Money Management: What's Better?

Discussion in 'Risk Management' started by Corso482, Dec 13, 2002.

  1. What's better? Using a fixed dollar amount of your starting capital per trade that stays the same no matter how the account fluctuates, then possibly adjusting that amount at year/month end etc., or using a percentage of capital that will constantly change as your account grows or shrinks.

    I have a feeling that a constantly changing percentage would be better, yet I suck at math and I wonder if I'm missing some complicated nuance of the two choices.
  2. cheeks


    This is the way I prefer to do it.
  3. Are you talking about back testing or real trading?
    You shoul use the percentage methods as described below for your backtesting. When real trading starts you might switch to Dollar and adust it from time to time (in relation to your overall capital).

    Are you talking about stocks or futures?
    If you talk about futures I would use the "Stop Based Risk %" method, by using this you always risk a certain percentage of your capital. On stocks a certain percentage of your capital will do. There are more complex methods that you want to look at that take loosing/winning streaks, drawdown a.s.o. of the individual symbol or the portfolio in consideration.

    Are you talking about one market or many markets?
    Also important, but use percentage is better, and always look at the draw down in percentage!

  4. Wow, I didn't know my question was so incomplete...I was talking about stocks. And why would backtesting be any different from real money? I'm also talking about one market, the U.S. stock market.

    You're right, you could take into consideration winning streaks, drawdown, etc. to optimize a strategy, but I was just speaking about in general.
  5. Well, so I asume you were talking about many US stocks. The difference between backtesting and real trading is that when backtesting and you start out with US$100.000 and use 5% per each signal from your over all capital you might be quickly going up to US$200.000 then you trade US$10.000 instead of US$5.000 per trade. By doing this you will keep your risk stable thru the testing period. When finally doing real time trading, you might calculate the percentage on a month to month base, which should be good enough unless you use some optimal F money management rules. Important is that you test as you would trade in regards of the stocks/symbols portfolio and the MM rule has to be applied to the portfolio.

  6. fsoft22



    I use Kelly, but it need some modifications in the risk.
  7. Fixed dollar when you are starting the system. If you have to decrease size in a drawdown it is too hard to dig your way out.

    If the system takes off and things go your way, then switch to the percentage method.
  8. man


    there is only one answer, unfortunately many daytraders refuse to accept it, namely percentage.

    consider this:
    you start trading and you decide to choose a fixed amount. this will inevitably be a pure function directly derived from your capital. Let us assume you have 100.000,- and you decide to risk 5.000,- on any trade. risky IMO but somehow reasonable.
    now consider you started with only 25.000,- in capital. you most likely would feel that 5.000 as a bet size is a little much. in essence this is the whole argument about the percentage issue. you must assume that you start everyday new, ignoring where you stood a week ago.

    If you do not change your bet amount with your account size you are in essence constantly changing your actual risk. rule number one is: "ensure that you can stay at the table". you violate this rule if you change your relative risk. if you still bet the same amount with only half the capital you double your risk in terms of probability to quit. if this probability is still bearable then there is only one conclusion: your bet size was too small for the original capital. if this probability is too high than you must not do it.

    there are only two good reasons for changing the bet size:
    1. streaks in your equity curve. if there is significant probability that you win after you lost than it makes sense to change your relative bet size = risk the same absolute amount as before.
    2. your account is so tight relative to absolute costs of trading, that you cannot afford to recude bet size, since relative trading costs would explode. in this case I would strongly reconsider trading.

    In all other cases a fixed bet size is not justified by rational arguments. It is just - comfortable for some. And of course the smaller the changes in your account, the smaller the effect.

    profitseer: with all respect, your argument with not being able to get out of a draw down spells: desaster. it is just a function of time until things blow out. signifcant draw down means: hard way to get back. increasing the relative bet size, without rationale argument, means gambling.

  9. Bob111


  10. If I were a young speculator with $50,000 to trade. I would take an initial position of no more than 10% of the total-$5000-and set exit points to limit potential losses to 10 to 20% of that-a $500 to $1000 loss.

    In other words, I would set it up so that my losses were no more than 1 to 2% of the total risk capital.

    Upon losing $1000 in the first trade, I would scale back my next opening position to $4000 and limit my losses to somewhere in the $400 to $800 range. And so forth.

    On the upside, if I made $2000 on my first trade, I would bank $1000 and increase the opening size of my next position to $6000, in effect reducing my initial capital at risk ($5000) by 20%, while increasing my actual risk capital by the same amount.

    That way, even if I lost on my next trade, I would still be up money for the period. Assuming that I was right in my market calls 50% of the time.

    I would make a lot of money employing this strategy. And I would make a respectable living being right on only 1 out of 3 trades, provided I maintained a risk/reward ratio of, at most, 1:3.

    In other words, if you pick opportunities so that the probable reward is at least three times greater than the objectively measurable potential loss, you will make profits consistently over time.

    Trader Vic-Methods of a Wall Street Master
    #10     Dec 13, 2002