Home > Technical Topics > Technical Analysis > The "Flipper" strategy

The "Flipper" strategy

  1. This thread is a takeoff on a thread by abogdan…
    http://www.elitetrader.com/vb/showthread.php?s=&threadid=26567

    In this thread he proposes a strategy which when applied to high beta stocks will, and I quote from his original post…“Effective return will be equal to appr. 1.97% a day which will give you annualized (compounded) about 100*Original capital. Not bad! And there is no risk! You are always "with the market".”

    Later in the thread he posts the formula in full.
    http://www.elitetrader.com/vb/showthread.php?s=&postid=409564#post409564

    Read the original post and you will see why I have called this the “flipper” strategy. I am always intrigued by strategies where “there is no risk” so after a series of comments I promised to code it up and post the results.

    ***********************************************************

    Ok, I had time this weekend to code up the abogdan’s ‘flipper’ strategy in Wealth-Lab. I will present a spreadsheet with the trade results at the end of this fairly long post(s) which will attempt to explain certain decisions I made during coding. To perform this test I downloaded over 100,000,000 records of bid/ask/trade data for KLAC. I picked this stock only because it is the original stock referred to in abogdan’s post. The data was scrubbed of erroneous (or irrelevant) ticks by removing any record that fell more than 0.5% away from the previous record.

    First let it be said that I am actively seeking comments on where I went wrong. Expect me to concede when you have a point and expect me to resist if I feel that you do not. In my experience (I am neither a newbie nor a Livermore) good trading systems end up making sense and trading logic does not magically disappear in a mist of mumbo jumbo. There are enough of us here that know how to move shares that we should be able to generally agree on that process. Let’s have at it.

    ***********************

    Continued...
     
  2. Continued:

    Things I DO NOT KNOW precisely about this system (or areas where abogdan’s statements conflict with his formulas):

    I am unsure how abogdan calculates his “guaranteed price move”. I will quote from his original post…”

    If you take daily price moves and for each day you take either (High - Open) or (Open -Low) which ever is greater then divide it by Open you would get daily guaranteed price move (Up or Down) For KLAC, for example, there were no days in the past 400 trading days when this move (up or down) was less than 1.045% (Let' say 1%).

    I find his formula in conflict with his conclusion. In the past 400 days, using his formula, I find more than a dozen days with less than 1.045% movement and 4 with less than 1%. (8/15, 8/25, 9/17, 9/19). Actually, there are more than that but I left out all published half market days because is seem a fair conclusion that with the added risk, this system wouldn’t be traded on such days. It is possible that my data is flawed and I would be happy for someone to point this out and back it up. Assuming good data and applying his formula to the data, the number that he starts with (1.045%) and uses for his current calculations would not have been used since March of last year and the number that should have been being used is 0.77%. This delta will play prominently in the success or failure of this system.

    Since his posted formula makes sense, I decided to ignore his conclusions and apply the formula as posted. I think what he wrote was merely in error and would be surprised if the formula is wrong.

    Other assumptions that I made in an attempt to make the results as “real” as possible…

    I thought long and hard about how to handle the transaction at the “flip triggers”. If I’m going long do I use the next offer as my transaction value, or do I take out as many offers as is necessary to get the needed shares, or do I use the following trade values etc., etc. For programming ease I finally decided to just use the FIR smoothed bid/ask values at the time of the trigger to value the transaction. In the case of smaller orders this is probably quite close. For larger orders I would expect the slippage to increase even above these calculations.

    For the profit target and trailing stops I decided to use a similarly smoothed average of the previous trades rather than bid/ask data. I consider this to be a better reflection of the actual market and probably would have used these values to trigger the flips if abogdan’s formula hadn’t specifically said otherwise.

    Continued...
     
  3. Continued:

    Abogdan did not post an exact formula for the share multiplier. I sure he felt he didn’t need to as it is pretty straightforward. Following is a brief explanation of how my share multiplier formula works. The intent of this formula is not unlike a Martingale strategy where one doubles each losing bet at the casino. Here of course one does not lose ones entire stake at each flip so doubling is not necessary. Abogdan originally mentions 1.1 as an example multiplier but of course later explains that that number will vary. I constructed the formula as follows:

    ExpectedProf = abs(OpenVal – ProfTarg)
    CumLoss = Total losses on trade series since open including commissions
    Needed Prof = ExpectedProf + CumLoss + 2xCommish
    AvailableRange = abs(TradeVal – ProfTarg)
    SharesNeeded = NeededProf / AvailableRange

    This formula gets us exactly where we need to be each time other than small rounding issues in WL that can become quite large when share reach epic proportions.

    All my calcs and the posted spreadsheet use a penny per share commissions. This will not exactly model every traders commission structure, but it’s close enough for pretend.

    The trigger delta referred to by abogdan was set at $0.06 for this test. I moved it up and down and for the test period this setting was the best. Close together obviously results in more flips and farther apart in more slippage.

