Why most of the system metrics are useless numbers

Discussion in 'Strategy Building' started by traderum, Dec 19, 2009.

  1. Why most of the system metrics are useless numbers

    People ask me what my MDD is, what my Sharpe ratio is etc.
    All these metrics are nonsense! I'll give the proof of that below.
    The only meaningful metric is what the end result is, ie. the P/L at the end.

    Imagine you have a long position and it's down by 10%,
    ie. MDD is -10%. Most people will say that's a too big MDD... :)
    You could have done better if you apply this trick:
    If it is say 5% down then double your position --> net effect is your DD is now only half, ie. -2.5%, so you have a cushion...
    By repeatedly using this method you can set and control the MDD you want. So, we see MDD can be tricked out.
    In an old posting I showed that also Sharpe ratio is useless as it can simply be tricked out. Same goes with the Win/Loss ratio etc. All nonsense.

    So, dear managers and funders, forget all these silly buzzwords like MDD, Sharpe ratio, W/L ratio etc. etc. They tell nothing about real trading.

    Trading is strategy: the same things a general in a war does, ie. applying many strategic decisions at their right times... Trading is dynamic, not static.
    What counts is the overall strategy one uses and all the many single decisions on the way for winning the battle, and of course the end result,
    not the result in-between.
     
  2. You have discovered what most of us who trade automation uncovered. Statistics do not give you the ability to trade better they only act as a model. It is in your trading plan that contains your “pain index” where the real work is done. For example an automated strategy that produces a max of 4 consecutive losses in the stats is not where it’s at. The real stat is in your trading plan. That says you will not accept trades that wipe out more than 6% of your account in a month.

    P.S. I hate Sharpe ratios. My successful automated daily reversal system never has a shape over 0.6. Yet it keep chugg’in along making profits.
     
  3. pwrtrdr

    pwrtrdr

    All true, however.

    Lets sy you cut MDD down to 2.5. And your long. Market now goes to a super new low. Not only are you stopped your fired.

    So yes you can play with stats, but a keene eye can also see your total risk position as well.


    There is no easy answer other than never have all in at once approach, you need to leave money to "buy more" if it makes sense. I.E. the company for example was not accounting scam, or Credit Default Swaps gone wild are not causing such volatility that you cant handle current position never mind adding more !!

    It timing and luck when your "managing" money.... disciplined luck



     
  4. What you're saying is mathematically true. It is called martingaling. It is a method that is used by some in gambling as well - "Just keep doubling up as you win." The problem with that method is that you need an extremely large bank roll or the probability of blowing up when the market throws a curveball at you becomes huge.

    I know because I've seen many traders blow up because of it. I've been involved in performing due diligence on both FX and futures traders for institutions for years (as well as for retail managed account programs) and I much prefer traders who are more accurate in their methodologies.

    Good luck. You'll need it when you go live with real money.
     
  5. Danger66, I think the op understands the danger of averaging down. He/she is just trying to make the point that someone can fuck with the stats if they wanted but it'd be all be useless come game time. I personally fell in to the trap of focusing on getting pretty stats without making sure the system really worked. kinda like marrying a hot chick with A bitch of a personality.
     
  6. I agree with you. Focusing on stats before you trade is like focusing on the chicken before the egg. That won't work. Nevertheless, some stats must be used to "measure" a trader. Furthermore, a trader who tries to fudge or rig the stats by doubling up, etc. won't last for long. It's impossible to create a "successful" track record of a long enough duration by rigging things because of a shortage of talent or ability.
     
  7. You still are ?

    What kind of institutions ?
     
  8. Yes; still are. Brokerage firms, hedge funds, etc. and also retail managed account programs for traders that aren't quite ready to get accepted at the institutional level.
     
  9. Said the person trying to get others to sign up for his "making over 5% per day" and then using a demo account to try and prove it.

    Only a newbie paper trader makes ludicrous statements like this. (as well as his aliases and paper-trading friends).

    Every serious institution throws you out the door when you say stupid things like this.

    Real traders look for "Risk-Adjusted Reward." Newbies look for "1,000% per year with minimal risk." And then discover things called "Risk of Ruin" and "Blowout" or "Publish false claims on ET and expect others to take them seriously."
     
  10. I agree with you 110%.
     
    #10     Dec 19, 2009