Trading Lessons/Insights From Coin Flipping

Discussion in 'Risk Management' started by tradingjournals, Aug 31, 2010.

  1. Nine_Ender

    Nine_Ender

    Two things about this :

    1. If we were to get actual trading results from people
    condoning coin flipping you'd see they either trade rarely
    or trade badly. A rare person might luck out and make
    significant dollars short term.

    2. Martingale theory has been proven to be a dreadful strategy
    for gambling, actually bankrupting several people who
    believed in it ( eg Roulette, Heads Up Poker ).

    Something similar to Martingale concerning trading risks has been the recent trading calls from Fly Down. Basically he's been selling calls at certain S&P levels on the basis the number will not be reached by expiry. Twice he's barely missed getting cleaned out by the Index.

    So how do we measure the suitability of his trades. On one hand, he can claim he made money so he made good trades. But on the other hand, the payout was far too small to justify the Black Swan risk in the equation. The trades were terrible from a risk reward perspective and this ultimately is what matters.

    Martingale is the same idea, guaranteeing a small return by increasing your bets on the idea the positive result will come. The problem is you only have to catch one statistically improbable run and you lose far more then the system would ever pay out in the long run. And if you drain your capital, you don't get that next 50% bet to save your ass.
     
    #321     Dec 26, 2010
  2. Good post. Everyone should read it carefully. +1
     
    #322     Dec 27, 2010
  3. Roark

    Roark

    While the trader/gambler is running a Martingale strategy, the party accepting the trades/bets is running a reverse Martingale. In the long run, the party having the larger bankroll usually wins.

    <img src="http://www.elitetrader.com/vb/attachment.php?s=&postid=3045305">
     
    #323     Dec 27, 2010
  4. This is only true if there are only two players in a zero-sum game. It does not hold in general if the anti-martingale strategy goes against the trend eventually.
     
    #324     Dec 27, 2010
  5. Good point, so how does the market maker handle the antimartingle trending phase against you? Is it a solely accumulation phase for them or just distribution of what they have gathered from many clients during an earlier rangebound period/session before the breakout to you? Or they simply have unlimited firing power?
     
    #325     Dec 27, 2010
  6. Roark

    Roark

    They essentially have unlimited firing power relative to the trader/gambler. Casinos usually have a maximum bet-size limit, but the martingale gambler can take his bankroll and find a table with a higher limit and continue betting according to the martingale strategy if a long losing streak is encountered.
     
    #326     Dec 28, 2010
  7. Maybe someone can enlighten me on this paradox:

    I am betting on the event "the prob of having 2 consecutive H or T within 10 tosses is 0.86" with limited martingale of just 1level, which is different from the event "the prob of next toss will always be 0.5".i.e 10tosses=one event I bet VS 1toss=another event.
     
    #327     Dec 28, 2010
  8. It depends on your expectancy no. If I win 8 out of 10 times, then limited martingale to 1 level is ok. I.e loss = 2x($1+$2)=$6, win=$8
     
    #328     Dec 28, 2010
  9. with all due respect to whatever trading edge you possess, you are a horrible communicator, I have not even read beyond this point and fully agree with Walter that your description of the problem was very misleading. What I find even more disturbing is that you are condescending without even commenting on his repeated requests for clarification. I fully agree with him that if the first card is deliberately chosen and not randomly determined then. How much longer do you deliberately chose to keep the confusion in your communication going or is it random?

     
    #329     Dec 28, 2010
  10. lol, really? I guess you then found the holy grail. Thats the same as saying daily weather conditions are skewed (which they are) and if you figure out the difference between this skew and normally distributed weather conditions then you become the master forecaster. I do not disagree that a correct estimation of probability distributions implied by the market provide edge but what news did you now provide? I thought that was made clear on the first couple pages in Hull???

     
    #330     Dec 28, 2010