I've never liked Expected-Value type calculations because they don't take variance into account. You could have a theoretical messed-up lottery where the payof

I've never liked Expected-Value type calculations because they don't take variance into account. You could have a theoretical messed-up lottery where the payoff is greater than the odds, where EV would tell you to blow all your money on the lottery even if the odds are totally against you winning. Similarly, this wouldn't really imply that coaches should go for 2 all the time (like... you've just scored a TD to tie the game at 6 at the end of regulation, and a tie would mean you'd miss the playoffs). There's a cool equation called the Kelly Criterion to solve this but it only works for single bets. I don't know how to use it to compare between two bets, or for a case like the extra point where you are essentially betting nothing for a 92% chance to get one point.