undefined | Better HN

0 pointsmaxbond1y ago0 comments

Re: how could they have handled this better, they could've caught this on the first or second day.

No one should need to tell you that an easily measurable and obviously critical metric is wrong. You should have alerting in place for that. A difference of 95% to 0% should light up like a Christmas tree.

If you release a game that's impossible to win, you didn't do a meaningful amount of testing or QA when it hit production. I assume a dev played at least one game as a sanity check, but they didn't sit there and play until they won. A QA person wouldn't sign off on something if they only observed 1 of the 2 states it should reach.

0 comments

utmostvanilla1y ago

I used to work on these sort of games and one would play a few rounds to get a feel that everything was working correctly and to see things like the transition to win states etc. Obviously, we wouldn’t be spending our own real money to do this: we would be playing against a dev server that produced the results. I find it hard to believe there would be zero QA before launch so I can only assume there was a mismatch between the dev and release versions of the backend.

JoshuaDavid1y ago

I wouldn't be surprised if it was something like, on the dev server

    SLOT_GAME_WIN_FRACTION=0.95

and then

    $win_fraction = (float) getenv('SLOT_GAME_WIN_FRACTION');

and then nobody bothered to set the environment variable on prod.

j / k navigate · click thread line to collapse