There were many reasons why the game tests were failing -

• Not using one single bot. This caused problems since different bots were keeping track of different scores across different tests... Scrap that, one bot is used across the entire session.
• animation_file_id actually changes every time the game score is edited, but the tests were checking if they were the same...
• @flaky on game tests were useless or actually made it more unstable as the same scores were being used, raising the error again.

Despite fixing this, there is still a chance that this can fail due to race conditions (many jobs running at once), so an xfail was added.

Edit: Also removed an unnecessary game test.

Onto the next test...

changes look good to me