I liked it and why I entered it at that time (MACD crossed 9days EMA)
     
  424. I've just noticed this thread is one year old today :D
     
  425. All thanks to thread-starter Neutral Al !

    And it has proved to be one of the best threads on ET! :cool:




    By the way, just where IS Neutral Al....? :confused:
     
  426. Cont'd:

    Through its several resurrections,
    traders continue to step up and
    make solid pairs strategy modeling
    contributions to carry the thread forward.

    The ever increasing number of "views"
    speaks for itself.


    Expect Neutral Al and more than a few
    others to re-emerge as even better traders! :cool:
     
  427. Hi seisan,
    Hi everybody,

    Thank you for your kind words.
    It is nice to realize that together we have created something that is of interest to so many people. The numbers of views and responses indicate that market neutral trading, pairs trading specifically, is not being overlooked by the trading community.
    Reading through the posts, I have noticed several “theme” questions that come up time and time again. Here are my thoughts on the subject:
    1.Why trade pairs instead of a single security/index? By its nature, pairs trading is not a “money making” instrument. It is a risk management tool. It was created, and is utilized the most by professionals who manage large portfolios and are responsible for other people’s money. If your goal is to preserve your principle and enjoy a relatively small but steady increase of your portfolio, then pairs trading is for you. If you are in the market for a quick and large profit, and able to stomach significant volatility in your account, then pairs trading is not your cup of tea. Personally, after doing it exclusively for a year, I find pairs trading to be the least stressful and the most rewarding trading strategy I have ever used. It gives me acceptable returns ( +38% for the year), and most importantly, life style that I want. I hate to spend all my days in front of the computer.
    2.Is pairs trading dead? It is very much alive with me, but I can see the point. Intraday scalping spreads lost much of its appeal after the decimalization took place. Increase in program trading with most of the volume attributed to a few major institutions makes it even more difficult. However, it sounds like Mr. Slavin at www.pairtrader.com is doing well. It takes a special talent to be able to do what he does. Positional pairs trading mean reversion has also slowed down a bit. I attribute it to this incredible complacency in the markets that we experience as of late.
    3.The LTCM story. Yes, they did pairs trading. However, that is not what killed them. They died of the same plague that kills any Market participant i.e. lack of discipline in adhering to the basic trading principles. You have to have STOP LOSSES!
    4.A lot of people ask me in the private e-mail whether the www.pairstrading.com ‘s manual is any good. I bought it. I used it to create my model. So far it worked fine for me. However, I can see a sophisticated stat arb trader being disappointed. There is nothing in the manual about cointegration, time series analysis, fuzzy logic or any other advanced stat arb tools.

    On the final note, let me know if you are interested to see my trading ideas. I’ll be happy to post them for your review.

    Good luck to all. Keep it going.

    Al
     
  428. Al, I would be more than happy to see your ideas. I've been trading spreads a while agoo and start to pick it up again.
    Any recent trading ideas / trading history would be a great review.

    Thanks in advance,
     
  429. Neutral_AI FYI...

    This came to me in an email from pairtrader.com...
    ______________________________________________
    Slavin Bids Farewell - I’ll Be Back by Steve Slavin

    As many of you know, I haven’ t been around much lately. August is always a slow month, but I’ ve taken this time to
    reorganize and consider new challenges. I’ ve had a horrible year consistently making money – it’ s night and day from last
    year. Almost every month I’ m on the positive side of the ledger, but with a wife, kids, a big mortgage, and investors to
    answer to, my wife and I have been asking “is it enough?” Unfortunately, it isn’ t.
    That’ s the reality of this work. There are times when the market is so generous and times when it shuts down. In the later,
    taking pennies from the market is darn near impossible without taking on extra risk. That’ s the key there – can you take
    business risk that’ s controlled and calculated. It makes no sense to risk $1.00 to make .20 cents in a dead market. In a fast
    and more volatile market you can get away with it more often since you have 20 times more opportunity to make it back,
    but we all know this hasn’ t been the kind of market we’ ve seen lately.
    By no means am I making excuses for my performance. Clearly I didn’ t adapt quick enough by lengthening my time
    horizons, but I never felt comfortable with the risk in doing so. It blows me away that I could go from making $20-40k a
    month to so little, but my story is not unique.
    So, what’ s next for me? I’ m exploring a job offer at a great institutional fixed income desk in Columbus, but I’ m not sure I
    can maintain two households in the short run. I’ ve also been offered a job running brokerage operations at a small bank.
    Do I see myself coming back to trading someday? Are you kidding me? Absolutely! I just need a more generous market
    and a bit more cash to buy me more time in the slower months. I certainly know I can make a ton with the right conditions,
    but for now I’ m excited about a new challenge.
    What’ s next for PairCo? To go into a new position where I use my Series 7, compliance departments are telling me I need
    to dissolve my interests in PairCo and PairCo Capital Holdings. Fortunately, Darren Clifford in Vancouver has really
    stepped up to the plate and made a further commitment into the concepts of PairCo and PCH by buying me out. Thanks
    Darren! Rob’ s assured me the door is wide open for my return someday, but for now, I’ m stepping down and letting them
    run the show. I expect nothing but great things from them! PairCo has been a great experience and shall always be in debt
    to Rob for his great leadership and friendship.
    Best of luck to everyone! Remember to keep your risk in check, and be prepared to keep changing your methods – over
    and over.
    Regards,
     
  430. TM Trader - now see what you started?! :cool:

    Neutral Al - thanks for the appearance!
    Everybody's been re-reading your post
    and
    waiting for you to show us
    some more trading ideas... :cool:

    vonk :cool: - thanks for passing on Slavin's particulars
    with his bottom line trading reminders to
    absolutely control risk management
    and
    constantly tweak one's strategy methods to the market.

