I would go farther and claim that 100% test coverage is worse than 90% test coverage. You're probably testing irrelevant stuff - which just means that you have

I would go farther and claim that 100% test coverage is worse than 90% test coverage. You're probably testing irrelevant stuff - which just means that you have more test code to maintain, and also likely means that your tests are bad (i.e. not testing business logic, but testing irrelevant implementation details). Last, but not least, 100% test coverage means that your code is bad/ you're not programming defensively (how exactly did you cover all those error-handling bits for cases of unexpected exceptions/ "stuff that theoretically should never happen, but in practice it might, so let's have some sensible treatment/ logging/ etc."? )