Hey Scorp - I've always thought the strategy tester had bugs - I've run things with optimization values that should make a HUGE difference, and the results were that it ran xx tests and there was no difference. I have problems believing that.
I also don't like this "accuracy of the model" idea. I want 100% accurate past values for pairs, even if it were only for the past few months. Is this so hard to gind? I want to run my EAs exactly as if they were on line for the time period specified, with exact results.
One small question about the strategy tester - is there a way to tell it not to close all positions at the end of the run? I sometimes take huge losses that aren't necessary because of closes of positions at the end of the test.
Is the new code generator ready for public consumption yet? I'd like to try some of my ideas in "two minutes." <grin>