Back Test Results Change with parameter change

[561timw]

I am back testing a strategy using NT 7 running on windows 7 machine prior to the new service pack. I noticed that my back test results were different each time I ran the test even though I did not change any of the parameters. What would cause this? I tried repairing the data base, rebooting, opening a new strategy analyzer each time I run the test, etc... I continue to get the same strange results.

Help is appreciated.

[NinjaTrader_Brett]

Hello,

Thanks for your note.

Did you develop this strategy? Is there any randomization involved in the strategy?

Also, are you sure no parameters have changed. As long as the strategy does not vary its results the backtest will return the same values each time.

To double check this can you please test this on SampleMACrossOver strategy on your PC.

I look forward to assisting you further.

[561timw]

Yes I developed the strategy and have been running live for a little over a month with perfect NT performance. However, when back testing the results change with no paramater change. I only have two variables and I keep them constant when back testing. I run the test and then immediately rerun and find different results. When I run the SampleMACrossOver strategy I get the same results each time. So it appears to be something about my strategy. But I still don't know why everything works beautifuly live and sometimes the backtest matches up almost identical to my live trading results give or take a tick and then I rerun it and I get completely difference results. Any other ideas?

[NinjaTrader_Brett]

Hello,

No other idea's would just need to have you send me the strategy code to support at ninjatrader dot com ATTN: Brett and reference this forum post and I can isolate the cause of what in your strategy is causing this.

I look forward to assisting you further.