    All others - keep up the posts - it's September, already...! :( :)

    djclif - yeah, you've got your own Conference Room
    and are practically running the company now... :cool:
    ...but, send us a post at least once a month or so!
    Everybody's just waitin' to kick it around a little.... :D
    ...'specially since you're now a bona fide ET'er...! :cool:

     
  431. Seisan,

    You asked for a post once a month, I will try to make it more frequent then that. My biggest problem is that I am too busy trading, working on our website, mentoring new traders, and writing automated trading strategies. I have a new personal assistant I will make sure she reminds me to be more active within these discussions.

    For those of you who are not familiar with my background, I am the CFO for pairtrader.com, a professional prop trader (through Bright Trading), I sponsor under capitalized prop traders, and write automated trading strategies. Enough things to keep me busy. Reading through some of the posts here it appears like the group has had some great creative discussions, I look forward to contributing and learning more with everyone here.

    To start, I should probably tell everyone the style of pairtrading I am using to make money in this market. I trade multiple pair strategies which include relative strength pair baskets, swing trading pairs, pair investing, trading tight 'arb' spreads, pair opens, and pair crutch trading. The list is broad, but I am using each of these strategies everyday while I trade. If anyone has any questions on these I would be happy to fill you in.

    Looking forward to some great discussions,

    Cheers,

    Darren
     
  432. We just knew you're doing a supreme balancing act with all that
    you've got going on - that's why we egged you on all the more... :D

    A trading pro and a great sport! :cool:
    Just a typical patience and discipline kinda' guy, eh....?

    Thanks for comin' aboard Darren.
     
  433. So, who's first up with a question or contribution sometime today.

    Too many sharp heads on this thread to leave the above strategies just layin' there on the table... :)
     
  434. Can anyone recommend books on the subject?
     
  435. OK, lets start basic then,

    You told in a prior post you do all kind of trading (spreads)

    I'm a retail trader , trading spreads since 2 years now, and I came to the conclusion that my returns are a lot better when only using EOD data and trading only on a daily timeframe (Might take some intraday stops / profits)
    So I created my own set of pairs that I follow on a daily base and so far this has been going great.
    However, I'm always looking to improve my results and for me I think the way is to optimize my entries / exits (not changing my timeframe)
    - What do you use for your entry /exit signals ?
    - How do you get your enties / exits, is there a way to optimize ?
     
  436. ... and tell us your pairs too :D
     
  437. I'm not asking for pairs, I'm asking for something else ,

    - How are your entries / exits calculated or formed
    - Do you optimize this ?
     
  438. Thanks for the question H2O.

    My entries and exits are determined using technical analysis, supplemented by fundamental analysis. The common plays I make are based on support numbers, resistance numbers, bollinger bands, and standard deviation plays.

    A couple things to remember: a lot of traders are gripping about the lack of volatility in the market (the VIX being around 20 and may go lower), what we trade is predictability, not volatility. You can have all the movement in the world, but if it is un predictable you can't make any money trading it. Give me less volatility and more predictability any day. With more predictability I just increase my size to make more money.

    How do we get more predictability? Look for compounding probabilities. This is where I use fundamental analysis to support my technical trading. I am not willing to hold positions for extended periods of time without having the fundamentals in my favor.

    As far as what pairs I trade, they change from week to week. I look at around 100 pairs every weekend to find the top 10-15 I want to trade. Some common pairs I trade include SBC BLS, MWD MER, FD TGT, VIA.B CCU. You can get a whole list if you sign up for our website www.pairtrader.com.

    Hope this helps,

    Cheers,

    Darren
     
  439. It was asked earlier if there are any books that can be recommended, if you go to www.pairtrader.com you can order the Art of the Arb Trading Manual. This is a comprehensive book from real traders who are actively trading pairs on a daily basis. The book covers everything from business plans, how to handle blowups, technical analysis, handling problem pairs, pair weighting, M&A pair trading, and generating pairs. It is the book for the course that Rob Friesen teaches. I used this book and this course to lead me into the trader that I am today.

    Even better than the book is the mentorship. I have been blessed trade beside some top traders who are willing to share things with me, and it made all the difference. I can stress highly enough the benefits that you get from talking and seeing someone actively trade the strategies you are interested in.

    Cheers,

    Darren
     
  440. pairs trading is losing its effectiveness as more and more people are doing it, that is why people like darren are mass marketing it through websites and books. don't get me wrong, you CAN make money pairtrading, but if it was such a great strategy, why are there people trying to get rich off marketing it instead of trading it??
     
  441. For me it's I can hold positions overnite with LESS risk than just holding a position.

    Before you comment, I know there's up/downgrades, 9-11's etc, but overall there's less risk in a combined long/short position than there is in a single long or short.
     
  442. Let say you have a great strategy and you making very good money with it. Would not you want to sell it for a few millions?
     
  443. Hi H2O,

    I opened a new position yesterday at 1400: Long PPH / Short XLK.
    Entry ratio 3.9552. Exit 4.3. Stop loss 3.89.
    Will keep you posted on progress.

    Regards,
    Al
     
  444. Here is the chart:)
     
  445. This is an excerpt from our recent pairtrader monthly news letter.

    //
    Changing of the guard.

    After talking to experienced traders over the last couple of months, I hear a lot of traders saying ‘remember the good old days’ or ‘if the market was only the way it used to be’. Personally, I am already tired of hearing this. I fully expect that two years from now the current market will be called the good old days. Learn to change, learn to adapt, this is the only way you will be able to succeed in the long run.

    Of all the things that traders lament it is the disappearance of easy money. Trades with low risk, high rewards, and instant gratification no longer seem to exist. Simply put, stocks are no longer giving the prints they used to. These prints allowed mediocre traders to achieve good profits in previous years. Pair traders, in particular, had a great advantage in this type of a market. They where in a position to capture a lot more of these prints since they where in so many trades to begin with.
    With the easy money being taken out of the market what is going to be the thing that replaces it? I think the answer lies in one of two camps: either trade with patience or trade spreads with little risk and little rewards.

