Monte Carlo for position sizing

Discussion in 'Trading' started by dotslashfuture, Jul 19, 2002.

  1. Is anybody out there using the Monte Carlo method for position sizing? After reading a lot, it seems that it is the best method in theory.

    I have also found a software product that will calculate the optimal fixed fraction of capital to risk based upon your trading method and upon your past trades.

    The guy wants $79 for a dos window program. Is there something else out there?
  2. m_c_a98


    try something called , I think that is the link. There is a free program there, I believe.
  3. I have done Monte Carlo simulations on Excel. I am a fully mechanical trader, and I used Excel to simulate 100 years of my trading, using my historical results. It was a lot of work, and resulted in about a 10MB excel file, but it was useful. One major caveat if you do this, or have someone do it for you: When doing a typical Monte Carlo analysis, you are assuming/applying some type of "hit rate", or win/loss % to come up with your results. The model then assumes that that rate is more or less constant over all of your trades. We know from experience that that is not true, i.e, if in a given year you average 50% winners, that 50% is a longer term average of shorter term "hot periods" and "cold periods" in your method. You may go 6-8 weeks with a 35% win rate, and later in the year, go 6-8 weeks with a 70% win rate. At the end of the year, it all averages out to 50%. HOWEVER....during your 35% "cold period" you are almost certainly going to experience a higher drawdown than a Monte Carlo model would predict, assuming a constant 50% win rate. Bottom line: In my experience, if you use Monte Carlo analysis to predict a drawdown for the purposes of position sizing, whatever drawdown your model predicts, DOUBLE IT before using it to calculate position size.
    Regarding the Optimal Fraction method, a.k.a, "optimal F", this model is VERY aggressive, and will give you BOTH huge profits and huge drawdowns. I risk 1/2% of my equity on each trade, and "optimal F" tells me I should be risking something insane, like 10%. No thanks. I'll take the tortoise over the hare every time when trading. Hope this helps.
  4. Yes, I would try it out. I have a high percentage system and my scheme is to when I lose I double up the next trade. My last QQQ trade failed. It was the other day on the monster rally and it triggered yesterday at 24.90. My deal was to double my share size to 5000 but I got chicken!!! Plus it was on the last bar of day.

    Please try it. I think there is something to it.