Say a game costs $16 million to develop, but then barely scrapes through, or fails QA, do you:
a) delay it and pump more $ into it in the hopes that it'll improve, increasing the risk and potential loss?
b) can it, losing all $16 million?
c) release it as is, hoping to minimise your loss?
A and C are really your only options (unless the game is so bad that it'll do so much brand damage that it'd be better off not even releasing... *cough* Dragon Age 2 *cough*).
Sometimes of course, you're also governed by set dates: financial calendars, competitors release dates, console launches and almost no matter how bad a game is it has to be out the door by then.
Shovelware is different of course; you release the best game you can on the lowest budget you can hoping to take advantage of the naïve.







