If I expect that my estimate will change in the future, why not change it now? I grant that it is highly likely that my estimates will change, but I don’t know whether any particular estimate will change upward or downward, so for now they stay put.
I suppose what anticipation of change in a probability estimate practically means is that you expect new pieces of evidence to come in and that you have a fairly good idea what the magnitude of evidence will be, just not the sign.
I don’t know which direction it will change, but for things I’m unsure of I expect more movement than for things I know more about. In bayesian terms, a weak prior.
If I expect that my estimate will change in the future, why not change it now? I grant that it is highly likely that my estimates will change, but I don’t know whether any particular estimate will change upward or downward, so for now they stay put.
I suppose what anticipation of change in a probability estimate practically means is that you expect new pieces of evidence to come in and that you have a fairly good idea what the magnitude of evidence will be, just not the sign.
I don’t know which direction it will change, but for things I’m unsure of I expect more movement than for things I know more about. In bayesian terms, a weak prior.