System Evaluation

Discussion in 'Strategy Development' started by man, May 26, 2003.

  1. man

    man

    I've tested a long only system for ND and would like to get some comments on the results.

    1995
    1996 16.3%
    1997 37.6%
    1998 12.6%
    1999 19.8%
    2000 63.9%
    2001 -2.4%
    2002 34.2%
    2003 17.2%

    3% riskfree
    28.38% return
    19.34% vola
    1.31 Sharpe Ratio
    -24% Max DD

    143 trades
    64% hitratio
    1.31 profitfactor
    0.00 ttest

    uses EOD data only.
    average trading length: 3 days.
    2 paramters, price data only

    I am worried about the MDD. It is not good enough for a stand alone.
    Same methodology works on sp500- and DJ- futures. But there is only little portfolioeffect between them.


    peace
     
  2. FinStat

    FinStat

    are you comfortable with the 1.31 profit factor?
     
  3. man

    man

    not really. but combined with the hit ratio of 64% it makes quite some positive expectancy IMO.


    peace
     
  4. Crude calculation (not including less than full stops):

    E = 0.64(1.31) - 0.36(1) = 0.48

    That's an excellent statistic... and not bad with a 24% DD...
     
  5. Foz

    Foz

    I can get the same "E statistic" with an unprofitable system that is profitable 78% of the time but has losing trades much bigger than winning trades and a profit factor of 0.9.

    E = 0.78(0.90) - 0.22(1) = 0.48

    So whatever you use E for I don't know, but it isn't a good judge of a system.

    If by 'E' you mean "expectancy", then... Expectancy is average winning trade x probability of winning - average losing trade x probability of losing. Because the average winning trade can be much larger or smaller than the average losing trade, the profit factor is not a good substitute.

    A 1.31 profit factor is low and this strategy has a high correlation with a traders non-trading equity investments. It's nice to see that the strategy made money the past couple bear years. If it isn't too time consuming and all transaction costs have been figured in, I'd trade it for fun while continuing further research.
     
  6. Are these results trading one lots?
     
  7. As another has said it is nice to see it made money the past few years but imo this profit factor is low. Most price based long/short systems have this profit factor and tend to rely on outliers for their profits. Good luck though.
     
  8. Just for kicks, here's a preliminary run of a short term end of day scalping method I'm working on. Note its for one measly contract of the mini, (don't you love the vendors that use 5 cars of the big boy as basis)

    1 possible trade a day (for now)
    no overnights
    this is an es test, but it seems to work on nq and ymm as well.

    some 'mild' optimization , system has only a few rules.

    what do you experienced systems guys think of the stats?
     
  9. Man when did the max DD occur?

    Let me guess you catch the falling knife with your system don't you?

    Let me guess further the max DD occur the first few trading days after 9-11?

    Have you tradet the first days after the has markets reopened?

    I thing regardless when the max DD happened - the results are good if slippage and comis are included.

    Stock777 look at this not as good as yours but go back 11 years ;)

    simple 5 day high low close data
     
    • sp.xls
      File size:
      75 KB
      Views:
      55
  10. man

    man

    Foz
    I do not know how an unprofitable system can have positive expectation value ... after having written that I just realise that this can be easily the case. I have never thought about it in that way, but what do think about this:

    assume a system that has these trades

    1 20%
    2 20%
    3 20%
    4 20%
    5 20%
    6 20%
    7 20%
    8 20%
    9 20%
    10 20%
    11 20%
    12 20%
    13 20%
    14 20%
    15 20%
    16 20%
    17 20%
    18 20%
    19 20%
    20 20%
    21 20%
    22 20%
    23 20%
    24 20%
    25 20%
    26 20%
    27 20%
    28 20%
    29 20%
    30 20%
    31 20%
    32 20%
    33 20%
    34 20%
    35 20%
    36 20%
    37 20%
    38 20%
    39 20%
    40 20%
    41 20%
    42 20%
    43 20%
    44 20%
    45 20%
    46 20%
    47 20%
    48 20%
    49 20%
    50 -1%
    51 -1%
    52 -1%
    53 -1%
    54 -1%
    55 -1%
    56 -1%
    57 -1%
    58 -1%
    59 -1%
    60 -100%

    hitratio
    82%

    profit factor
    2.0

    expectation value
    1.45

    anyways you end bancrupt.


    in the end there is no single figure that tells you how good a system is ... or did I go wrong somewhere?


    NYSE
    it is definitely not trend following.

    Candle
    I feel it could be something, though daytraders like yourself are usually bored with these numbers I guess.

    to you all
    thanks for posting. this kind of feedback is very helpful.


    peace
     
    #10     May 27, 2003