I think that one in particular is inexplicable. (View Post)

Scotland was SNP 56 of 59 seats on a quasi-majority of the vote in the era before Unionist tactical voting. That's the kind of result one can predict pretty well without an intricate regression … (View Post)

My specific concerns this time are that it is easier to extrapolate a UNS-type number (i.e. relying on errors cancelling each other out, each way) when the overall two-party swing is around 1% than w… (View Post)

How I think about it is that UNS works - in so far as it does - by hoping the errors are similar each way across a pretty smooth distribution of seat majorities in the critical area for seats' c… (View Post)

Which side was which? (View Post)