It's a math problem. You make some assumptions, and then you solve the problem with respect to those assumptions. Of course all models of the world are flawed.

It's a math problem. You make some assumptions, and then you solve the problem with respect to those assumptions. Of course all models of the world are flawed. But some are useful. If you're wondering which restaurant you should stop at on your walk from a to b but aren't familiar with the area, you might assume you should pick the next place better than all previous after walking a third of the way home. That would be a reasonable application of this model. Regarding your last point: I wouldn't base serious life decisions exclusively on an "optimal game theory" move. That would be nuts! This obviously doesn't perfectly model reality, so why act like it does?