Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I think it is more about the "risk of ruin" than infinities.

Here is the idea: you go in a casino with $1000, that's your self-imposed limit, you will not die if you lose it, but it means you will stop playing, and you will have no way to recover from your losses, time to go home. That's what ruin is.

And this is something you have to take into account. For example, imagine you find a game where you have 50% chance of losing your bet, and 50% of winning your bet +2%. You can bet up to $1000. In theory, you get the best expectancy ($10) if you play $1000, but since it is your limit, you actually have a 50% chance of losing it all with no other chance to play this profitable game again, the result is you can expect to leave the casino poorer. If instead, you start with smaller bets, slowly ramping up as you play, you will be able to take full advantage of the game indefinitely, and almost certainly end up richer.

Note that in a real casino, games have a house edge, so from a pure money perspective, ruin is not a bad thing because it prevents you from playing more negative expectancy games, and it will happen eventually. But since if like me, you just go there to have fun and consider the house edge to be the price of entertainment, then lowering the risk of ruin is actually the main goal.



You are correct. For economic opportunity evaluations, it often superior to treat 0 as having negative infinite utility (e.g. ruin aversion). My comment was about more everyday situations where that is clearly not how people behave.





Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: