"We wrote the first scoring algorithm at Goalee based on the red-green light. Within a couple of weeks we made our first algorithmic change, and made several quick fix releases to update the scoring methods in the following weeks. By the end of the month, a whole section of our testing suite was almost completely worthless. In a world where cash is king, I wouldn’t mind having those dollars back."

What I don't understand about comments like this is that a whole section of your code, both runtime / deliverable code and test code had become worthless. But, you only seem to view the discarded test code as wasted effort. Either the tests have value or they don't. And, if you write tests, and then discard the code they test, you'll likely also discard the tests. But, that doesn't change whether or not the tests had value, nor whether the new tests that you'll write for the new code have value.