Metrics and risk management

Discussion in 'Risk Management' started by harrytrader, Feb 14, 2004.

  1. In the thread "The Most Annoying Thing About Trading "
    by Flashboy he says:

    http://www.elitetrader.com/vb/showthread.php?s=&threadid=28422

    "The Most Annoying Thing About Trading
    When you hold onto something for a bigger profit it comes back down and stops you out.

    If you bail early to protect your profit it winds up going in your direction to your entry point..

    This happens over and over.. I know many of you have experienced and still experience the same thing.. guess you can't pick and choose which ones to hold..

    Just releasing some frustration.. can't trade right now.. too emotional."

    That's why precision and so prediction counts. That's the reason, as I am putting my system under "state control" (I use the term of Shewart the father of statistical quality control) I need to define precise metrics.

    For example let take the break zone of last friday:
    Normally my model had a breakdown zone at 10678.4 theorically see picture below (MnPFS1 acronym for Min Projection on Future Scale 1 - I remind all the models charts are for spot even when calculated for future). What I will mesure is the highest pullback after the market broke this level, and also the garanteed profit zone for a chicken trader that would cut his profit too short :D (green vertical line zone). As a benchmark this can be compared to the break of the low of the day (LOD): you can clearly see (on picture 2) that the noise was worst. Once I will have enough datas I will show them on a beautiful Shewart Control Chart which will aim to predict the maximum risk for this kind of break.

    <IMG SRC=http://www.elitetrader.com/vb/attachment.php?s=&postid=431251>