    When I said earlier that I expect the current market to be the good old days two years from now, here is why: currently, my business plan with spreads in general is as simple as being patient in executing my first leg of capital, be patient in taking profits, competent management of positions, and making lots of money. Two years from now I will be begging for the days when you made a simple business plan, stuck to your guns and made money.
    //

    You can still make money in this market, and you can still make money trading pairs. The problem is you can't make easy money right now. You can make a million a year still, but you need to take more risk to do it. The people who will succeed in this market are the people who are able to understand risk reward relationships, trade probabilities, and stick to systems.

    Cheers,

    Darren
     
  446. Darren,

    Several questions for you, if you don't mind.

    1) How long have you personally been trading pairs?

    2) What time frame do you use for your trades? (i.e. typical trade held for hours, days or weeks)

    3) What annualized percentage return do you expect from pair trading (max 2x leverage in your positions)?

    Thanks,
    -Eric
     
  447. Darren,

    I would also like to ask - intrading pairs, do you add on to a losing pair position or do you adhere to a stop loss? If you adhere to a stop loss system, how far should a pair go against you before you exit?
    Also, is your chatroom still in operation?
    Thanks.
     
  448. It was asked how long I have been trading, this is month 16. Not too long, but the joy is that I am trading profitably in this market where others are struggling.

    As far as the time frame of my trading, I trade everywhere from 20 second trades, to holding for two months. Generally I am not trading longer than this. I am looking to increase my time frame with some of my trades as I have a bigger capital base to work with now. The market seems to be favoring patient people.

    As far as an annualized percentage return with two times leverage, I would expect returns between 25 and 30 percent.

    With intraday stop losses, I do use them but in a different way then most people. I will never sell the bottom unless I really see something in the tape to make me want to get out of the position. I will wait until I get retracement to a two minute moving average. Lighten up on some of my position and then revaluate. If it continues in my favor I will be patient to take off the remaining position. If it goes against me I now I have known resistance point. Remember: the reason why you get into a trade justifies where you get out of a trade. Whether for profits or for losses.

    Cheers,

    Darren
     
  449. A lot of traders (now ex-traders) I had traded with blew out their accounts by pair trading the "traditional" way - meaning keep adding on to a losing position until the pair either comes in for you or ends up hurting you in a very painful way. I was toying with the idea of observing a predetermined stop loss if a pair goes against me instead of adding more units to the position indefinitely. I plan to only hold 1 unit for each pair, close out most positions at the end of the day, and close any position once my stop loss is hit.
    Has anyone tried this method of pair trading? I like the advantage of being hedged but cannot stomach the idea of adding to losers.
    Any ideas, anyone?
    Thanks for any input.
     
  450. Reg, I know a lot of traders who have blown out their accounts over the last year trading pairs in the 'traditional way' as well. Most of these traders did a very poor job of understanding risk reward relationships, and 'business as usual' environments.

    Unless the probabilities are highly skewed I never risk $1 to gain $0.2. I am amazed at how often spreads develop common consolidation areas, or have tendencies to trade like individual stocks with no trend. I look at my historical chart and figure out where the next area of support or resistance is and calculate my risk to there. Some times this is dollars away from my current level, in which case I should not be adding to my position every $0.3. Only add when you would take the trade from scratch. You need to be able to justify why you are in the trade and why, based on probabilities you are adding to a position.

    Most trader add to a loosing position because they do not like the red p/l they see on the screen. Hide your profits and losses while you trade and only trade probabilities. For me the only time my profits and losses enter into my trading decision is when I am down half of my account, at which point I cut everything to make sure I can trade tomorrow.

    The 'business as usual' environment comes from spreads trading within their historical ranges. If I see that a spread has traded in a $30 range over the last three years, it should not seem that odd for the spread to trend $10 in a month and revisit one of its historical highs or lows. My business plan needs to alot for this, and I design my stop losses accordingly.

    The rule for adding with me is actually simple; first: don't take the trade unless it stands on its own. Second, stop out of a spread when you are not willing to take the loss all of the way down to the next support or resistance level. For the most part I have been having great success in this market by being patient, being willing to accept some pain and extending my exit points to justify the risk.

    Hope this helps,

    Darren
     
  451. Thanks Darren. Great explanation.
    Is the pairtrader chatroom still in operation?
     
  452. Yes, the chat room is still working. There is a limited number of seats available. You can access it by going to www.pairtrader.com and clicking the portal link on the right hand side of the page. On the next page click the join chat room button on the left side of the page. You have to download some software, but this happens automatically. Feel free to come and join us.

    Darren
     
  453. is pair trading considered statistical arbitrage, quantitative arbitrage or what?? i think its stat arb, correct me if i'm wrong.....and what would quantitative arbitrage be? is doing arbitrage on mergers quant arb??
     
  454. Bigbeech,

    Statistical Arbitrage is a form of Quantitative Arbitrage. Quantitative analysis is any form of analysis that uses numerical analysis, such as statistics. The opposite to this would be qualitative analysis.

    Cheers,

    Darren Clifford
     
  455. Did you know pairs trading is indirectly related to Jeff Bezos?

    The Shaw referred to below is founder of D.E. Shaw, a massive quantitative arbitrage hedge fund. Jeff Bezos helped get it off the ground before he embarked on the Amazon.com project.

    The more I learn about who is running with the world's money, the more intimidating trading becomes. This your competition with pairs strategies....

    [​IMG]


    In 1986, Shaw was hired into the Analytical Proprietary Trading group at Morgan Stanley. This was a top-secret, "black-box" operation that worked behind a guarded door on the nineteenth floor of the Exxon building in midtown Manhattan. They were trying to exploit anomalies in stock market prices. Their job--aided by Shaw's computational expertise--was to isolate recurring patterns in financial data and use them to beat the market. The group was run by Nunzio Tartaglia, a former Jesuit seminarian with a Ph.D. in astrophysics, and Bill Cook, who was head of information technology at Morgan Stanley. It was Cook who later founded Tech Partners and tried to lure Norman Packard to Wall Street in 1991.

