Scott writes

. . . most people are excessively impressed by unconditional forecasts.  When dealing with business cycles and financial markets only conditional forecasts matter.  People win the lottery every day.  I’m no more impressed by an economist making an unconditional prediction that turns out correct than I would be if my plumber won the lottery.

I take it Scott’s point is that since knowledge takes the form X given Y, if you simply guess X and get it right you must have gotten lucky on Y.

Which is fine, but getting the right answer is still information. Unless X is actually more likely under not-Y then correctly getting the unconditional conveys the same basic information as the conditional forecast, its just attenuated by the fact that X has some probability on not-Y.

About these ads