BattleCard 56
Could you expand on how parameter risk has a large impact on Loss but a small impact on Discounting? Using too high of a discount rate leads to under-detection of RT and too low of a discount rate leads to over-detection of RT, so wouldn't parameter risk impact discounting in a RT test quite a bit?
For the Loss part, is this referring to undiscounted loss? As in the loss distribution?
Comments
The section on Parameter Risk explains this (page 19 in the text).