    The secret being exploited at Morgan Stanley was something called pairs trading. Pairs trading traces its lineage back to Jesse Livermore, the famous speculator whose life story is recounted in Reminiscences of a Stock Operator, first published in 1923. Pairs trading is based on the idea that prices of related stocks should be correlated. Ford and General Motors will tend to fluctuate in price around the same news events. But what if an unusual gap develops, where Ford lags in price and GM pushes ahead of their normal relationship?

    In this case, a stock speculator might go short, or sell, GM and go long, or buy, Ford. The market can crash. The market can soar. But if the gap between the two stocks remains unchanged, our speculator will neither win nor lose money. Such strategies are called market neutral. The bet is being placed not on which direction the stock market will move, but on company-specific or sector-specific correlations. If the gap between GM and Ford narrows, with GM falling and Ford rising in price, as predicted, our speculator will make money. How much money? One can only guess, but if Morgan Stanley was hiring astrophysicists and tracking additional 0s onto the salaries of college professors, the strategy must have been rewarding.

    A key trick to statistical arbitrage, or stat arb, as Wall Street refers to pairs trading and other forms of algorithmic hedging, is to get one's strategy truly market neutral, not only to fluctuations in the stock market, but also to fluctuations in interest rates, foreign exchange movements, and global economic risks that can come at you faster than a shark going for red meat. Wall Street is filled with former employees of "market-neutral" h edge funds that went belly up because they weren't neutral enough.

    Shaw worked for a year and a half at Morgan Stanley before striking out on his own. He wanted to expand pairs trading to include multiple classes of securities--bonds versus stocks versus options--and multiple markets hedged against one another in a computer network designed to be global market neutral. Shaw was going to stat arb the world; he was going to hedge the universe. Today, his computers are so powerful that Shaw, along with locating market inefficiencies, also detects rigged markets and other forms of cheating.


    The Predictors pp. 106-107
     
  456. Bigbeech, mergers arb is normally known as risk arb, I think quant arb and stat arb are pretty interchangeable terms.

    Mac
     
  457. Just read the entire thread, pretty interesting and only have a few things to add.

    First off, I can't imagine why there was talk about averaging down and whether stops should even be used. Trading the spread is still trading, the rules still apply. It's a losing game, if the position you are in is incorrect, get out, dont hope! I believe Neutral was trading with stops from almost the beginning and was successful for over a year (I don't know whether he is still successful, lets hope so) while others were struggling and complaining about lack of volatility. I also don't understand why the reluctance to trade trends, if a spread seems to be fluctuating around a mean, then look to catch a move towards the mean, if a spread is trending, trade with the trend. Another thing that I wish to have some discussion on is using options. Both on using options instead of buying/shorting underlying in position trades on pairs and maybe even trading spreads between options for different strike/months/underlying? The idea just came to me while reading the thread and was wondering if any of those more experienced with options or spread trading could tell me if there was any use to these strategies.
     
  458. Has anyone tried pairs trading using single stock futures?
     
  459. last time i checked the SSF there was no real trading volumes. some stocks have volume, but it is all concentrated in a single block trade per day


    OLiver
     
  460. What is happening with the pair traders this year. I haven't seen any additional info from anyone.

    I doubt the edge has fully disappeared, but I was just looking for an update in the forum....Neutral Al?

    Nitro? have you come back into the fold?
    DaveN....?

    Anyone......bueller?
     
  461. Pairs trading is dead...Just look at spread charts. There is no mean reversion, they start moving in one direction and continue to do so day after day after day. This makes it nearly impossible to trade pairs and make great money. I think automation and low volatility have taken much of the edge out of trading pairs.


    The good news, if you are flexible and change your strategies to adapt to current market conditions you can make great money. I am making great money by taking advantage of the lower volatility and taking position trades, either long or short.

    Mike
     
  462. Let me see if I can give some clarification to the words of the imortal mschey: pairtrading as we knew it is dead.

    People are having a hard time trading pairs intraday with much success. Whether it is trading off Bollinger Bands, average daily ranges, or pair support resistance numbers, few strategies seem to be providing a big enough edge to overcome commissions on their own.

    By changing strategies to use compounding probabilities, and be very patient with capital allocation, we are finding continued success in the pair field. I have been very successful trading pairs in this new manor, and I have been successful in mentoring others to trade pairs in this fashion as well.

    Cheers,

    djclif
     
  463. I think one problem here is that people often define for themselves the meaning of something.. pairs trading clearly means different things to different folks... its impossible to declare anything definitive this being the case really... some regard it as 'statistical arbitrage' or something... whatever the hell that is.. dont tell me please! sounds way to complicated for me.. and anyway there is little to compare between intra day pairs trading and swing based pairs trading.. so many other examples also..

    For myself, I am now trading a portfolio of sector based pairs.. chosen by basic ta for strong longs and weak shorts... swinging as long as they take... and have been doing better this year since I started than any of my trading previously, which was mostly daytrading..... so its too varied in my opinion to draw any meaningfull conclusions... altho it is clearly true that it has little interest to people here.. who seem mostly devoted to swift trading and rapid profits.. no idea what happened to those who were posting earlier, hopefully they have thriven and prospered and are too content and happy to bother posting here.. :)
     

  464. up 1,9% for first quarter. not much downside, not much upside. boring, boring.
     
