Hello,
If I know the win rate, payoff ratio, and profit factor of a trading system, how do I set up a spreadsheet that can do Monte Carlo runs to estimate the likelihood of a drawdown of given % of capital or larger, for bet size of X% of capital? If this has been posted before, or there's an easily understandable website explaining it, then a link is fine, thanks.
How to do Monte Carlo
-
aaron
- Posts: 0
- Joined: Thu Jan 01, 2004 12:00 am
How to do Monte Carlo
You want a simulation, not Monte Carlo.
A spreadsheet is not the best tool for this, but it can be done.
The three parameters you specify have only two degrees of freedom, that is you can compute the third from any two.
The distribution of drawdowns will depend on much more than your parameters. It depends on how many trades you put on, how you adjust to losses, the time series behavior of the parameters and other factors.
Step one is to model the actual P&L. For example, suppose every day there is a 25% chance to enter into a new position and for each open position each day, there is a 5% chance you close it out. When you close out a trade you either lose L (set as X% of your capital at the time you put the position on) or win Payoff_Ratio * L, and you win with probability Win_Ratio. You could make all of these independent.
This particular example is not realistic, you should be able to do better. Even if you don't do better, however, this will give you some idea of the first-order influence of Payoff Ratio and Win Ratio to drawdowns.
Next you might put dates for the simulation period in rows and simulate the strategy returns in columns, using the spreadsheet random number generator. X will be a parameter. You compute the maximum drawdown in each column to get your simulated drawdown distribution.
A spreadsheet is not the best tool for this, but it can be done.
The three parameters you specify have only two degrees of freedom, that is you can compute the third from any two.
The distribution of drawdowns will depend on much more than your parameters. It depends on how many trades you put on, how you adjust to losses, the time series behavior of the parameters and other factors.
Step one is to model the actual P&L. For example, suppose every day there is a 25% chance to enter into a new position and for each open position each day, there is a 5% chance you close it out. When you close out a trade you either lose L (set as X% of your capital at the time you put the position on) or win Payoff_Ratio * L, and you win with probability Win_Ratio. You could make all of these independent.
This particular example is not realistic, you should be able to do better. Even if you don't do better, however, this will give you some idea of the first-order influence of Payoff Ratio and Win Ratio to drawdowns.
Next you might put dates for the simulation period in rows and simulate the strategy returns in columns, using the spreadsheet random number generator. X will be a parameter. You compute the maximum drawdown in each column to get your simulated drawdown distribution.