Would You Trade This System?

Discussion in 'Strategy Building' started by feng456, Sep 18, 2013.

  1. dom993

    dom993

    By design, any filter will improve the strat performance metrics, but it is difficult to assess the "quality" of a filter just from that angle.

    If anything, I am suggesting to look at filters performance separately from the resulting strat performance.

    WRT "smoothness" of equity curve, it relates quite directly to the information ratio, which is a nice & simple way of "measuring" that smoothness.
     
    #21     Sep 19, 2013
  2. Sergio77

    Sergio77

    A MC simulation is a terrible method of analyzing system risk due to the many assumptions that must be true for it to be valid.

    For example is there serial dependence in your trades? Are all the trade sequences equally probable? Why do you expect that the future will be i.i.d. with the past?

    What about the returns of the market when you are not holding positions? Does your MC simulation take those in account too? Or by MC you mean a randomization of system trades only? Well, this is too naive approach. If you happened to be out of the market during the flash crash but next time it happens your system (I hope not) trades against the move, the MC simulation results will be way off. In short, the system will be bankrupt.
     
    #22     Sep 21, 2013
  3. dom993

    dom993

    May-be, you could kindly suggest something better than MC simulations then - and highlight the pros & cons vs MC sim.

    That would be a useful post.
     
    #23     Sep 21, 2013
  4. feng456

    feng456


    I already ignored that guy. I scanned his history...all negative condescending posts about how everyone is wrong about everything.
     
    #24     Sep 21, 2013
  5. slumdog

    slumdog

    Profit factor and worst case drawdowns are my favourite metrics these days.

    pf>=2.0 and max drawdown of 15% is what i personally aim for.

    What is the profit factors for this system?
     
    #25     Sep 21, 2013
  6. Sergio77

    Sergio77

    I was trying to help. I understand that people with no formal background may find the statements outrageous. You must understand that before you can use MC you must know that all random sequences of trades are equally probable. This is not always the case and the results are meaningless. This is also the case with averaging returns and any other parameter.

    To me it is interesting that this forum did not ignore someone with a 5K account and a claim of a 80% expected drawdown. You may be better off flipping burgers in MacDonald's.
     
    #26     Sep 22, 2013
  7. Sergio77

    Sergio77

    Did I say I am a tutor? I don't think so. You must do the work yourself. You may start here with this seminal paper and good luck:

    http://www.cs.rpi.edu/~magdon/ps/journal/drawdown_RISK04.pdf
     
    #27     Sep 22, 2013
  8. That's insane dd. I wouldnt do more than 10% for dd
     
    #28     Sep 26, 2013
  9. feng456

    feng456

    In case anyone is interested I am done with this system. The performance this year has been shit. The DD IS terrible, it was only tolerated because the expected return was good. Not so this year.

    What kind of reasonable DD profit factor and other factors should I aim for?

    I really value your inputs.

    Thanks!
     
    #29     Sep 26, 2013
  10. #30     Sep 27, 2013