    The profit target for this test was consistent throughout the test period…0.65% (0.73% * 0.9)

    The simulation begins on 9/16/03 and ends a few days ago. There are two or three days missing in that time and they are either short market days or the data for that day had obvious problems and I deleted results. I start with an initial 1000 shares assumption only because that is what abogdan referred to in initial post. You will notice that none of the days start out with 1000 even shares – that is the above formula doing it’s job. The more slippage incurred (wider spread) before initial purchase, the more shares it will buy to end up with the expected profit. You will occasionally see less then 1000 shares as the initial purchase as negative slippage is encountered.

    The last three columns in the spread sheet display profit for three different strategies. Unlimited flips, three max and five max. These columns are totaled at the bottom. At one point in the original thread abogdan mentions allow a day to go to 11 flips. Clearly according to this test (which is just that) that would be ruinous.

    Abogdan ask for impressions of his system and early in the thread I summarized my thoughts this way…

    >Generally I proposed that the noise will at times cause an
    >unduly large number of flips on some trading days and that
    >this will kill the gains you make on other days.

    >Generally I proposed that your slippage will at times be greater
    >than can be 're-earned' through upping your position (especially
    >on the short side) since of course the larger the position you take
    >to overcome the slippage loss the more slippage you get.

    >I still stand by both of those general positions.


    Currently the test bears those initial impressions out and so I still stand by those general positions. I am however very willing to entertain the discussion of flaws in my testing. Questions about my coding are also welcome.

    JB
     
  4. 1. From one flip to another, there is like 10 cents of room, what is that for? For example...


    Short KLAC-11-05-2003-0.657 1,242 1.242 11/05/03 09:30 am 59.01 11/05/03 09:36 am 59.13 -178.13 Flip #1 -178.13
    Long  KLAC-11-05-2003-0.657 1,783 1.436 11/05/03 09:36 am 59.13 11/05/03 09:37 am 58.94 -381.98 Flip #2 -560.11
    Short KLAC-11-05-2003-0.657 3,956 2.219 11/05/03 09:37 am 58.94 11/05/03 09:43 am 59.14 -855.79 Flip #3 -1,415.90


    This was picked randomly, the same issue exists for every trade, where is the flip line here? You are shorting 59.01 and covering 59.13, long 59.13 sell 58.94, short 58.94 cover 59.14. What's going on? 10 cents in slippage/coms or am I missing something? If the flip line is 59.01, for example, you would go long 59.01, then when it came back, short 59.01, then when it came back up you would be long 59.01 (of course I would take a tick off or whatever for slippage/spread). There is something grossly wrong here, there is a reason for the multiplier, to cover the losses to slippage/coms of the previous trade :)

    Edit: Only way I can see this not being incorrect is if KLAC has such a huge spread (8-15 cents!!??), if so, I appoligize, I do not day trade stocks
     
  5. Hi Gary, it's a good question.

    >If the flip line is 59.01, for example, you would go long
    >59.01, then when it came back, short 59.01, then when
    >it came back up you would be long 59.01 (of course I would
    >take a tick off or whatever for slippage/spread).

    I think you have a misunderstanding of his system. If we were to use a single line (say the 59.01 as in your example) there would be a huge number of flips just from noise. Abogdan's system instead uses *two* lines (the space between he calls the "delta"

    From abogdan's post:

    Calculate Delta:

    Delta = Abs(EntryPrice - ((Ask+Bid)/2) - Commissions - slippage)

    Find optimum Delta over 250 days that has minimum losses
    using the formula Shares*Delta*Flips

    If filtered BidPrice is higher then upper limit of delta then go long
    If filtered AskPrice is lower then bottom limit of delta then go short.


    I have attached the chart shot from the initial entry and first flip for the day you chose. The red line is the smoothed ask average, the green is smoothed bid, the blue is smoothed trade and the two horizontal purple dashed lines represent the upper and lower flip trigger values and are 3 cents off center each way.

    You can see the trades trigger as the dashed lines are crossed.

    JB
     
  6. Oh ya, forgot about that. Ok, I understand now. Those numbers are using the optimal delta?
     
  7. >Those numbers are using the optimal delta?

    For the test period, 0.06 was the most profitable.

    JB
     
  8. Turok:

    I really appreciate the work you've done. I need to go thoroughly through all of this and get back to you with adjustments that are needed for the system. I'm on the road right now and will try to work on it as much as I can. It might take a few days, but that is what the situation is. On the end, we are running the system successfully in our offices so, I'm sure we can make it work in WLD.
    Thanks again,
     
  9. Ok, I have 15 minutes. First of all, it is really hard for me to follow your code but I'll try to make general comments that jump out right away.

