Can't repeat previous backtest results

[AnotherTrader]

Sometines, when I run a backtest, and then run it again with exactly the same parameters, I get different results between the runs.

Is this a memory caching problem, or the result of not clearing/resetting the database in some way between backtests?

[NinjaTrader_Bertrand]

Are you sure you're running then on exactly the same data set?

Please see if you get the same result when running consecutive same parameter backtests in disconnected state.

Thanks,

[AnotherTrader]

Yes, I am disconnected, and it's the same data as far as I am able to judge.

The only way to get the same results that I have found is to restart NT after rebooting the computer. Then the same parameters give the same results. Otherwise, the results will differ, in my experience.

I wondered whether there was some stack that wasn't getting re-set, or database temporary table that wasn't getting flushed, etc ...?

[NinjaTrader_Bertrand]

AnotherTrader, I was not able to reproduce it on my end here in testing - the same parameters for example on the SampleMACrossOver produced exactly the same results on consecutive runs for me in the SA. Are you using any custom resources in your code that are not properly released perhaps? Would you see the same outcome using the default SampleMACrossOver strategy in your testing?

Thanks,