(nods) Absolutely.

Unfortunately, I came installed with a fairly broken evaluator of chances, which tends to consistently evaluate the probability of X happening to person P differently if P = me than if it isn't, all else being equal... and it's frequently true that my evaluations with respect to other people are more accurate than those with respect to me.

So I consider judgments that depend on my evaluations of the likelihood (or likely consequences) of something happening to me vs. other people suspect, because applying them depends on data that I know are suspect (even by comparison to my other judgments).

But, sure, that consideration ought not apply to someone sufficiently rational that they judge themselves no less accurately than they judge others.