Wow. I didn’t realize it had been a month since I wrote on here. While this Alpha has been the best one, as far as hours, I guess it still has preoccupied me. Our scheduled Alpha date was March 7 and the team did a great job pulling all their tasks together in the last sprint. The goal was to hit Alpha on time or at the very least come in better than last year, which was 7 days late.
In order to hit this goal, the idea was to leave the last sprint (3 weeks) as a “contingency” sprint. While this sounded good on paper, it was hard to know when to use the “contingency”. It seems like we used it for anything and everything and in the end, it appeared that we were scheduled right up to the end of the last sprint. The reason it got this way, was there was always “one more thing” that needed to get in or that was half done and it would be a shame to release a feature that wasn’t 100% complete. In the end we have to ship a good game with awesome features. I don’t think the concept of a “contingency” sprint works very well because you end up using it like a sprint, scheduling right up to the Alpha date. Next year I think what I’m going to do is have a two week period before Alpha which is completely unscheduled. It will only be used for finishing something that didn’t get completed in the last sprint and fixing any issues with the game which might be hindering QA’s ability to test the whole game.
Despite having that last sprint completely scheduled, the team did a great job leaving that last week of the sprint available to verify all the features were in a good state. We ended up being 4 days late for Alpha, being declared on March 11th. There were a few crashes in one of the game modes which would hold up QA from testing. EA has a checklist of things that need to be complete before a game is declared Alpha. Stability of all game modes is one of those. Even though we had done a great job on the stability of our game, there were a few timing specific bugs that held it up. Friday the last fix went in QA declared us Alpha.