  465. In PairsTrading you just trade pairs of something that used to behave similarly and at the moment for some reason there is a gap in them. Let’s concentrate just on stocks (though you also can trade options, futures, currencies… whatever).
    Even just with the stocks there could be numerous if not infinity of possible strategies – when you open the position? when you close it? and what you actually do when there is a gap? You may guess that this particular gap will shrink, but another one will get bigger, so you can play trend in that case.
    So, if some particular sub-strategy does not work, one can’t say the whole PT does not work. No – just THIS strategy does not work. And maybe even never worked – have you tested it before going into it?
    Here is a latest article on CBOT about PT.
    http://www.cbot.com/cbot/pub/cont_detail/0,3206,1070+18827,00.html
     
  466. I know lots of people who backtest and develop systems, and not one of them makes any money trading. On the other hand, my actual trades demonstrate that I having many winning strategies, that produce consistent, daily, results. And that is all that really matters!


    Mike
     
  467. So, how is it dead then? :confused:
    Result is all that matters, right?
     
  468. Hi all,

    I saw the thread active again. Here is my update...
    I finished the year with 24% gain doing pairs trading exclusively. This year I am up 7.4% so far with the Sharpe ratio of 5.
    As far as death of pairs trading, I think it is more alive than ever. When I started a couple of years ago, I was able to find only a couple of articles on the subject. Today there are dozens of articles and books on trading pairs. Even the main stream Industry publication such as Futures Magazine features an article on pairs trading this month. http://www.futuresmag.com/library/contents.html
    Get a copy, educate yourselves…
    What looks dead, however, is the action of pairs trading. It is slow and boring. Just the way I like it.

    Good luck to all,
    Al
     
  469. Why are you encouraging others to arb away your edge?
     
  470. What are some of the pairs that you trade? its all equities right? Any pairs you would like to share?
     

  471. that is very, very good figures. you are trading systematic? intraday?

    good trading
     
  472. Hello,


    24% of what amount of capital, if I may ask? thanks



     
  473. I trade equities and some options pairs spreads. Some of the recently closed positions are ALTR/XLNX from R=0.539 to 0.60(gain of 11%) and MSFT/IBM from R=0.269 to 0.3 (gain of 10%). I presently hold 18 pairs. Ex: CTX/DHI, CECO/APOL.

    Al
     
  474. I work with several traders who trade pairs, mainly oils, gold, and banks. Oil traders have been much much more active this year and have done rather well....basing pairs on 5 day has been most successful. Golds have been a little more difficult this year and again shorter term indicators on the five day have been more profitable than running longer time period indicators. As for banks its been a banner year and one that will be tough to duplicate in the near future. Any of you guys trade oil service pairs or gold stocks? If so I would like to talk.


    Chad
     
  475. thats how i trade all my pairs
     
  476. My group currently mentors 23 traders, with all but 1 profitable this month. Our traders are focusing on two separate strategies: scalping the mergers, or trading a portfolio of pairs. The pair portfolio has its foundation in technical analysis, supplemented by fundamental analysis. We are using a swing trading model that requires the ability to leverage larger amounts of capital (this is why we trade with Bright Trading), to create a larger portion of profitable trades.

    If you want more specifics about our trading strategy visit our website at www.pairtrader.com, or if you want more information about sponsorship and mentorship email me at darrenclifford@pairtrader.com.

    Cheers,

    Darren
     
  477. Hi Djclif

    Pair trading is quite interesting specially in the summer months due to smaller ranges. My concern in not trying it is that the skillset required to pair trade violates several no's no's you learn in trading 101 such as never average a loser, when u are wrong get out instead of hoping for mean reversion. Basis your experience are there traders who pair trade AND momentum trade at the same time. To me , it is sorta like practicng Karate and Tai chi at the same time.. insights are most welcome .Thanks

    Do you use options to hedge "hairy" positions also...
     
  478. thats why you have to trade pairs based on technicals as well as all that (mean and blah blah blah stuff)
    pick your spots and trade
     
  479. Are most of your traders being mentored at the PairTrader.com office recent college grads? Or that age group?
     
  480. GATrader:

    Generally we do not trade pairs from a momentum or trend trading position. This, of course, does not mean it can not be done. Some things to consider if you wanted to pursue trend trading or momentum trading pairs: look for stocks that have trends in two separate industries (at least you diversify away from general market risk). You could mix strategies and be in one trade for momentum and one trade due to support/resistance. This way you have justifiably reduced market and sector risk. We have found it to be easier to teach a contrairian trading style to new traders.

    No we do not use options. Institutions do an excellent job of pricing them so all transactions are giving the edge to the institutions.

    Seisan:

    Our traders vary from being young and just out of college, to sixty year olds that have been around the block a few times. Most of the younger traders we have in the m&a trading, and the older ones are into portfolios.

    Cheers,

    Darren
     
  481. I've made my way through most of this thread... congrats to all the folks who have made it work successfully. You must actually find it unfortunate that pair trading has become a fad, because the more exposure, the less profit opportunity as the market becomes more efficient.

    My thinking on pair trading is this:

    a) The "easy" opportunities are gone, and have been disappearing over the last few years. I have talked to someone who worked for a fund-of-funds group (allocates to hedge funds) - he commented that many stat arb / quants have been having a hard time lately. Simply because, now anyone can download a price series and calculate correlations & mean reversions in Excel, provided they have specific knowledge. Back in the 1980's it cost thousands to get the kind of data now on the internet for free. This sounds like it corresponds to the experience of many individual pair traders as well (decreased performance).

    b) Given that the "easy" opportunities are gone, pairs trading is just like any other style of trading that is widely employed - eg momentum - in other words, it is *hard* to find an edge. Various strategies have their day in the sun, go away, come back, etc. The long term results of trend following CTAs bears this sort of phenomenon out...

    c) I don't think I could ever bring myself to have a trading plan that requires averaging down. I know some people do it - and not saying that it will happen to all - but I think the risk that one day you get hit hard with a really big loss, that wipes out many small gains, is there (statistically, it is a strategy with a negative skew). I think new traders find pair trading appealing because it dangerously implies you can avg down with abandon, which is generally a trading no-no. (To confess, though, it's not like I have never succumbed to the urge to avg down - just that I rationally know I shouldn't)

    d) Therefore, the only type of pair trading I could consider (and this is a personal statement until seeing someone's account record converts me :) ) is one where there are stops and no averaging down is allowed. It is fine to trade countertrend with this, same as one can trade a straight price series countertrend. It is just that you define the loss point as somewhere nearby, past support or resistance. But personally, I would rather pair trade using a trending relationship, where the benefit is that the market risk has been taken out. I would definitely consider trading stocks in different sectors against each other, or even baskets of sectors - eg the SMH vs some other ETF...


