Frequentists think that ignoring this problem means it goes away.
They don’t.
I don’t mean the negative-answer problem. I mean “the confidence interval simply is not the posterior probability full stop” problem.
Well, sure. But whither calibration?
They don’t.
I don’t mean the negative-answer problem. I mean “the confidence interval simply is not the posterior probability full stop” problem.
Well, sure. But whither calibration?