The original poster on this kuro5hin article starts out trying to cast doubt on the Ohio numbers from the election. Fine, nothing new there. But, someone who actually knows how to crunch numbers pipes in in the comments section with some pretty interesting numbers and, well, basically says "I don't see anomolies, but if you do, refute this data".
This is why numbers, and not emotions, matter in elections. Because when you get down to it, you can spin "voter attitudes" and "I remember 2000, and this year was different", but if you can't quantify it, it doesn't count.