Hey guys, thought Iâd jump back into the fray. I was very intrigued when this journal was started and paper traded alongside the system for a month or two before making a total of 55 trades last February, March, April. My final tallys were 69% profitable and 31% losers ⦠but my losers averaged about 1.5 times the size of my winners. (My winners were smaller). Overall, still nicely profitable. This was compared to stats I had collected to the same point as Johnny, who had (when I stopped counting on 4/10/2009) 65% winners vs 35% losers over about 100 trades; but had average losers that about equaled average winners â so he was much more profitable. Initially I was doing great from a profit perspective even with my smaller winners but realized that while our losers were about the same size my winners were much smaller; meaning I was taking more seemingly safer trades. At that point I had something like 75% winners; I remember writing in this journal that I had something like 17 straight profitable trades. Being dumb or aggressive or both I decided to start trading what Iâd consider wider patterns to increase the size of my winners. This led to a handful of big losers and meanwhile I got involved in a new business ventureâ¦so I stopped trading. I decided to get back into it as of this past Monday and would like to share what Iâm doing as I learn. I still consider myself quite a novice. First, my system I think is pretty true to the original system on this journal. I try to find ratio charts that are fairly horizontal and tend to bounce from one high range to the other. I also try to find stocks that have a history of being highly correlated over longer periods of time (150 and 200 days) but have fallen off somewhat recently. My entry, true to Pairtrader (at least a year or so ago) is when the moving average of the ratio exceeds the standard deviations of the ratio. I originally used pairtrader too (and own it) but I also recreated the signals using AmiBroker and find it much simpler for me to scan and study. I was using the pairtrader charts as confirmation (since for some reason they just look a little different even though they are the same formula) but I broke my version of pairtrader by putting in too many pairs and have to reset the database and havenât taken the time to do that yet. My personality is to overcomplicate things so I originally said I would completely ignore fundamentals, news, and just trade the charts. I did this and the only thing that I recall deciding I would change this time around was earnings announcements. A few of my biggest losers last time around were due to one stock having earnings after I had entered into a trade that caused it to go nuts in the wrong direction. I also favor the concept of a large basket of smaller positions so even after only a few days I have about 25 pairs right now. Assuming I keep finding pairs some will begin to rotate out and others in so I will probably hold 30 to 40 pairs most of the time. I favor this approach in part because following the original journal I did not use stops and waited for an exit signalâ¦and I wanted to minimize my risks on any one trade. And, everything Iâve read about pairs says its all a numbers game. I mentioned I did not use exits but I have since read at least one study that said the optimal time to exit a pair trade is 11 days and one thing I found was that most of my winning trades closed in less than 10 days and a losing trade after a week or so tended not to get too much better so Iâm thinking I may set an 11 day stop..or at least using discretion after 10 or 11 days. I also have included some ETFs to see how they work out this time around. I had not done that before. My ultimate goal is to make this fairly mechanical which is why Iâm still trying to avoid incorporating news and other information (other than not trading a stock during earnings announcement)â¦otherwise it would take me hours a day which I donât have. I have limited time tonight so I will share some trade I entered into this week (many of which are not profitable yet so still offer an opportunity to enter) and some I found tonight and may enter first thing tomorrow. Ongoing I am not always at my computer during the day but will try to regularly post what Iâve found for that day. Assuming I donât completely crash and burn I will try to keep posting. In the future I will try to include some screen shots. I feel like I still need to remove some rustâ¦as I review my earlier entries this week the charts donât look as clean as I had originally thought but for better or worse, here are some I am in at this point⦠(long is first) HCP/VTR BNS/BMO ESS/MAA USO/OIL XES/IEZ VDE/DIG LRY/NNN (perfect ratio chart for last 6 months for me) ARD/BTU NHI/OHI ICF/VNQ (playing two ETFs I think.. I also love this 6 month ratio chart) NU/WEC.. (also like this 6 month chart) EQT/NBL XLK/XLB A few that look interesting tonight include: IWV/VTI PWE/PVX STR/HAL AVB/BRE UTX/CBE CLR/UNT Mike
