That's all very nice, but I would be wary of nice smooth equity curves...
I had a case of the 'too good to be trues' with my GA, it turned out that it was the order management options that I was feeding the GA. Anything that is 'tight' in terms of stops or take profits will be exploited by the GA to give erroneous results, only with tick data can one test these types of things. With 1 min data the back-test and forward test would diverge chaotically, one only really finds this stuff out by forward testing with small size.
I had a case of the 'too good to be trues' with my GA, it turned out that it was the order management options that I was feeding the GA. Anything that is 'tight' in terms of stops or take profits will be exploited by the GA to give erroneous results, only with tick data can one test these types of things. With 1 min data the back-test and forward test would diverge chaotically, one only really finds this stuff out by forward testing with small size.
The breaking of a wave cannot explain the whole sea.