Assess my strategy, please

Discussion in 'Strategy Building' started by doublewobble, Aug 6, 2021.

  1. Thanks for your thoughts. Yes, my data starts in the April 1st 2004.

    RedDuke, you made a valid point about the bull market. Admittedly, the performance of the recent years is the best. However, for the 5 year period from the April 1st 2004 to the April 1st 2009, during which the SP500 fell from 1126 to 790, the algorithm performs as following:

    Initial capital 100000.00
    Ending capital 171549.68
    Exposure % 20.35%
    Annual Return % 11.39%
    All trades 1109
    Winners 626 (56.45 %)
    Losers 483 (43.55 %)
    Max. system % drawdown -3.41%
    Sharpe Ratio of trades 1.59

    Not stellar but also not too bad IMHO.

    Daniel.a, I am sorry for the stupid question, but what is the average trd%...?

    The algorithm trades hourly. The Open prices mentioned above are the 1 hour bar Open prices.
     
    Last edited: Aug 6, 2021
    #11     Aug 6, 2021
  2. Daniel.a

    Daniel.a

    Myself i am not that worried as most will tell you that (yes it has been working just because its been the words best bull market etc) if it works now and outperform, it works, does it not ?, then squeeze out what you can from it until it does not work anymore, its not more difficult than that, just keep track of your metrics and when it goes above or below that you take notice and shut it down if need be, does it need to be more difficult than that ?, now i am reasoning from the point that this is one out of many strategies you run. If you shall drop every systems that works great now (just because it does not work great in bear markets or that you dont have data for bear markets etc) then you can just as well wait forever to trade anything. There is no single system that works at all times. If you have made everything you can to validate your system, then trade it, first paper it, then if metric and execution matches then switch it on gradually. Avr trd, iis what you find under All trades, section in AB, Avg Profit/loss % .. this is your margin
     
    #12     Aug 6, 2021
  3. Daniel.a

    Daniel.a

    From my own experience, its not so much strategy breakdowns that causes most issues, its poor execution of the strategies that you chose to run. This causes a lot of drag and can eat up a lot of your Avr trd %.
     
    #13     Aug 6, 2021
    comagnum likes this.
  4. SunTrader

    SunTrader

    Once again, I'm probably sounding like a broken record to some folks, but using data bars higher than individual ticks even if tick data itself is being used to plot them (i.e 60 minute bars in this case), assumptions are made where trades might take place.

    Allowing for slippage of a penny or two is nowhere sufficient in most cases.

    Even then there is nothing like forward "testing" with money on the line to confirm what you got.
     
    #14     Aug 6, 2021
  5. @Daniel.a
    This is my first strategy that kind of works and for the time being, the only one. However, I fully understand the reason why one should have many of them. I have just not had enough time to create more. A short only strategy following the same train of thought will be the next one.

    That's exactly what I am doing here, asking about validation of my strategy. I sincerely hope that I have not missed anything truly obvious.

    Avg. Profit/Loss % 0.46%
     
    Last edited: Aug 6, 2021
    #15     Aug 6, 2021
  6. @SunTrader
    Could you advise how to include slippage better? As short as I can understand, the slippage is a function of traded volume. Is the 1% rule not sufficient?

    My initial capital is fairly low, 100 000 $ (close to reality).
     
    Last edited: Aug 6, 2021
    #16     Aug 6, 2021
  7. Daniel.a

    Daniel.a

    But thats not that bad for an intraday strategy, so thats your margin for error in your assumptions. If your strategy buying breakouts etc your slippage will be higher than for example Mean reversion types, Anyway, test it, connect it to your paper trading account and check if your assumptions are correct, in terms of your metrics.. swap your 0.005 commission to something higher to provide some room for errors and drag etc.
     
    #17     Aug 6, 2021
  8. @RedDuke

    I appreciate your comment. Please refer to the performance for 2004-2009 in one of my earlier posts and feel free to express your opinion.

    I just found out that @ works at this forum.

    Thanks in advance, etc.
     
    Last edited: Aug 6, 2021
    #18     Aug 6, 2021
  9. @Daniel.a

    Changed commission to 0.01.

    The results:

    Initial capital 100000.00
    Ending capital 2754165.87
    Annual Return % 21.21%
    Avg. Profit/Loss % 0.42%
    Max. system % drawdown -7.31%
    Sharpe Ratio of trades 2.00

    What value do you recommend for the commission?
     
    #19     Aug 6, 2021
  10. Daniel.a

    Daniel.a

    Why dont you just start to test it? papertrade it, then live, check your real slippage compared to your assumptions.. test slippage as 0.1 perhaps instead of 0.01.. you are going to get 324234 million opinions in here anyway. Start and test it small if you believe you have something. Good luck
     
    #20     Aug 6, 2021