Thanks, Mike. Best of success to you. The list of stock looks interesting. I'm looking forward to you posting charts with the pairs -- it really helps. I'm not wanting to enter each pair just for a chart. However, if you end up boasting some fantastic results, I might do the extra work to ride your coat tails. And 40 pairs at time, 80 stocks! I'd find that impossible to keep track of. Like herding cats. Very expensive tempermental cats. I reduced my pairs from 6 or 8 to 5 now, and it's much more comfortable - for my style. And also, I'm not seeing more that are any good to trade. My universe of pairs that I scan must be in the area of 200. 40 pairs out of my 200 would be 20% in trade at any one time. Your universe must be tremendous. How did you get that universe? Walt B
Walt: I do have quite a few...around 3000 in my database. Because the scan i built looks for a few key criteria i only get maybe 50 or 60 to look at on any given day. The first time around i spent quite a bit of time w/ pairtradefinder using the industry or sector lists (don't remember which) and going through his method of finding correlations but it was very time consuming. this time i used 3 or 4 sources including past trades, a site that offers stock correlations (don't recall the name right now) and just kept adding them to the list. There is actually no harm to me in having a large list as the end of day scan still only takes 20 or 30 seconds to pull out the 50 or so that are candidates. If I begin to see that certain specific patterns or charts are better I may decide to be more selective in the future but right now it is hard for me to (based purely on the charts) select among the 3 or 4 best I see each night so i just feel right now i increase my odds by splitting my money between them rather than trying to pick 1.
Gee, was looking so good, it hit my profit target with lots of room for more before reverting to the mean, then news hit this AM. The fiz is gone. The pair's gone flat. The profit has drained. NEW YORK (AP) -- Coca-Cola Co. has little room for its stock to go higher, and the company is being pressured by volume declines in its developed markets, a JPMorgan analyst said Friday as he downgraded the world's biggest soft drink and juice maker. The Atlanta beverage company's shares fell $1.04, or 1.9 percent, to $55.15 in morning trading. The shares have traded between $37.44 and $59.45 over the last year. John Faucher said in a client note that Coca-Cola "is doing almost everything right," but with the stock trading at an 18 percent premium to the beverage group, the shares are not likely to increase much more. Walt B
Sounds interesting. What can I do to get more specific? I'm interesting in using your method of scanning a larger universe. Do you have the time and are you willing to share with us specifics so we/I may duplicate you system? Software, data sources, spreadsheet formulas, methods? It's asking a lot, I know. Walt B
BPZ/NBR is looking good, but the ratio is stretched out so far I'm worrying that it may be the start of an unprofitable trend. May not revert, but take the MA mean down with it. Here is one I've had on the radar since 12/18/09 when the ratio was on the falling side of the mean. Since then it has moved to the better side and is looking good. Long NOC Short HON RSI 60 and rising sharply. I'll wait until above 80. Ratio on profitable side of trend, and rising. I'll wait until I see an indication of reversion. I don't know of any news problems. Nice cyclical history. Caution: Earnings reports are out Jan 29 and Feb 2. Walt B
This one is coming into a good trade soon. A couple of days yet, maybe. Short KBH Long RYL On good side of the trend. Wait for RSI to top 75. Wait for signs of reversion. I haven't check for news or pending earnings, yet. Walt B
Hi Mike, Again, thanks for the long list. Of those, I narrowed them down to these that are most interesting to me. I know you like to have your risk spread over many small trades, but most of the favorable pairs I saw had low profit potential. These have potential for greater profit, so I would favor putting more of my account into these two. Over a hundred days, I'd have the same number of trades as you would over 10, but mine would have a higher potential for the same dollar. XLK long XLB short ARD long BTU short Walt B