Oh, no question the obstacle is high at 100K. But the Secy of State site gives the raw numbers for 2000 PBs. It was about 100K and 86% were valid votes.
My calculation is 155K (this year's PB total) X 0.86 = 133000.
That 133K number is smaller than Bush's lead, but that's not the threshold of concern. That lead is about 2.5% of the vote. I am under the impression there is an automatic recount at 0.25% of the vote, which is 10 times less, or 13K. So if the gap closed 133K - 13.3K = 120ishK, we're in dangerous territory.
There may be recounts allowed if requested maybe inside 0.5%, but with the concession I discount that concern.
So, there are 133K (if the 86% holds true) valid PBs and 120ishK must go Kerry. That's 90% and no, that split is very unlikely. But . . . it's mathematically possible.
I want it impossible.