Reply #30: I don't estimate. I use actual data. Your approach is inadequate and obfuscates vote-switching.

[L. Coyote]

The third-party cross-votes are concentrated in specific precincts,
at locations where switching Kerry votes to Bush is NOT possible, a fraud pattern

In precincts where all Kerry cross-votes count only for the third-party
candidates, the rate is from 0.81 to 1.43% for Badnarik and Peroutka respectively:



So your belief "that the problem is real and still not believe that it affected
the outcome" is ill-founded on a misperception and lack of analysis of the data.

Secondly, the vote-switching rate is NOT determined by the third-party cross-vote rate.
That is particularly so in a fraud scenario, where the idea is to switch Kerry votes to Bush.
In fact, lower third-party irregularites help obfuscate the real crime, vote-switching.

What impact do these facts have on your skepticism, and your estimate?

Also, one has to account for non-votes!