negative r squared

I have fitted the line y=x which passes through origin to a bunch of data by using cftool. I got a negative r squared for the coefficient of determination. Could someone shed light on what that means? I know that y=x is not the best fit but I want to see how much is the determination coefficient if I force the data to be fitted to y=x.

Thanks.

Answer 1

[Oleg Komarov]

<http://www.minitab.com/en-IT/support/answers/answer.aspx?ID=1037&langType=2057>

Regressing through the origin doesn't give the R^2 you expect.