System Metrics Evaluation

Discussion in 'Strategy Building' started by Maverick1, Jul 7, 2016.

  1. People need to understand that only metrics based on real performance count. One can easily cheat both IS and OOS performance with repeated backtests. Those links were pretty good info.
     
    #11     Jul 11, 2016
  2. Sergio77

    Sergio77

    I agree with point about actual performance.
     
    #12     Jul 11, 2016
  3. i think a system drawdown is the most important factor .and capability of generating profit is the "least" important item . consistency & a smooth equity curve is much important than big profits.why ? because big profits need big risks and huge drawdowns .when you stop losing money very fast chance of increasing your account goes up .
     
    #13     Jul 11, 2016
  4. Maverick1

    Maverick1


    I would agree with that to an extent. There's no point talking profit when the risk of ruin is high.

    So all in, I would take it that there's no such thing as a profit factor significantly higher than 2 on a system with a very large sample of trades (high hundreds to 1000s)...

    If I recall in a Jim Simons interview he talked about his edges being thin and needing to stack them up, and given that he's more than likely not a trend follower for the most part, that can't possibly imply a large profit factor?
     
    #14     Jul 11, 2016
  5. some of seasoned traders use counter trend setups because they risk is low an also their profit is low. but they maintain their consistency .because lack of big profit means lack of big risks and lack of drawdown
     
    #15     Jul 11, 2016
  6. Maverick1

    Maverick1

    makes sense CME_trader, I can see how that approach would help them 'fly under the radar' so to speak. It reminds me of a comment someone made on an old ET thread where they talked about deliberately trading/going for the system with a profit factor of 1.4 because it traded a LOT and would not draw attention
     
    #16     Jul 11, 2016
    CME_Trader likes this.