Interesting issue with optimizer

[ScottB]

Hi,

This may have been discussed somewhere but I have not found any reference to it.

The genetic optimizer appears to rerun at the end of the optimization whichever iteration proves to be the best. I can print out the iterations from the genetic optimizer (in the Type folder) and after all the iterations are complete, I get one extra strategy execution with the parameter set from the winning iteration.

Is there a reason for this or is it "just the way it works" or am I misinterpreting the results?

I noticed this because I am writing the trade results to a SQL Server database in the genetic optimizer and kept getting duplicate key errors at the very end of the optimization run.

Best Regards,
Scott

[NinjaTrader_RyanM1]

Hi Scott,

Thanks for the post. I was not able to reproduce this. Attached script, output, and screenshot. The winning result of 7, 25 is only outputted once. Can you think of any changes needed in this test where we could see this occur on our side?

[ScottB]

Ryan, thanks for getting back to me. I didn't mean it was in the output window (the one you showed in the screenshot) but rather the code executes a second time.

If you put a print statement in OnStartUp you will see it executes once per iteration and then once again at the end of the optimization.

Best Regards,
Scott

[NinjaTrader_RyanM1]

I moved the Print statement to OnStartup() but still get the same results. Winning result 7, 25 is only shown once.

[koganam]

Hi,

This may have been discussed somewhere but I have not found any reference to it.

The genetic optimizer appears to rerun at the end of the optimization whichever iteration proves to be the best. I can print out the iterations from the genetic optimizer (in the Type folder) and after all the iterations are complete, I get one extra strategy execution with the parameter set from the winning iteration.

Is there a reason for this or is it "just the way it works" or am I misinterpreting the results?

I noticed this because I am writing the trade results to a SQL Server database in the genetic optimizer and kept getting duplicate key errors at the very end of the optimization run.

Best Regards,
Scott

I do not think that it reruns the best result; it just reports it as a summary. As that summary report is just a duplicate of the actual test run report, it would probably flummox anything that is recording all output, if the records are required to be unique.

[ScottB]

koganam, you may well be correct. I know OnStartUp executed again and the error came out of a method only called in OnBarUpdate but as always, I may have it wrong. Under any circumstances, I was able to fix the issue it was causing with the database.

I am always loath to fix an issue that may be masking a more subtle yet more serious issue.

Best Regards,
Scott