[gowld]

Are you saying that eyeball tests are wrong? That's an extreme claim.

[spekcular]

Eyeball tests are often misleading, or fail to detect weak correlations (or deviations from model assumptions such as heteroskedasticity). That's why we check with more formal methods.