Forex Software — Question about R-Squared

Re: Question about R-Squared

2019-01-18T08:57:14Z

Great work! Thank you!

Re: Question about R-Squared

2019-01-18T07:22:27Z

Thank you very much for your outstanding effort Mr. Popov. Will reload my instances and give them a try.

Best regards,

Re: Question about R-Squared

2019-01-18T04:49:34Z

I fixed some issues with the formula and it covers better some special cases now.

Here is an example of comparison of the EA Studio calculations and an online statistics calculator.

Re: Question about R-Squared

2019-01-17T22:04:42Z

Yes, I also think there is still something wrong with the R-Squared implementation in EA Studio. Especially strategies with very long duration trades seem to get a much higher R-Squared value, even if their curves look completely "bad" and the first one in your example surely should not have a higher R-Squared than the second one. I am sure the R-Squared formula itself in EA Studio is fine, but that there is a bug of how EA Studio counts / "forwards" the data points to the R-Squared formula.

Mr. Popov?

Question about R-Squared

2019-01-17T19:25:12Z

Dear Mr. Popov,

The addition of R-Squared is a great improvement. However I have noticed many times weird (non linear) balance lines with very high r-squared values.

The following image illustrates it very well. The below strategy is clearly more linear than the one above it and has a lower r-squared value

Is this a normal behavior? Is there a way to make EAS filter-out the strategies that have such stepped appearance?

Thank you in advance for your help.

Best regards,