I feel like you'd want to create an index that's something like (number of features) + (number of unit tests)^2 + (number of adjacent integration tests)^3 And

I feel like you'd want to create an index that's something like (number of features) + (number of unit tests)^2 + (number of adjacent integration tests)^3 And maybe normalize for each by the lines of code for each component and some constant factor after twiddling. Basically LoC by itself is meaningless. When you gauge responsibility you want to look at how many features are owned, how complex those features are (unit tests) and how critical those features are to the system/product (integration tests). With complexity/criticality being weighted much higher than feature count. Of course like all metrics this is easy to game. But that's what code review is for.