Monte Carlo simulations best way to project drawdowns/returns ?
- Thread starter Saxio
- Start date
If you have only the outcome of 100 trades, you can not find the best way to project future drawdowns/returns. You need other information like, what kind of system is used, what is the market doing at the time those trades are made, who traded the system, and so on. Only go to Monte Carlo simulations if you know your system inside out and how it performs in any kind of market. And then you need to acknowledge that the worst and the best is in the future of your system and has not happened yet in your data set.
100 trades is too small a sample. 1000 is better. I like to use several thousand.
For returns you can take the standard deviation. And see where 2 or 3 standard deviations put you. Like if you make 50% after 100 trades and your standard deviation is 25%, then after 100 trades your returns could be anywhere from -25% to +175%
I am not talking about trades from back testing but actual trades taken in the market
Agree. There are different variations of markets like consolidation, trend,reversals. 100 trades is not large enough
It is not a system , it is discretionary trading , 100 trades in a live environnemt
For example Darwinex that is looking for talented traders to trade their capital and that of investors , calculates the statistical significance of your trades and for that you don’t need 1000 of trades .
For example , if you can buy 50 days in a row the daily low and sell at the daily high , imagine how statistical significant that would be, on the other hand suppose another sample with 1000 trades in the same 50 days that barely turns a profit , which trader would you give your money ?
Last edited:
It is not a system , it is discretionary trading , 100 trades in a live environnemt
For example Darwinex that is looking for talented traders to trade their capital and that of investors , calculates the statistical significance of your trades and for that you don’t need 1000 of trades .
For example , if you can buy 50 days in a row the daily low and sell at the daily high , imagine how statistical significant that would be, on the other hand suppose another sample with 1000 trades in the same 50 days that barely turns a profit , which trader would you give your money ?
Ok, I know by experience that you do not need to backtest or even do Monte Carlo simulations if you know your stuff. Then 100 trades are enough to make some calculation what you can expect. If you are at that level, I would only suggest that you track past trades and use position size that lets you sleep at night. I would not try to optimize it. I would only look at what is comfortable to do it, day in day out, and under trade. Your goal should be optimal stress-free trading.