Good morning. First, the basic points: the polling data were fine, but my follow-up assumptions about net undecideds and turnout were wrong. The statistical approach is fine, and indeed still the best way to look at the data. The final result will be very near, and perhaps exactly equal to the decided-voters calculation. Finally, now that votes are mostly in, we have real results to compare with pre-election polls. I will be doing that in the coming days. This is the point at which something real can finally be learned.