    Just my thoughts & I welcome any feedback...
     
  482. I generally agree with your points but you won't be able to use simple "stops" with pairs trading. You'll need something more sophisticated like a real time download of prices to Excel to allow you to calculate the price delta between pairs and then an upload to execute your stop orders.

    There are still pairs opportunities out there but you have to be patient, particularly if news or other events cause a sudden shift in the relationship. I have had a high percentage of winning trades (about 95%) but the loosers have been killers. So I haven't been as active as I once was with this strategy.
     
  483. could you explain your option spreads strategies ?

    Here in Brazil, I´m implementing pairs based on ratios of implied volatility with different strikes (implied volatility surface spread). The idea is to go Vega neutral in the pairs when they cross 2SD.

    Are we talking about the same subject ?
     
  484. Hi Everyone,

    Have read this entire thread, and wondering why there has been no input in almost a year.

    A few questions to any pair traders.

    1. Is there an auto spreader that can be used with equity pairs-- similar to TT's xtrader that can help with execution/slippage issues?

    2. Is pairstrader.com a useful tool that is still being used by anyone. e.g. Are there other sites that are out there.. searches haven't turned up much.

    3. Any suggestions for essential reading for learning more about pairs.

    Thanks for any input that anybody out there might have.
     
  485. I have got through half this thread so I am not sure if this question has already been answered, but can trading pairs be executed on a intraday basis and be profitable? I'm trying to develop a pair trading model for day trading, becuase there has to be more to day trading then just buying low and selling high, but thats just me.
     
  486. Speaking about pairs trading in general, it relies on the assumption that there is a convergence (or divergence) between two securities that will occur sometime in the future due to some measured data (stat arb, for example). As such, if you're talking about two non-similar securities, then it's tough to profit from an intraday pairs trade. Now, if you're talking about the old RD Shell, for example, where you're doing a pairs trade on an identical security then intraday is a great time to make a profit as supply and demand will cause the pair to trade at a premium or a discount. So, in short, you can do it intraday, but I doubt it will be as profitable as a longer-term view (if they are not identical securities). Just my 2 cents....
     
  487. Hi all,

    I've read all of this thread and have noted a few issues.

    Everyone mentions correlations and ratios, but has anyone tested these for their statistical significance?

    I'm coming from the statistical arbitrage perspective, and based on the regressions I've done, it is more important that the pair are cointergrated. Then it is just a question of plotting the standard errors and identifying by eye what level to put the trades on. Hence no use of technical analysis or fundamental analysis.

    I think this is why some have blown up doing pairs trading, because they are thinking that if the pair are correlated and have been trading in a certain ratio, then it is a good pair to trade. What you have to bear in mind is when something is correlated, all it means is if it goes up, so does X and if it goes down, so does X. There is no reason they should converge or diverge. Or put another way, just because they are correlated, doesn't mean they are cointegrated.

    Competition has taken away alot of the profits available in pairs trading. Given the 10,000 pairs I've looked at and their profitability of at best 8% per annum in the last year, you might as well invest in an index tracker, and save the effort.

    If you want to know how it is done properly, read Pairs Trading Quantitative Methods and Analysis by Ganapathy Vidyamurthy.
     
  488. With all due respect...
    You miss the forest by focusing on a tree...
    That hoary old correlated versus cointegrated mind f*ck...
    And your conclusion is simply not correct.

    There are 1000s of quality Listed pair combinations...
    And a pair must satisfy a mathematical corelation such as > 0.90 over medium term 3-6 months...
    And make sense...
    Meaning one would rationally expect these 2 securities to be highly corelated.

    Also in the real world...
    Quants traded basket of corelated securities...
    So the focus is never on a single pair.

    It's best to corelate a universe of stocks to an "index"...
    Such as a 100 gold mining stocks to the price of gold...
    To give a very volatile, but probably viable, example.
    One might be long 20 gold stocks and short 20 gold stocks...
    And trade in and out in high volume... all in the context of the price of gold.

    It's really all about volume, volume, and volume...
    In post-decimalization 2006.

    But there is no easy money lying on The Street...
    And this requires lots of very expensive propriatary software...
    And execution MUST be in the hands of a very experienced traders...
    Because in the above example universe (which I do not trade)...
    There would be many pitfalls ** specific to ** the gold mine universe...
    That could never be screened out in a quantitative manner...
    But require highly experienced decision making.

    Why?

    Because AI software... specifically "expert systems"...
    Are not anywhere close to the point...
    Where it could beat a pro poker player...
    Or match the complex and real-time decision making of a pro trader.
     
  489. Yes, but not in the traditional way. Not in the way that the article suggests.

    nitro
     
  490. Yes, it is possible.

    nitro
     
  491. I read the Louis Capital article 3 months ago...

    They are in the business of selling financial products based on pairs/basket trading...
    Very expensive financial products...
    Where 100 hedge funds may be getting the same information...
    Unless you pay them extra to get special treatment or data.

    This is way, way inferior to developing similar systems in house...
    Which is what I have been doing for > 10 years.

    I'm posting record profits for my firm this summer and saturated...
    Have way, way more good pairs/basket trading opportunities than I can possibly execute...
    And might go from 200 trades/day to 400 trades/day in the next year through internal growth...
    But could easily scale 10:1 with some infusion.

    This is probably the one must read thread at ET.

    Unfortunately, most people are ** hopelessly unqualified ** to be in that Top 5% that beats Zero Sum games...
    And too many people are determined to beat their head against a wall in a dead end...
    While many others sabotage their trading and life... because they fear success.
     
