TakeProfit difference between 1.20 and 1.22

I took workshop 4_1 script and tested on 1.20 and 1.22.
Same script, same data, results are identical.
Now add just one line:

TakeProfit=4*ATR(100);

This changes everything: 1.20 is in big profit, 1.22 is in big loss?

I don't think that it matters much for your results if you're using Zorro 1.20 or 1.22. But the same script can produce very different performance when something else is different, such as the test periods, or account parameters.

When you don't know why system A produces a different result than system B, just compare the trades in the log - that's the easiest way to find the reason.

Well,
entry times and prices are the same in both cases,
exit times are the same,
exit prices are different.
looks like 1.2 is exiting at the close of the bar and 1.22 is exiting at the open of the bar.

1.2:

Short EUR/USD 2564112 1 25.03.13 13:00 25.03.13 17:00 1.29394 1.28396 7.5441 0 Target
Long EUR/USD 2583413 1 04.04.13 17:00 04.04.13 17:00 1.28779 1.2943 4.9497 0 Target

1.22:

Short EUR/USD 2564112 1 25.03.13 13:00 25.03.13 17:00 1.29394 1.28519 6.6086 0 Target
Long EUR/USD 2583413 1 04.04.13 17:00 04.04.13 17:00 1.28779 1.28755 -0.18238 0 Target

Yes, it's the exit price. The trade is not exited at the open or close. Hitting a stop or profit target normally happens inside the bar. If both are triggered in the same bar, the exit price can't be determined, but must be estimated. I suspect that this the reason of the difference here.

AFAIK Zorro 1.22 uses a more pessimistic algorithm for estimating intrabar exit prices, with a larger weight on the begin of the bar - that seems to cause the differences in this case. When the performance depends heavily on profit targets or stops, set the TICKS flag. It gives more precise estimates for intrabar simulation.

ok,
so with TICKS set:
1.20 result is about the same AR 59%.
1.22 much better AR 15% (was -24%).

The 15% is the most realistic of the three results here, i.e. the return of the highest likeliness in real trading. Even with TICKS set, when stop and target are hit in the same 1-minute bar, the inaccuracy is equal to the high-low difference of that bar.