[statistics] linear regression

R2 is just a rough measure of reliability, but we can't say a regression with R2 = 0.999 must be better than a regression with R2 = 0.990. For example, the value of R2 could be misleading if there is outlier.

it is better to cacluation the SSE (sum of square of residual) and then use either t-test or f-test to evaluate the reliability of the regression. This method is called ANOVA (analysis of variance)

You could search lots of details of ANOVA from web.

I would like to ask,
how can we measure the reliability of a linear regression line?
Shall we use r^2 (where r is the sample correlation coefficient) to do so?
Thank you!