  492. Just to add some critical analysis the Louis financial product...

    When I was slaving away on my Computer Science degree in the early 80s...
    "Data mining" was a ** pejorative term ** describing a misuse of statistics...
    To dredge up meaningless patterns from mountains of data.

    Today... according to Wikipedia...
    The terms used for this fallacious methodology is "data dredging" and "data fishing".

    When I read the Louis sales pitch...
    Scanning 180,000 stocks to find the best 10 pairs or whatever...
    It smells totally like "data dredging"...
    Which will produce many spurious pairs mixed in with solid, tradeable pairs.
    And who is qualified to weed out the dogs?

    I wouldn't take Louis Capital's data... if they gave it to me for free.

    My approach...
    Where you limit your analysis to a homogenous universe...
    Like the above ** for example only ** gold mining stock universe versus gold futures...
    Is infinitely superior...
    And almost completely avoids the problem of "data dredging" and spurious pairs.

    But it requires hardcore expertise on the part of the traders...
    Which is always in very short supply.
     
  493. I only posted the link(s) because they were "real world" examples. Not saying that they are good or bad, but just an outline to understand from reading/book knowledge about pairs to seeing a client and fund real world example.

    I do agree that with anything involved for trading that you will have to search and develop for an "edge."

    Thank you for your insight. What can you recommend for an individual investor in terms of programing language and database for getting started? VBA, MySQL?

    Is there anything thing else you can pass along without giving away too much?
     
  494. As a professional software engineer who became a trader at age 34...
    And has spent most of his life obsessed with developing quantitative trading software...
    It's hard for me to imagine how a "hobbyist programmer"...
    Could have the "intellectual training" in structure and logic and software design to make this work well.

    It's the real-time software using APIs and Automated Systems that is beyond the capabilities of a hobbyist...
    But using Excel and VBA, I suppose, a determined hobbyist could develop a perfectly viable ** small scale ** basket trading system.

    A cab driver may deliver a baby...
    But is totally f*cked if any complications arise.

    On the other hand...
    This is not rocket science...
    And the language and tools you choose don't really matter...
    And about one second latency in real-time systems is OK...
    Because you are never gonna be competing on the millisecond playing field with the Big Players.

    And more critical analysis on Louis Capital and the like:

    (1) Pairs of very liquid securities are completely worthless... the financial markets are too efficient.

    To use an extreme example...
    If the US 2 Year Note and the 5 Year Note drift away from a historical ratio...
    It's common knowledge... and in no way a trading opportunity.

    (2) Great tradeable pairs are mostly found in complex or illiquid securities...
    What I call orphan securities... because no one is interested in them.

    But if Louis Capital gives an illiquid pair to 100 hedge funds at 8:00 AM...
    Both stocks are gonna totally gap at 9:30... and trade unnaturally for days/weeks.

    For sure there's lots of front-running going on...
    Like the people paying $1000 and getting the list on Tuesday...
    Are buying stock from people paying $10,000 and getting the list Monday.
    Even if this was illegal... so what... hardly no one ever gets caught doing anything fraudulent.

    That's why this kind of thing must be developed from scratch in house.
    The idea that you can get winning trades from a Third Party is ** always an illusion **.
     
  495. Surprisingly insightful for ET.

    Another thing I would add is that perceived liquidity can be incredibly misleading.

    Also, it's important to mention the issues associated with trading illiquid securities in this manner

    1. Impossible to put on size without giving up a lot of edge, thereby putting a nice cap on the money you will make trading this way

    2. Once you have size on, it's almost impossible to get out without giving up a lot of profits, especially when the illiquid leg is the one moving against you

    3. Illiquid and complex securities are illiquid for a reason - one of them could be their complexity... many times there may be a perfectly reasonable explanation for the divergence that evades you
     

  496. In response to your points:

    (1) One does not trade illiquid securities "in size".
    Rather... one tracks 100s of illiquid securities and the overall volume adds up.
    So the absolute profit "cap" might be somewhere in the low 8 figures for NYSE listed...
    Which means the Big Players and black boxes are out of the picture...
    And, often, it's just me and the Specialist... and some widows and an orphan or two.

    But my methodology is very adaptable...
    If it ever came to wrestling with 8 figure "problems".

    (2) I always have 2 things in my favor:

    (a) entering into a mispriced position due to market inefficiencies with reversion to mean on my side
    (b) the bid/ask spread...
    which is still typically $0.05 to $0.10 in a stock that trades 50K-100K shares/day

    (3)
    > many times there may be a perfectly
    > reasonable explanation for the
    >divergence that evades you

    Yes.

    That's why I have repeatedly said...
    That this only works well in the hands of very experienced traders.

    The main benefit of experience...
    Is that one has suffered through every possible pitfall indigenous to a class of securities...
    And therefore avoids repeating such losing trades...
    That are not easily screened out using quantitative methods.
     
  497. Thank you for your continued expertise and insights.

    For you positions and trading model, can you state the average holding period? Holding overnight?
     
  498. This applies to NYSE/AMEX stocks...
    With average daily volume between 5,000/day and 100,000/day.

    Holding period minutes to hours to 1-2 days...
    To absolute max 1-2 weeks for trading out of problem positions at a loss.

    If you are scalping for small profits...
    You have to be absolutely ruthless in closing out or hedging losing positions.
    This is probably the hardest thing for a trader to learn to do.

    You cannot trade lower volume stocks without holding overnight...
    Because you will be forced to liquidate too many positions at sub-optimal prices...
    So that is a significant limitation in terms of leverage.
     
  499. is anyone aware of a way to do this in tradestation besides putting both of the stocks on a chart and visually viewing the spread? What is the best way to run the correlation? Is there an outside program that you need or can it be run in tradestation?
     
  500. nobody?
     
