I'm not a fan of this (and I probably already posted about it in the past). Here's why: We had many complex tests at $job-2 and working on a ticket usually meant spending more time fixing the tests than implementing the changes themselves. Some tests used a framework where counting the tests wasn't even possible - and several times, we found out too late that a change caused some of the tests being skipped, but if they hadn't been skipped, the would have failed. Therefore, I prefer counting the tests, even if I need to fix the number often: non-running failing tests are worse than missing tests, because they give you a false sense of security.