[jodrellblank]

That's the sort of analysis I'm questioning - it's also 95% likely that I'm not in the second-last 5%, and the same for the third-last 5% and for any 5% slice.

How can we usefully reason anything from a start like that?