MonteCarlo Drawdown smaller than sample

[fercho]

Hi I've backtesting a Strategy for a mix of instruments that report 13000 trades in total.

Drawdown is 84.65%

In a Montecarlo with 100000 simulation, percentage DrawDown results only 35% at 0% percentage.

If Montecarlo simulation it's a trade-scrambling why I get a 1/3's DD at worst-scenario ?

Thanks in advance

[NinjaTrader_Bertrand]

fercho, thanks for the report, understand your point. Which instruments / timeframes are you working with here? Would this only occur for your MultiSeries script(s)?

[fercho]

Hi Bertrand, thanks for reply.

I'm running a 300 multi-instruments strategy.

Day data only.

At the moment only occurs with this config.

[NinjaTrader_Bertrand]

Thanks, will check into in more detail and get back to you.

[NinjaTrader_Bertrand]

fercho, we believe the root to be that 100k simulations on 13k trades is not allowing for enough permutations to happen to be meaningful. It's just an extremely high amount of trades to simulate MonteCarlo with, for example if you were to manually calculate the permutations for the trade amount in Excel > it would not be able to do it. So running a higher # of simulations to make the MC test meaningful isn't a good option here, we would rather lower the trades amount used.