That’s not really underconfident if you’re wrong on about 20% of such estimates.
Wait, wouldn’t it be “if you were right on about 20% of such estimates”? To be calibrated correctly with 20% confidence you should be wrong about 80% of the time, right?
I’m not trying to be obnoxious about what might be a typo, it’s just that you have 3 upvotes so I’m worried I’m missing something obvious.
If there’s any status in comparing excessive underconfidence, I think I take the cake—one year off with 20% confidence in the interval. Good survey.
That’s not really underconfident if you’re only right on about 20% of such estimates.
Wait, wouldn’t it be “if you were right on about 20% of such estimates”? To be calibrated correctly with 20% confidence you should be wrong about 80% of the time, right?
I’m not trying to be obnoxious about what might be a typo, it’s just that you have 3 upvotes so I’m worried I’m missing something obvious.
Nope, you’re right.