    When you go in (let say Long) in order to determine your shares you have to use "Remaining Percent" which is = TargetPercent - (CurrentAsk - OpenPrice)/OpenPrice. This gives you an idea of how many shares do you need to hold to pay for all the flips plus your target profit. The way we calculate this is following:

    Target Dollar Value of profit = OpenPrice*PercentTarget*InitialShares.

    So at any time your Holding shares should be calculated from the following equation:

    Target Dollar Value of profit + All Losses That occurred before this flip = CurrentEntryPrice*SharesToBuy*Remaining Percent

    Another thing is:

    If you have reached Max amount of flips (in our case it was 11) you are not looking to make profit anymore rather your target is the losses that you encounter so far. This eliminates a lot of losses including days when there is not enough percentage gain.

    There are some other things that we do to make it better, but I need to spend more time to describe them. Please let me know whether this was helpful. Again it does work. It just a little bit more complex than the rough concept that I have described.
    Cheers,
     
  10. turok,

    interesting analysis!

    however, it seems to me that you have a mistake in your "Profit" calculation formula in cell K351: there you are summing *all* cells in the K column. however, the numbers in that column are *cumulative* profits, not *trade* profits. to get the grand total you should sum the daily profits instead.

    or i'm i missing something here?

    - jaan

    EDIT: i can see now that the problem in K351 calculation is not that relevant to the analysis, as the limited flip columns (L & M) do not have that problem.
     
  11. You are absolutely correct jaan. Thanks for pointing that out.

    So, of the last three columns on the right, the total at the bottom of the first one (K) is incorrect and the totals of the last two are correct.

    I didn't catch that as I didn't even pay a whit of attention to that column since I can't afford the 15million dollar account that it takes to buy that many shares.

    Thanks again.

    JB
     
  12. Oh, when I said "losses that occurred before current flip" I mean profits also which sometimes reduces the capital exposure.
    Cheers,
     
  13. that does not make any difference, because after 11 flips your loss is several orders of magnitude greater than your target profit.

    in other words, you are suggesting that after being hit with a $10000 loss, you should "just" aim for making that 10k back, and not worry about making the $300 profit on top of that.

    - jaan
     
  14. Roughly, after 11 flips your shares are at app. = (1.1^11)*InitialShares = 2.9*initialshares. Your loss is about 0.06*11*2.9*InitialShares = $1800 on 1000 initial shares. Your target was 1000*0.6 = $600.
    Cheers,
     
  15. Hi abogdan, thanks for weighing in.

    >Ok, I have 15 minutes. First of all, it is really hard for
    >me to follow your code but I'll try to make general
    >comments that jump out right away.

    Not sure what you mean by "code" since I didn't post any. I'm assuming you mean my formulas.

    You detail your share formula and though I didn't just now go through and test it, I have no doubt that you can and did construct one that does the job. I am also confident that I have done the same. I'm unsure if you are saying that mine is wrong, but for the sake of clarity I will walk through an example using mine and someone point it out if I've gone wrong somewhere.

    Assumptions: (numbers chosen for ease of use)

    1.111% guaranteed price move
    $50.00 open value (OpenVal)
    1000 initial shares
    penny per share commissions
    3 cent spread

    So...

    Profit target = (50 * 0.0111 * 0.9) +/- $50.00

    Our ProfitTarg(s) are 50 cents on each side of our OpenVal or $50.50 for longs and $49.50 for shorts.

    and...

    >ExpectedProf = abs(OpenVal – ProfTarg)

    So...
    Expected profit = (50.50 - 50.00) times initial shares

    Or: Our ExpectedProf is $500 bucks per winning day

    Any problems so far?

    For the first set of calcs lets assume it's the open and we have no trades yet and thus no cumulative losses.

    >Needed Prof = ExpectedProf + CumLoss + 2xCommish

    Our NeededProf is $500 + CumLoss (0) + 2xCommish (unknown until shares are set)

    We just use the $500 number to start and calc commish later.

    Our short and long flip triggers are set 3 cents off center so when our smoothed bid average crossed the upper line (50.03) we lift and offer and go long. With our 3 cent spread our TradeVal is 50.06

    >AvailableRange = abs(TradeVal – ProfTarg)

    So the AvailableRange is 0.44

    >SharesNeeded = NeededProf / AvailableRange

    So the SharesNeeded are 1136

    Now we must calculate commissions...(as you will see this ends up not quite perfect but very close with just one round of calcs)

    1136 * 0.01 = $11.36

    Divide the 11.36 by the AvailableRange and you can see that you must purchase 26 more shares to cover the cost of commissions

    So SharesNeeded = 1136 + 26 (or 1162)

    Using the above assumptions and formulas, my code will purchase 1163 shares @ 50.06 and hope to sell them at 50.50 for a ~$500 profit after commissions.

    At each successive flip the code just adds in the losses so far (including commissions) and recalcs the shares.

    If there is a flaw there I'll happily correct it.

