After market hours (10:pst) I'm sitting here toggling between streaming and not streaming.. and watching my "Net Profit" change 100%. What are you guys doing with dataset data in streaming vs. none streaming mode to affect the outcome of a 'strategy?'

(fwiw) This occurs when the bad-tick-filter is turned on or off. No streaming options selected. Such an obvious bug this has to be a non-intuitive feature... right?

(also fyi) This is for 2 years of INTC downloaded data from Fidelity (e.g. Scale, 10 mins & range 2 years...). Dataset is 1 minute data. Streaming vs. non-streaming backtesting nets the same 'net profit' from 1 day to 11 months. Between 11 and 12 months (or at 12 months) is when the difference starts occuring.

This brings up the big question the community should be asking themselves: Can wealth lab actually be trusted with backtested data longer than 11 months old?

QUOTE:
(e.g. Scale, 10 mins & range 2 years...). Dataset is 1 minute data.

I suspect a problem with the historical data. Create another DataSet with the native 10-minute INTC data (not 1-minute). Now, run the strategy with streaming turned off with 12 months loaded on both these datasets. Is there a difference in net profit?

Looks like the problem is intermittant so I don't have a 100% repro like I thought I had. I'll have to wait until the weekend to find a consistant repro. Is your suggestion to give me a work around or to help isolate the problem Eugene?

I *did* resync the dataset from the "Symbol Details" dialog for the symbol (e.g. download new dataset) but haven't created another dataset.

QUOTE:
Is your suggestion...

To help isolate the issue.

When backfilling a streaming chart, WL would load the data in the native bar scale (10-min) if it existed. The 1-min and 10-min data files could be different due to a backend error (like missing bars, spike in the historical data), explaining what you saw.