> I never claim that 100% coverage has anything to do with code breaking.

But what I care about is code breaking (or rather, it not breaking). I'd rather put effort into ensuring my test suite does provide a useful benefit in that regard, rather than measure an arbitrary target which is not a good measure of that.