    JB
     
  16. Turok,

    I'm not sure how hard it would be, but if easy for you I would love to see the results of the following changes to your test (shown in your spreadsheet):

    1) Start every initial trade with a fixed 1000 shares.
    2) Use a fixed multiplier of 1.1. In other words, have the net position AFTER the first flip be equal to 1100 shares, upon a second flip the position will be 1210 shares, etc.
    3) With these changes, the share size doesn't ramp up nearly as quickly, thus you can more safely use the suggested "max 11 flips" as the stop loss exit criteria for trading during that day, and see what the results are.

    If this is relatively easy to do, and you don't mind, I would be interested to see the results of this in the same spreadsheet format as you have posted the latest results.

    I just found this thread and the prior linked thread today, and I sure appreciate all of the thought and discussion that has gone into it so far. This is the kind of thoughtful discussion that makes ET worthwhile, despite having to frequently dig though a lot of crap (on other threads) to get to this sort of interaction. I don't think this ongoing thread would have been thought stimulating if it weren't for abogdan's willingness to share his overall idea, as well as to welcome the views of 'opponents' of his idea and the occasional ET idiots that never have nothing worthwhile to contribute. In addition, this thread would not have been nearly as useful without Turok's valuable thoughts and healthy skepticism to continue the discussion (in a non-threatening way, no less) for the benefit of all.

    My hat is off to both of you. Thanks for providing a very stimulating discussion. I wish there was more I could contribute to this discussion from the backtesting point of view, but this is just not one of my strengths.

    Thanks again,
    -Eric

     
  17. >I'm not sure how hard it would be, but if easy for you
    >I would love to see the results of the following changes
    >to your test (shown in your spreadsheet):

    Hi Eric,

    Thanks for the kind words toward abogdan and myself.

    It is not that difficult at all to code the changes you describe (it wouldn't be done in the spreadsheet actually as that is mostly just a display medium for me...the coding is done in Wealthlab)

    I am currently working on my next post on this subject which I think will be of interest to you as it deals with abogdans 1.1 multiplier. After reading that post, if you still want me to do a run with 1.1 I will consider it (each run of 80 days takes an overnight period for WL since it is almost 30million candles of data to crunch)

    Get back to me after.

    Thanks
    JB
     
  18. thx for your work turok.

    here is a small spreadsheet i made. i dont have access to bid/ask prices, but this is a theoretical sheet. i assume delta+commissions is $0.10, profit target is $0.55, and the multiplier is 1.15. as you can see, you start out with maximum profit if you win on the first trade, however as you progress, your winnings once you do get to your 1% target get less and less.

    however, you can fix this by adjusting the flip multiplier, raise the flip multiplier high enough and you can win the same amount no matter which flip it is (or even more with each successive flip)....however, you're also exposing yourself to greater losses if it never gets to your target (see the cumulative loss column).

    however, you were saying that out of 400 days, there were only 12 days that it didnt reach its target, so this seems ok to me :)

    im only sending this because i noticed on your sheet that some of the multipliers were very high, around 1.5 to 2, which will ruin you, i've found 1.25 is about the highest you should go unless youve got a lot of capital and balls.

    let me know if any of my calculations are incorrect.


    just u/l a new copy of the attachment, i forgot to offset the loss columns up one row.
     
  19. Hi SimStim. Thanks for your contribution to the thread.

    >im only sending this because i noticed on your sheet
    >that some of the multipliers were very high, around 1.5
    >to 2, which will ruin you, i've found 1.25 is about the
    >highest you should go unless you've got a lot of capital
    >and balls.

    I haven't had time to review all your work (headed out for an evening with friends), but I do wish to comment on one aspect of your post.

    We as traders don't get to dictate the multiplier. It's not a matter of picking a number that fits our "capital and balls". As is demonstrated by my posted and detailed formula, the multiplier is dictated by the delta,slippage and expected profit. The multipliers in the spreadsheet are the result of those factors and if you disagree with those numbers you are free to challenge any of the assumptions that are made there. (and I will happily show you screen shots of any trade you question)

    Saying "i've found 1.25 is about the highest you should go" is equivalent to the statement "if the spread is too wide I don't take my stops". We don't decide the spread and we can't know the spread in advance. For this system to work one MUST continue adding shares at a rate that will cover previous losses, commissions and still end up with the expected profit.

    So, I'm still happily entertaining any flaws in my formulas and assumptions, but setting a limit on the multiplier is a sure route to disaster (and as I noted to Eric, I will back that up with a code run if I don't convince you all otherwise).

    Again, thanks for the contribution and I will review your work later this evening.

    JB
     
  20. Turok:

    That is an excellent first step! Now I have to give you the rest to make it work! So, you can see from the algorithm that the winning days is not a problem its losses that erase everything. That is exactly how I started. Now, how many days do you think are there that do not meet 1% target within Max allowed flips? In my trading journal for KLAC it was 18% an average with 11 flips.
    Agree? If after 11th flip you start minimizing your losses you'd see that the average loosing day is app. 2*winning day. So your win/loss is already 64%/36%. Adding trailing stop to the win days makes it roughly 72 - 76% success rate.
    Now the good part! You can have 100% success rate! How?

