in policy, you can never get it just right. you should try to get close, but perhaps more importantly, you have to choose in which direction it will be less costly to err, then condition your choices so they will err — hopefully not too much but inevitably — in that direction. going for symmetrical optimality is foolish hubris.