SQ 3.8.1 – Discrepancy between Retest and Optimizer ‘Original’ Results

Hi Mark

 

While processing this week’â„¢s databank of strategies I came across a discrepancy (bug) between the Retest strategy results of the attached strategy file and the ‘Original’ set of strategy results generated by Optimizer from the same file. Normally these results are identical, but in this instance they are not.

 

See attached screenshots:

 

Screenshot A – 20170218f_Strategy 18.2 – SQ Optimizer Before Run.png

This screenshot shows ‘Original’ row SQ ‘Optimizer’ reported results + equity curve for attached strategy file, as seen upon import into Optimizer from Retest. As expected, the results and equity curve are identical to that seen in SQ Retest.

 

Screenshot B – 20170218f_Strategy 18.2 – SQ Optimizer After Run PTP=0.5.png

This screenshot shows ‘Original’ row SQ ‘Optimizer’ reported results + equity curve for the same attached strategy file, but after Optimizer has been run. Note that the ‘Original’ row results have completely changed along with the equity curve. This is clearly incorrect, pointing to a bug in Optimizer.

 

Screenshot C – 20170218f_Strategy 18.2 – SQ Optimizer After Run PTP=0.png

Upon investigation, I noticed that SQ Setting Parameter ‘ProfitTrailingPips’, which is displayed as an ‘Integer = 0’, is actually saved in SQ as 0.5. Identified by copy/paste parameter values into Excel. If I change this parameter SQ to = 0, the results change but not back to the Retest values: See Screenshot C. Again, this is clearly incorrect, further pointing to a bug in Optimizer.

 

Screenshot D – 20170218f_Strategy 18.2 – SQ Optimizer After Run PTP=92.png

Upon further investigation, I identified that the correct value for ‘ProfitTrailingPips’ is 92, same as parameter ‘LongProfitTrailingPips’ & ‘ShortProfitTrailingPips’ (*). See Screenshot D, identical to Screenshot A. So at least I have a work-around to progress processing in Optimizer.

(*) Not sure why SQ shows 3x different parameters for ‘ProfitTrailingPips’ when Strategy Options was set for ‘ËœLong/Short Exit Rule symmetry’â„¢ (?).

 

Screenshot E – 20170218f_Strategy 18.2 – MT4 TDS.png

This screenshot shows results obtained from MT4 with TDS for MQ4 file generated from same strategy file. Clearly indicating that the MQ4 file generates strategy result in MT4 TDS very similar to SQ Retest (Screenshot A), rather than Optimizer (Screenshot B or C). Further pointing to bug in Optimizer.

 

While this is the 1st time I’ve encountered this particular issue, clearly there is a bug within Optimizer that is causing this discrepancy upon processing some strategy files.

 

Please can you investigate what has gone wrong and ensure that this Optimizer bug is fixed for SQ4.

 

Thanks

AC1962

Hello,

 

thanks for reporting this. I will check it and forward to Mark

I found the problem, it was not in Optimizer. The strategy for some reason had duplicate trading stops, one with value 92, another with value 0.5. 

This was probably messed up during genetic evolution, although I don’t know how it could have happened.

 

The duplicate value was not used in any tests and ignored also in export to source code, but as Optimizer goes through all the parameters it overwrote the value 92 with 0.5 that it found later.

 

I’m sending corrected strategy in the attachment.

 

Great, thanks Mark for the quick response & explanation.

Much appreciated!

 

AC1962 

Hi Mark

 

Similar to the SQ 3.8.1 bug reported topic above, please see attached another strategy file where the WFM ‘Original’ run results significantly differ from the Retest run. Upon investigation it appears that in Retest the parameter ‘ProfitTrailingPips’ is a Float = 0.6, whereas in WFM ‘ProfitTrailingPips’ is a Integer = 0. As you can see from the attached screenshots the ‘Original’ equity curve results are totally different in WFM compared to Retest. Using the strategy in MT4 TDS it is clear that the WFM ‘Original’ results are incorrect.

 

Is there anything you can do to correct this strategy file, and return it to me, so that parameter ‘ProfitTrailingPips’ is correctly recognised as a Float = 0.6 in WFM, and not an integer? Further, is there any chance of fixing this WFM bug with an up-grade patch to v3.8?

 

Thanks

AC1962

Hello,

 

thanks for reporting this. It has been forwarded to developers to make sure it will be fixed once the new version StrategyQuant is released