    Imagine you have a portfolio of stocks (we use 200 stocks per portfolio). You allocate your capital spreading it evenly through 200 stocks. Then you start flipping. You will quickly notice that some of the stocks go to the profit right away. So, you lock your profits and reapply the capital that you did free up to the stocks that are still struggling. Slowly you work through the stocks and on the end you'll end up with 18% of stocks that did not make it. But you already locked the profit to cover this loss by other stocks! So, 100% success rate (meaning that every day you end up being positive).
    Cheers,
     
  21. >Imagine you have a portfolio of stocks (we use
    >200 stocks per portfolio).

    I'm currently of the opinion that it doesn't do any good to multiply an unprofitable (or barely profitable) system by 200. I may become convinced of this later, but for now let's just make sure that we are doing what we should be doing with KLAC. Once we get to that point we can look at spreading it around.

    Later.

    JB
     
  22. No, not multiplying, spreading! That is the key! You make your efficiency of the capital use by far higher!
    Think about it.
     
  23. turok, i disagree with you there.

    the multiplier has a simple formula, it is 1+(slippage/profit target).
    so if your slippage is $0.10 and you're trying for a target of $0.50, your multiplier should be 1.2. this gives them an even amount of profit no matter which flip it is (assuming they get to the target)

    my reason for saying capitals and balls is that if someone wanted to, they could raise this multiplier (say to 1.4), which would actually give them a higher profit amount with more flips (also a greater amount of loss if the price didnt reach target). but if it is rare that the price never reaches target, then this would be worthwhile.

    please disregard my last spreadsheet, i noticed another error.

    if you notice, i put:

    multiplier 1.1
    delta+comm: 0.06
    profit target: 0.60
    this gives you a constant $600 across flips, if you decide to quit on the 10th flip of the day because it hasnt reached target, you would lose $956.25.

    say for sh & gi you decide to raise the multiplier to 1.2:

    multiplier: 1.2
    delta+comm: 0.06
    profit target: 0.60

    you see now that each successive flip you get more profit. if you were to finally reach target on the 9th flip, you would have $1589.95. however, if you did not make it by the 10th flip, you would have lost $1557.52 (greater than the scenario above). this is what i meant by someone have greater capital and balls could change the multiplier.
     
  24. I think you are misunderstanding Turok. He understands the value of diversification, but he is just not yet conceding that the strategy itself is profitable over the long run with a single stock under question. Certainly, if successful, then diversification will help. But, if the strategy itself fails over the long run, then diversification will not help at all.

    I agree with Turok at this point, and think that we should keep the discussion focused on one stock to ensure that he is able to duplicate similar performance to your results using his backtesting capability.

    That said, if successful, the logical extension of this is certainly to expand the system to numerous similar stocks to diversify results and gain an even better (smoother) equity curve.

    Thanks again,
    -Eric

     
  25. I have an idea..

    How about enough of the theoretical crap.. and someone start actually trading this system live with real $.

    When u actually put $ behind something u will learn a lot more than u think.



    ---MIKE
     
  26. Hi SimStim. Thanks for your participation in the thread.

    I appreciate your thoughts but I would have to say that Turok's algorithm is by far closer to the actual thing that we are running. The reason we have chosen this approach is because it is much easier to implement in real life. You do need to calculate your actual losses that sometimes through a curved ball!
    Regards,
     
  27. Ok. I'm also very interested in the capabilities of WLD. This is a good experiment for all of us! If WLD can match my results then it is going to be worthwhile for me to learn it.
    As you get older you need to apply a lot more efforts to learn new things eh?
     
  28. Ok, have to go now. I'll catch up with you all tomorrow.
     
  29. TrendFader:
    >I have an idea..

    >How about enough of the theoretical crap.. and
    >someone start actually trading this system live
    >with real $.

    >When u actually put $ behind something u will learn
    >a lot more than u think.

    Hi Mike, I'll let you start.

    For myself I find it better to hammer on any mechanical system for a while to get a reasonably good understanding of it's behavior before braving the shares. Especially a system such as this which appears to be able to go on for quite some in a profitable mode only to whack you REALLY hard when the going gets rough.

    Even if someone says that they are trading it profitable (which abogdan does) I prefer to watch over my own account by investigating it for myself.

    Onward.

    JB

     
  30. hi abogdan, i didnt mean to usurp turok's algorithm in any way :) he did all the work, i just wanted to give him a sheet to look at to see the multiplier at work. it seemed in his sheet that that was the 'weakest' part of it. there were days that there were cumulative losses of nearly $70,000 because the multiplier he was using was so high, when that shouldnt be the case.
     
