As a software engineer you should surely know that in-house testing is pointless- coders know how it should work and aren't good at rigorous testing. User testing also falls down because they do miss things. What if their testers never finished it 100% and did parts of it? What happens if between testing and release a minor change affected the end timer and nobody realised they were linked. It's not uncommon for such things to happen when there are lots of coders working on a large scale code-base