** Update ** The October 7 model in the Manchin (D) vs. Raese (R) WV Senate race has Raese with a 64% probability of winning. The Sept. 30 model, which I wrote about below had it the other way around. If you’re a betting man, and you like to bet on the favorites, you’d vote for Raese (if the election were held today, of course).
Karl Rove and Company have produced an interesting map of 2010 Senate races. Rove says that if the election were held today, Democrats will lose 10 seats (down to 49), Republicans will gain six (up to 47) and four toss-ups.
Interested in those four toss-ups, I checked out Nate Silver’s FiveThirtyEight.com blog (hosted on the NYT website) run by Nate Silver (great website and great writing). FiveThirtyEight averages a variety of surveys conducted in House, Senate, and governors races across the country, but they also calculate the likelihood that a particular candidate will win his race*. For example, Arizona governor Jan Brewer has a 98 percent chance of defeating Terry Goddard (Brewer is up by 15%). Doesn’t mean Brewer’s certainly going to win by 15%, but it does mean that she’s almost certainly going to win.
So, let’s take a look at the four Senate toss-ups:
Patty Murray (D) is up 50% – 47% over Dino Rossi (R) and has a 74% chance of winning.
Harry Reid (D) and Sharon Angle (R) are statistically tied at 48% each. However, Angle has a 55% chance of winning.
Alexi Giannoulias (D) and Mark Kirk (R) are about even (Kirk’s up 49%-48%), but Kirk has a 54% chance of winning.
Joe Manchin (D) is barely beating John Raese (R), 50%-48%, but Manchin’s chances of getting elected are 63% to Raese’s 37%.
What to make of all this? Today, it looks like the Dems will retain the majority in the U.S. Senate by the slimmest of margins, 51-49.
* What Silver’s chance of winning percentage means: The chance that each candidate will win based on 100,000 simulations with random variation in the local and national political environment.
We will continue to track these four races and other key races around the country.