  31. To all contributing to the thead:

    Thanks for the input so far. There are quite a number of questions that are being asked and theories being presented and I appreciate all of them. I have put my next (long) post on hold to address a few of these questions. I will try to address these in some kind of logical order so if I don't get to yours right away, please be patient.


    Thanks
    JB
     
  32. Hey Simstim,

    Thanks for your contributions to the discussion and I don't consider anything you are doing to be "usurping" in any way. Keep the thoughts coming.

    You mention a disagreement and that is fine. Before I address that particular, let me comment on your (second) spreadsheet that you posted. I think I understand what you are doing there and I believe there is a significant flaw. Let me demonstrate using your assumptions (6 cent delta/spread/commish, 0.60 target)

    Let's start with a $60 stock for discussion. Your targets are $60.60 long and $59.40 short, agreed?

    We will arbitrarily say that we go long first and we buy 1000 shares @ 50.06. Agreed?

    For the sake of this discussion the stock immediately turns and heads down. Our trigger to go short will occur with a sell price of 40.94, agreed?

    Buy 1000 @ 50.06
    Sell 1000 @ 40.94

    Loss = 0.12 per share

    So, on that first trade (and also all subsequent trades) the loss is exactly DOUBLE of what you have in your calcs.

    Agreed?

    With the double losses, the multiplier must be higher and the shares get bigger and the losses get bigger and on and on.

    JB
     
  33. Ok, I'm going to see if I can steer this conversation back to the actual testing and away from the assumptions that we have been making the last few posts. I know we have been using nice round numbers for the sake of discussion, but part of the problem is that those numbers often look better than my testing revealed.

    Here are some hard numbers from my 100,000,000 records of KLAC data.

    Average spread (difference between smoothed bid/ask) at time of trades: $0.036

    Most profitable delta: $0.06

    Average trade price: $57.01

    Profit target % (using abogdan's formula) 0.65%

    You guys are wondering where the big multipliers come from? Well, run some numbers and it's not hard to see.

    $57.00 stock * 1.0065 = 57.37 (long target)
    $57.00 stock / 1.0065 = 56.63 (short target)

    Ok, there is 37 cents to play with in profit. (notice it isn't 60 anymore like in SimStims example)

    We can't negotiate the spread...it's there in the data for all to see. We already negotiated the delta to it's most profitable point. We can negotiate commission, but if the system won't stand up to a penny a share I'm not interested.

    So for each flip we have a "fixed" cost of delta + (spread*2) + commissions.

    That's 6 + 7.2 + 2 or 15cents rounded off.

    So out of each aborted attempt (flip) at 37 cents we lose 15 of it to those costs. That's over 40% right there. So perhaps now you can see why the average multplier is ~1.45


    JB
     
  34. SimStim:
    >my reason for saying capitals and balls is that
    >if someone wanted to, they could raise this
    >multiplier (say to 1.4),

    Well, if you were to do that you would be trading a different system that isn't being tested here. This system clearly as set out by abogdan uses a formula to determine the size of the multiple as opposed to an arbitrary number.

    (I know this is actually a bit confusing because abogdan occasionally uses a number of 1.1 for discussions sake, but if you read his posts you will easily see that the multiplier is not fixed but rather moves with the situation at hand.)

    JB
     
  35. SimStim
    >however, you were saying that out of 400 days,
    >there were only 12 days that it didnt reach its
    >target, so this seems ok to me

    Does it? I'm not sure how you can say that without determining what happens on a day that the target isn't reached.

    If you decide to stop at 11 flips (as abogdan says he does) then take a look at where your losses are at 11 flips on the spreadsheet. Are you prepared to take 20k losses for each of those 12 days in exchange for the $3-400 days?

    JB
     
  36. Ok, is anyone taking any issue with the example used in this post?

    http://www.elitetrader.com/vb/showthread.php?s=&postid=424142#post424142

    If not I will assume that that is the correct formula for determining the proper number of shares to get the desired profit per day.

    Just so you all know why my confidence is high on that formula, my first round of testing did not include a trailing stop...it just bailed at the target. No matter how many flips we took and no matter how the spreads moved around, we always ended up within a few bucks of the ~$370 profit target per day.

    The trailing stop code makes it impossible to see this result (but did increase the systems profitability significantly) but I can assure you that it is spot on. (and I'll prove it with a data run if I have to)

    Any questions?

    Night all.

    JB
     
  37. I believe you are overstating your 'fixed' cost of each flip. You should include the cost of the spread, but not double the cost of the spread. One trade (a buy at the ask, subsequently followed by a sell at the bid) will cost you a single spread.

    -Eric
     
  38. Thanks all for the efforts, lurking and learning here. Personally I am from the outset sceptical about a double up (albeit by multiplier) strategy as it gets really hard to mentally execute on those days when flipping losses add up and you need to put on a bigger and bigger position just to "make-back" the amount lost. Quite the opposite of the old advice of "never add to a losing position".

