Backtest totally different (again...)

By custom session do you mean Limit time range from 8-22 ?

As I said, I tested it also with this setting and it matches 100%.

So I'd say the problem is probably in different settings between MT4 and SQ4.

Please make sure that you use the same data, you use real ticks in SQ4 and your trading options are set the same.

Can you perhaps attach also MT4 result .htm file, not just image?

You test it in MT4 with broker that has different time zone than Ducascopy.

It might not be a problem if you didn't limit trading to hours from 8-22.

But it seems your broker timezone is +2 hours from Dukascopy, so it means that 8 hours set in SQ4 mean 10 hours in your broker.

So when you test your strategy in MT4 you should modify the range from 10-24.

Otherwise you'll miss trades hat were opened between 8-10 and it could lead to totally different results.

Or try to run both test swithout limiting time range and check if the results are similar.

This strategy is not very good, it uses extreme values to look back when checking rising/falling conditions. This problem in strategy generation will be fixed in RC6.

> if you use dukascopy data to generate strategy you should use dukascopy broker to execute in live?

not necessarily, ideally the strategy should perform similarly with data from different brokers - that is one sign that it is robust.

But as you see if you start using some time conditions in your strategies you can run into problems with different timezones.