Quote Originally Posted by Shibi View Post
Yeah they tested it, but the test environment wasn't close enough to a real server to be a valid test.

A bit like taking a driving test on an empty city street, the results won't be the same as if it's peak hour traffic.
And that's the problem, right there. They tested it with ONE person only bidding in ONE lottery.
Not really a test of anything more than:
'can I bid?'
'yes'
'ok, it works'
...
There was no full phases test.
No beginning to end test.
Because even if there had only been ONE bidder and they ran the actual lottery, even if they shortened the times for the tests, they'd have run afoul of the bug and known about it beforehand.
I'd say it is obvious they didn't do that but the possibility still exists that the bug was introduced by their emergency patch that changed the parameters of who could and who could not bid in lotteries that they released on day 2 of the lottery WHILE it was in progress.
Programming 101 says you don't CHANGE parameters of a program while it's running.