What is the statistical likelihood that so many high-profile races can finish by such close margins? Bush/Gore, Coleman/Franken, Gregoire/Rossi, etc. Someone is finagling the votes.
The likelihood of close races in a lot of these cases is very good. Two of the three you mentioned, Gregoire/Rossi and Coleman/Franken are in heavy Democratic states where it is good that the Republican can compete.
Also Bush won in that contest and I believe Coleman is going to win this one.
As to the Gregoire/Rossi race, that was bogus they way Gregoire won.
But the answer to your question is that there is no violation of statistical principals/probablities with those three races.