Attention Support: Big non-sense issue in Strategy Analyzer

[pstrusi]

Due to incoherent backtest results under same set of variables, I did a detailed zoom in ( 1 tick data ) about trades, internal calculations...etc; what I've found was just shocking to me. It turned out that the Strategy Analyzer "changes" the original tick data making it different; of course with that weird data, the results are logically different. The curious thing is that this issue happens when I make bigger the backtest's timeframe. In this specific example, when I backtest from 2-7-2016 until 2-19-2016, results are perfect and according to genuine data; however the same backtest ( same set of variables ) but expanding the timeframe until 2-26-2016, the supposedly genuine tickdata is already changed. Obviously it doesn't have to do with the Strategy or Code, because the problem is that the data used for the backtest is changed !. Attached to this post, there are the two tickdata charts to show you the difference.

I can't emphasize more the terrible of this issue, it stops my work cause the backtest results are not reliable. I hope your developer team starts to review and debug the SA module.

IMPORTANT NOTE: I closed NT and start it over once again, now the issue is gone and the backtest didn't change the original tickdata. Can you explain to me why is this? How can I be sure when backtest are not distorted ?

I'm looking forward for your response

[NinjaTrader_PatrickH]

Hello pstrusi,

Thank you for your post.

Were you connected to your data feed provider? Who do you connect to for Forex data, both real-time and historical? When you restarted did you connect to your provider?

And I assume this happens with the SampleMaCrossover? Have you seen this occur before?

[pstrusi]

Hi Patrick, thanks for your response on Saturday. I'll answer below all your initial doubts. As I've post earlier, the non-sense backtest issue disappeared when I closed NT and start it over again, without any provider connection, just the NT platform; so obviously it had to do with some "data error in cache" that was clean out when restart it. It has nothing to do with data providers or Strategy programming. At this moment it 's not an urgent matter cause it was solved by a simple reset, but I'm aware that after a bunch of backtest it's a good rutine to "refresh" NT daily to avoid unreliable test.

Were you connected to your data feed provider?

No, I was not.

Who do you connect to for Forex data, both real-time and historical?

Interactivebrokers

When you restarted did you connect to your provider?

No, I didn't, just NT.

I assume this happens with the SampleMaCrossover?

I didn't a backtest at that moment the SampleMaCrossover, but in a eventual future, I will.

Have you seen this occur before?

No, I have not, ever.

Thanks for your attention