Monte Carlo results when using scale outs

[NickyD]

I've created a strategy that enters the market via market order with quantity = 3. That position then gets managed with three separate stops and three separate targets for scaling out at different levels.

This seems to be causing problems with the monte carlo simulations for things like Max Drawdown, Max Consecutive Losers, etc . The max drawdown that actually occurred during any given backtest is always in the 95% percentile or greater (bad) outcomes from a monte carlo simulation of that backtest. I take this to mean that the monte carlo is working with the trade exits as individual events instead of the trade entries. And thus is skewing the results more favorable than reality.

What can I do about this? I'm providing a signal name for all the entries, I'm not sure what else to do.

Thanks

[NinjaTrader_Jesse]

Hello NickyD,

When you scale in or out with the managed approach you do need to use individual entries with unique signal names so you can tie part of the exit to a specific entry, that will treat those as 3 separate trades. The monte carlo simulation deals with trades which are paired entry plus exit so that sounds like it is dealing with that use case correctly. What result were you expecting specifically, to have a single entry with 3 exits combined into an average fill for the exits?

[NickyD]

I'm using unmanaged orders.

I believe what I'm asking is that I would like it to be aware that the position came from a single named entry execution, regardless of how many exits existed. So, yes, I think your proposal is a solution to that.

Here's a sample of my activity if it clarifies - first pic is the "trades" tab from the strategy analyzer. The second pic is the executions tab of the strategy analyzer. I entered a position short @ 4/18 8:57:18 AM, that position had 3 exits. I entered position long @ 4/19 1:00:29 PM, that position had 3 exits. It's 2 entry decisions driving the initiation of 2 positions.

The monte carlo results get distorted if you don't consider this 2 trades, instead of 6...?

[NinjaTrader_Jesse]

Hello NickyD,

In that situation you still have 3 individual trades because you have 3 separate exits even though they are tied to a single entry. For the trade performance to calculate it requires a Trade which is defined as a paired entry + an exit, to combine them into 1 result would require using 1 entry of 3 quantity and 1 exit of 3 quantity. It will otherwise display in that way because you are exiting 3 times for that entry which generated 3 total trades. There would not be a way to have the performance combined differently in that situation because each exit needs an associated entry to be considered a trade and part of the performance calculations.

[NickyD]

I'm not sure I fully appreciated how much this affects almost all the performance metrics. The takeaway here for others reading this is that if you're using scales in your trading, or you flatten a variable number of contracts at session close, or anything else that causes a different filled Qty on your entry and exit, you need to export your back test data, aggregate it, and do your own analysis on it outside of NT. Most everything presented to you on the strategy analyzer besides is skewed and not easily adjusted when the quantities vary.