Any changes in Backtesting function?

[billr]

I have recently migrated to v7.1.4. Also recently, I have noticed that the consistent results from a strategy I backtest often have changed significantly. I know it sounds a little nutty, but have there been any changes in the backtesting functionality or logic? This would normally be the last thing I would consider, but for quite a long time I have been using a strategy as a sort of control test for other strategies to make sure the strategy analyzer is returning consistent results. Lately, my formerly reliable control strategy is no longer backtesting anywhere near the same. All appears to be correct when I look at trades, charts etc, but the results are dramatically different. (the time period is consistent with the past tests).

Unfortunately, I don't have any saved backtest results to compare with the latest tests to compare or send to you. Any possibility that the newer versions of NT have tweaked the backtesting functionality or the way it handles data?

My backtesting is usually in the EUR/USD forex pair.

Thanks.

[NinjaTrader_Brett]

Hello,

Thanks for the forum post.

There where two changes that could have made some small changes. But should not be anything major like where your seeing. The only thins else that could make these differences is your data you're running on. For example not useing a correct session template or a number of days back are different. Also, who is your data feed provider?

Fixed
4434
Strategy Analyzer
Optimizer could have skipped the last iteration when using the default optimizer.

Fixed
4418
Strategy Analyzer
Changed logic for determining which session template the secondary bar series in a multi-series strategy will use to match that of running a multi-series strategy from a chart

[billr]

hello Brett,
Thanks for your response.

My data provider is Kinetick and the session template I use is 24/7 insteatd of <default instrument>. I usually backtest and live test using forex.

[NinjaTrader_Brett]

Hello,

I guess at this point I would need to have you watch it and see if you notice anything off with the trades being taken. As I dont see anything that should make a difference here and I'm not seeing any other tickets on this at this time. Without the results from .4 and settings from previous runs I really dont have any way to compare really.

You could give .4 a quick download and install and re run the strategy and see if you can duplicate the changed results. However you should not see any differences. Please also run the test with sample MA cross over as I did here on two PC's and didnt see anything off.

[billr]

Hey Brett,
I will do that. I am currently on .4 but still have a .2 on a different computer. I will see if I can reproduce the 'old' results and compare. I just wanted to eliminate the remote possiblity that the backtesting may have changed.
thanks.