Suppose you are right 90% of the time in some activity, where you gain $1 for being right, lose $1 for being wrong
Suppose a new strategy is right 80% of the time, wrong 20%, but with ± $2.
Your expected value change is +40c.
Moral: There is such a thing as being too right 🙂
Conversation
Show replies
Replying to
and you can prove anything if you're allowed to change all the variables randomly :)
Replying to
If you increase payout and decrease hit rate linearly it peaks at around 75% right with +-2.5 payout. Optimal truth/bullshit ratio?