    Although I conceede that screening for ideal stocks should help this strategy by a certain amount.

    Best

    J
     
  39. Hey SimStim, I just notice that I screwed the numbers all up in a post that I left for you...not enough Red Bull I guess. The principle remained the same but I just was writing what now looks like random numbers. My apologies and following are the corrected numbers.

    *************************
    Hey Simstim,

    Thanks for your contributions to the discussion and I don't consider anything you are doing to be "usurping" in any way. Keep the thoughts coming.

    You mention a disagreement and that is fine. Before I address that particular, let me comment on your (second) spreadsheet that you posted. I think I understand what you are doing there and I believe there is a significant flaw. Let me demonstrate using your assumptions (6 cent delta/spread/commish, 0.60 target)

    Let's start with a $60 stock for discussion. Your targets are $60.60 long and $59.40 short, agreed?

    We will arbitrarily say that we go long first and we buy 1000 shares @ 60.06. Agreed?

    For the sake of this discussion the stock immediately turns and heads down. Our trigger to go short will occur with a sell price of 59.94, agreed?

    Buy 1000 @ 60.06
    Sell 1000 @ 59.94

    Loss = 0.12 per share

    So, on that first trade (and also all subsequent trades) the loss is exactly DOUBLE of what you have in your calcs.

    Agreed?

    With the double losses, the multiplier must be higher and the shares get bigger and the losses get bigger and on and on.

    ***************************

    I'm sure that is a lot easier to agree with than the gibberish that is previously posted. LOL

    JB
     
  40. correct, but i assumed turok's multiplier (about 1.45, according to him), as your suggestion was targeted towards improving his system. anyhow, i don't want to get in the way here -- so please ignore my remark and continue discussing the "multiplier issue" directly with turok. oh, and thanks for a great thread!

    no, eric, i think turok's correct. for example, a "flipped long" trade unfolds as follows: 1) go long at the ask, 2) wait until the ask drops the "delta" *plus* the spread, 3) cover/go short at the bid. so, effectively, you pay an additional spread in step 2.

    ---

    finally, allow me a short comment on this system from "philosophical" point of view. clearly, if the underlying instrument followed a random walk, this system would have a negative expectancy (because of the spread and commissions). so the system would not work on *any* instrument. but what are the requirements then?

    i think the system's assumption is that the underlying instrument is trending. or, more precisely, the system assumes that every time the price has penetrated the "delta+spread" band, the probability of the move continuing in that direction is sufficiently bigger than the probability of reversal. here "sufficiently" basically means "sufficient to cover the spread, commish, and capital costs".

    so, effectively, the system is just a breakout system in disguise. which means that whenever the underlying instrument is not "trendy enough", it will lose money.

    - jaan
     
  41. EricP
    >I believe you are overstating your 'fixed' cost of
    >each flip. You should include the cost of the spread,
    >but not double the cost of the spread. One trade (a
    >buy at the ask, subsequently followed by a sell at
    >the bid) will cost you a single spread.

    Ok Eric, I'm willing to be proven wrong here, but you'll have to work at it harder than a simple assertion. Please review my example in the following linked post and then tell me where it goes wrong?

    http://www.elitetrader.com/vb/showthread.php?s=&postid=424334#post424334

    Also, review the image posted below which nicely shows a real trade with the delta and the 2x spread as it should execute in real life. The posted trade is the 4th flip on 12/19 if you want to correlate to the spreadsheet. On that trade we go long @56.03 and flip short @55.87.

    Now, I don't really care how you want to divvy up the loss, but the fact is that trade lost 16 cents BEFORE commissions (which is slightly above the average loser for the test period). We know that 6 cents is from the delta so if you want to call it one spread at 10 cents or two spreads at 5 cents I don't care -- it's still the same loss either way.

    (just for the sake of the argument, I'll say that I completely disagree with your premise of "only paying one spread per trade". On a simple long trade, if the spread is 3 cents at both ends and you lift the offer to enter and hit the bid to exit, you will have lost 6 cents (2x spread) to the guy who bids in and offers out (considering exact same trade times). Do the math.)

    Would be thrilled to be proved wrong as it would be a real boost to the system.

    JB
     
  42. >so, effectively, the system is just a breakout
    >system in disguise. which means that whenever
    >the underlying instrument is not "trendy enough",
    >it will lose money.

    I couldn't agree more.

    JB
     
  43. turok, you're right about the larger spread. however, even with .12 spread and a .60 target, the multiplier is still just 1.2, which is doable. if you were to say a .15 spread it would be a 1.25 multiplier.

    btw, where did you get bid/ask data? i have access to wld so i would like to test as well if possible.
     
