Not sure about you, but my university was very strict with cheaters because of how widespread cheating was. Gaining an extra 10 points on tough exam wasn't worth the risk of expulsion. I suppose that deciding whether to cheat or not is an optimization problem in and of itself. If no one else is cheating, the gains are huge, but if there are a lot of cheaters, it might not end up being worth the risk.
Actually, there never was a time where I felt that cheating was worth it. On the contrary, the amount of time I'd need to prepare a strategy (compile cheat sheet, figure out how to use it, where to sit, etc.) probably exceeded the time I'd take to just study for the exam!
I agree, always following the rules isn't beneficial to society, but I'd argue it's better than the case where no one is following the rules.