You can't plan for unknown unknowns.

If you develop a feature and it turns out the feature is not great, you can choose to ship a game that's not great, or fix it.

Once you have committed to a ship date and moving the date is no longer an option, fixing it means working over time.

These crunches often come from the team themselves who are identifying that things aren't as good as they could be, want to fix it, but don't have the power to push dates back.

Update: One problem is that creative teams are rarely satisfied, and would love to polish things forever. The production team needs to be constantly pushing back on the creative team to leave things alone and call it done.

Working out what is "good enough" is really really hard.

Update: Also, I wrote this comment before reading the article, and I don't mean to defend 6 months of crunch. That clearly _is_ bad management and exploitative. When you have a few bad features you can take a few weeks or a month to fix them. When you are 6 months out and only now just announcing a release date, there is lots you can do. You can cut the bad bits, you can hire more staff, you can move the date.

All software projects have changes in scope. And yes, that scope evolves. You get new customers (internal or external), new feature ideas, and functionality that is discovered to be unwanted or too complex.

And deadlines are moved or features are cut in order to make them. “Pizza during weekend at the office” is just bad planning.

For a game the scope evolves since you can't know how things play or feel until you test it.

So if your plan for the game is not fun when you play it, you should just ship a turd sandwich?

That's what Ubisoft does.

Like any project, sometimes you have to drop the game and work on something else. Ideally using the assets you have from the failed game.

It's not an unusual tack to take even for regular software developers. Not all feature work. Smart folks will cut bait and focus on something that does.