  44. I just asked one of our programers to give me an idea of what our average share increase per trade is. The number in the past three months was 1.262. That included all the cost of trading. As I mentioned before, our commission rate is $0.005 a share with IB. We use a fairly elaborated order management system to keep our costs down. We also use spread limitation techniques where the limit orders are fired instead of market ones. We also have a box that helps us with the short selling slippage.
    I hope it helps.
     
  45. >turok, you're right about the larger spread. however,
    >even with .12 spread and a .60 target, the multiplier
    >is still just 1.2, which is doable. if you were to say a .15
    >spread it would be a 1.25 multiplier.

    If the ".12 spread" you refer to includes the delta and commissions than your numbers work out fine. However, throwing made up numbers around doesn't cause our accounts to grow. KLAC is at the very tippy top of the range of movement stocks and it's target only ends up at .37. That target number isn't just arbitrarily determined, but rather derived from the behavior of the stock. Apply the .60 to KLAC and the performance of the system get's WAY worse.

    >btw, where did you get bid/ask data? i have access
    >to wld so i would like to test as well if possible.

    I started out in 2001 downloading gigabytes of data from several different sources (some of which are no longer around) and burning them to CD. You can get b/a from Qcharts back 6 months or so I believe, perhaps even further.
     
  46. >>>If the ".12 spread" you refer to includes the delta and commissions than your numbers work out fine. However, throwing made up numbers around doesn't cause our accounts to grow. KLAC is at the very tippy top of the range of movement stocks and it's target only ends up at .37. That target number isn't just arbitrarily determined, but rather derived from the behavior of the stock. Apply the .60 to KLAC and the performance of the system get's WAY worse.<<<

    im not sure what you mean by the target being .37? i thought the target was supposed to be 1% or 0.9% of the stock price? i think this is where im not clear, and where abogdan can clear something up. if we're trading a $60 stock and buy/sell levels are 59.94 and 60.06, are our targets $59.40 and $60.60 or are they $59.34 and $60.66?

    if i misread what you researched above, but didnt you mention that out of 400 trading days, only on 16 days did it not reach a 1% target (excluding the half days)?

    however, abogdan mentions above that their avg. share increase per trade was 1.262 including the costs of trading, so they must be getting good fills i guess...
     
  47. >im not sure what you mean by the target being .37?
    >i thought the target was supposed to be 1% or 0.9%
    >of the stock price?

    Nope, you haven't followed abogdans formula...

    From:
    http://www.elitetrader.com/vb/showthread.php?s=&postid=409564#post409564

    O = (OpenBid + OpenAsk)/2;
    UpSwing% = 100*(O - LowestAsk)/O
    DownSwing% = 100*(HighestBid - O)/O

    If UpSwing > DownSwing Then
    MaxSwing = UpSwing
    Else
    MaxSwing = DownSwing
    End

    Find the Min value of MaxSwing for at least 250 days. (GuaranteedSwing)

    ProfitTarget = 0.9*GuaranteedSwing


    In other words, the target is 90% of the minimum move over the last 250 days.

    >however, abogdan mentions above that their
    >avg. share increase per trade was 1.262 including
    >the costs of trading, so they must be getting good
    >fills i guess...

    abogdan didn't say that this was his average increase on KLAC, but rather implied that it was an overall number. Remember he is running on 200 stocks. Also he has the ability to limit the spread though techniques and boxes. This is clearly an advantage that you and I don't have.

    JB
     
  48. youre right, i didnt see the guaranteedswing part of it. maybe this is where i disagree with abogdan, i think it should be more of a mean value of the past 250 days and not min value, of course using a mean has its drawbacks as well.
     
  49. >i think it should be more of a mean value of the
    >past 250 days and not min value

    I think it should be the most profitable value and the mean ain't it (been there tested that).

    Good luck on your own data download and research. As I'm sure abogdan can confirm, analyizing b/a data over any significant period of time is a non-trivial effort. There is just so damn much of it to claw through.

    JB
     
  50. I stand corrected.

    Average spread (difference between smoothed bid/ask) at time of trades: $0.036

    So for each flip we have a "fixed" cost of delta + (spread*2) + commissions.

    That's 6 + 7.2 + 2 or 15cents rounded off.

    So out of each aborted attempt (flip) at 37 cents we lose 15 of it to those costs. That's over 40% right there. So perhaps now you can see why the average multplier is ~1.45



    I was thinking that the "cost of delta" was the delta between the open and the upper band, and so the cost of delta would need a 2x factor and the spread would not need the 2x multiplier. Thinking through the example, I now see why the spread would need the 2x factor, and I assume that your "cost of delta" is the difference between the upper band and lower band (not distance to opening price). Thanks for the explanation.

    -Eric



     
  51. >Thinking through the example, I now see why the
    >spread would need the 2x factor, and I assume that
    >your "cost of delta" is the difference between the
    >upper band and lower band (not distance to opening
    >price). Thanks for the explanation.

    On a "flip" transaction the above is correct. Of course the initial transaction only costs one half of the delta plus 1x spread.

    JB