When designing a game, most people at least give lip service to the concept of iteration: that is, no one gets it right on the first try, EVER, so build some time into your schedule to make major changes that you'll find along the way. It's the game developer version of the saying from engineering, "build one to throw away". I'm sure other fields have their own jargon for essentially the same thing.
Of course, with games (especially for console), once you ship your game you're done. No more iteration for you! If you're lucky and the original game is Good Enough, you might be able to improve some things in the sequel... but you can never go back and actually change the first one. "Remakes" of old games are incredibly rare.
Designing a class is just the opposite. I taught Game Industry Survey once already, and I'll teach it again two more times this year. Each time, I get to look at the results from previous classes and iterate on the design, improving weak points, tweaking the grading system, updating the content, and generally making the whole thing better. And I get to do this again and again, with fresh students each time who don't have any previous experience with earlier versions.
This property of classes encourages risk-taking. I can do something crazy like make each homework worth a hundred thousand points, and if it doesn't work then I can just apologize to the current class and change it next time around. It's quite a different environment from the risk-averse environment of the game industry, and the only thing that changes is that I get to redo my work after it "ships".