Topic: Randomly Close Positions Effect - Monte Carlo

Hi Guys - I want to share with a strange situation I had been coming across.

Many times that I would build a system which had a positive result over a period of time - something like:

Then I would put it into the Monte Carlo simulation as a cross check - and it would happily turn into:

So I really was confused why a good performing system on the same data, would drastically change. So I started investigating the different simulation options within the Monte Carlo simluator.   Randomly close positions has the following effect  (Picture in next post due to 2 link limit).

So I would like to find out:

1)  Is it bugged , or working correctly?
2) Do you use this part of the simulation?
3) I assume that running an EA on a VPS would have minimal of these kinds of errors.. what do you think?

Re: Randomly Close Positions Effect - Monte Carlo

Randomly close positions turned off effect:

Re: Randomly Close Positions Effect - Monte Carlo

Monte Carlo is designed to set conditions that will break your strategy performance. This principle is valid also for the "Randomly Close Positions" simulation.
When we close a position randomly we actually test if it is on a profit at a random moment. Imagine that a strategy has a very good entry logic and it open positions just at the beginning of  a trend and the positions go on a profit after the entry. Let's also this strategy has so good exit logic that it takes the profit on time and doesn't allow the positions to stay at a loosing state too long. The Randomly closing position in this case will not have big negative effect because the positions will be on profit most of the time.

On the other hand, if a strategy keeps the positions on a loosing state and only an effective exit rule close them on a slight profit, you can have a nice balance curve from the backtest. The randomly closing of  a position will have a devastating effect on such strategy.

Re: Randomly Close Positions Effect - Monte Carlo

Hi,

Probably your systems exit is very important or it can be curve fitted (overoptimized). I just had one strategy with similar situation with randomly closing positions. In my opinion it does not mean that your strategy is shit. With my strategy example I was using a lot bigger Stop Loss than take profit, but also I used custom indicator to trigger exit logic before that SL is hit. So randomly closing positions through montecarlo test is not correct, because correct logic is not applied To put it in other words exit had big edge in this strategy and montecarlo just ruins it.

Have a nice day

Re: Randomly Close Positions Effect - Monte Carlo

In my opinion it does not mean that your strategy is shit.

Had to LOL at that.

Re: Randomly Close Positions Effect - Monte Carlo

Thats "why" the minutest details matter in quantitative trade systems

@DoCZero
Your doubt is very useful thank you for point it out