  501. Excel or OpenOffice equivalent.
     
  502. Tradestation is limited in what it can do with pairs. You can plot both symbols on the chart and create an indicator which is the spread between the two. They have an included indicator spread difference and one for ratio. If you want to beable to create a custom spread, say .7 wmt to 1 fd, you'll have to modify the indicator yourself.
     
  503. hey, thanks, do you know what the indicator they included is called?
     
  504. Anyone recommend a blog/journal website centered around tradeable pairs listed from a broad spectrum of industry groups?

    Somebody must have a trading blog on the web for pairs trading.

    Then again, this thread seems to have completely died.

    Trying to add to the usual suspects:

    HD/LOW
    INTC/AMD
    CVX/XOM
     


  505. How can something be centered around something else? Anyway, there is a reason you seldom hear about pairs trading, and it is the same reason this thread is long dead. It does not work.
     
  506. Thanks for the self-centered double vote of bunk and negativity.
     
  507. Sorry to burst your bubble, but there is no edge in pairs trading. There has not been in a very long time. The purveyors of pairs trading books and software, along with the brokerages who charge double commissions, they would all like you to think otherwise. It has nothing to do with having a negative attitude.
     
  508. Actually... you are 100% wrong.
    Your speech would apply to TA, though.

    Pairs and basket trading is so profitable ** if done right **...
    (And it's not easy to do it well)...
    That no one will talk about it in detail...
    It's like "throwing pearls before swine."

    You have to be BOTH a great quant... and a great trader.
     
  509. I am going to go out on a very long limb here and say that you have no idea what you are talking about. No one will talk about stat arb in detail?

    Inference and Arbitrage: The Impact of Statistical Arbitrage on Stock Prices Tobias Adrian MIT

    Testing Market Efficiency using Statistical Arbitrage with Applications to Momentum and Value Strategies S. Hogana, R. Jarrowb, M. Teoc*, M. Warachkad 2003

    The Cointegration Alpha: Enhanced Index Tracking and Long-Short Equity Market Neutral Strategies ISMA Discussion Papers in Finance 2002

    Cointegration and Asset Allocation: A New Active Hedge fund Strategy Carol Alexander 2001

    A Computational Methodology for Modelling the Dynamics of Statistical Arbitrage Andrew Neil Burgess 1999 Phd thesis

    Statistical Arbitrage and Securities Prices" Oleg Bondarenko University of Illinois at Chicago

    High Frequency Pairs Trading with U.S. Treasury Securities: Risks and Rewards for Hedge Funds Purnendu Nath London Business School 2003

    I have over 100 papers on the subject going back to 1988. Yes sir, there is definitely a lot of money still on the table here.
     
  510. You are going out on a "very long limb here"...
    Because I make a very good living pairs/basket trading for 14 years...
    And profits lost through decimalization have been made up through low costs and higher volume...
    And the changes in the last few months have had little impact on quants...
    In fact, may still end up being a net plus.

    And at the heart of the Bright operation is a strong focus on this...
    In fact, the web site pairstrading.com = Bright Vancouver office.

    The fact that you can find a bunch of academics to address an issue is a Big Yawn...
    They are amateurs.

    Try getting pro traders to talk about it.

    Look at this thread.
    90 pages and 110,000 views...
    And virtually no useful, specific information on this area.
    That's what I mean.

    Getting this kind of approach to work is a matter of trial and error and experience...
    And expert application is everything...
    Meaning (a) proprietary software and (b) expert traders.

    Other than basic concepts...
    You cannot get this out of a book...
    Or run off-the-shelf software...
    You have to invent and build and create and constantly tweak over the years.

    I could come up with literally dozens, if not 100s, of workable groups of stock to trade/hedge...
    Some would work better than others...
    But I have more than enough action to keep me busy for the foreseeable future.

    Even is I posted a 10 page exact explanation in detail of what I do...
    Maybe 1% of the people here could duplicate what I do...
    And it would take that 1% at least 6-12 months to be up and running.

    But because the basic idea is always that you are arbing less efficient areas of the markets...
    Certain niches regularly get saturated and cease to be profitable...
    So, in that sense, your point is correct.

    For example...
    I traded convertibles very profitably for 2 years 2000-1...
    But since the convertible niche got hammered by the tech meltdown in 2002...
    And the energy meltdown in 2003...
    It's not very tradeable... and I have long since moved on.

    Staying ahead of the curve here is paramount...
    And that requires ongoing creativity.
     
  511. I started doing pairs trading last year and I wrote about my strategy on my blog last june. Scroll down and you will see my trading idea. Also there are links to www.market-topology.com. and www.spdrindex.com/correlation.

    I like subsectors of XLE and XLF for this type of strategy. I don't currently use this strategy, but think I may go back to it because it keeps you trading in the right direction and if you stick to a scalping approach with this strategy you are less likely to have big losing days when you do make trading errors.

    Here is my blog with the pairs strategy idea:
    http://highprobability.blogspot.com/2006_06_01_archive.html

    #1 trade- in the direction of the VWAP for ES futures contract
    #2 trade- in the direction of the VWAP of the sector your trading
    #3 trade-scalp slow to catch up stocks following your sector in the direction of the VWAP
    #4-dont buy on nyse tick highs or short on nyse tick lows
     
  512. Pairs trading no longer profitable? Don't tell my pnl that. :p
     
  513. Seriously.

    dribbling must not work in basketball because I can't do it as well as kobe.
     
  514. hahaha LOL.
     
  515. I have been using this Tradestation PlugIn for 2 months to day trade pairs and have had really great results. Definitely not the holy grail of trading, but it supplies a logical, easy to understand trading system for pair trading. It has real time alerts to get you in trades as well as out of them. Also, the developer is great and answers all of your questions in a couple hours. Its the best $50 I have ever spent!

    www.daytradepairs.com
     
  516. Just wonder if there is broker / platform that has an edge for executing pair trade orders?
    Thank you for your insights
    Berbo