-
…in reply to @allenholub
allenholub Regression tests really do have the best ROI because current bugs are a good predictor of future bugs. Tricky code stays tricky. Paradoxically, high density/coverage becomes evidence of low confidence code (AKA candidates for refactoring).