Long-time list member Marek R. wrote in to ask about Monte Carlo simulations (name used with permission).
Marek R.:
I’ve recently discovered Monte Carlo simulations. How do you approach this? Do you use this kind of random order of the trades to confirm the edge and avoid curve fitting to only the historical data sequence?
Dave:
I think Monte Carlo simulations are a useful tool to have in your toolbox.
I don’t use them frequently, but occasionally I’ll use them to visualize different ways a strategy might have played out.
The idea is you take your backtest, which is a list of trades in order by entry time, and reorder them randomly.
You can do this multiple times and then compare the resulting equity curves against each other.
The ordering of your trades is somewhat random, so when you reorder them, you are simulating plausible scenarios that could have occurred, but didn’t happen just by chance.
You can better visualize what could have happened, and therefore what sort of variance you can expect going forward.
Is this useful? Yes.
Is it a silver bullet that you can use to improve your strategy or prevent curve fitting? Probably not.
But it’s certainly worth knowing about and having in your bag of tricks.
Good question, Marek – thank you!
-Dave
