**In the Senate:** In critical states there are so many polls that fairly exact statements can be made. **The probability of the Senate remaining Democratic is nearly 100%*.** Assuming polls are unbiased on average, with 85% probability the following **exact** outcome is predicted: **51D-49R.** Knife-edge races:

**WA: Murray (D)** over Rossi (R) **by 2.0 +/- 1.0%** (mean +/- SEM).

**NV: Angle (R)** over Reid (D) **by 2.0 +/- 0.5%**.

**CO: Buck (R)** over Bennett (D) **by 1.0 +/- 0.7%** (this is the real nailbiter).

**IL: Kirk (R)** over Giannoulias (D) by **2.0 +/- 0.8%**.

**WV: Manchin (D)** over Raese (R) **by 4.5 +/- 1.1%** (statistically a little suspenseful because only n=6 for recent stable polls).

All nominal win probabilities are over 90%.

**In the House:** **The Republicans will take control.** By the same methods as before, we’re at **230R-205D**. The 95% confidence interval for a gain of 50-54 seats to 228-232R. Pollster currently has 27 “toss-up” races with a total of over 100 polls, an amazing number.

P.S. For the record, as of 10PM tonight, two model-intensive comparables are FiveThirtyEight Senate 51.7D, House 233.1R, and StochasticDemocracy Senate 52.2D, House 237.7R. Both House results are outside my 95% confidence interval. I am interested in how far my simple-minded approach can go toward top-line characterization. But who knows, maybe there’s unseen complexity.

**This goes down to 90% if there is a 50-50 split and Lieberman jumps ship again. 100% if he stays with the Democrats this time.*

Matt McIrvin// Nov 3, 2010 at 12:47 amIt looks to me as if the unknown correlated systematics that Silver was waffling about do exist–but that they’re much, much smaller than Silver thinks they are!

Dave Rutledge// Nov 2, 2010 at 11:51 pmIf current indications are accurate, it seems we will have some interesting dissection of the hypothesis tomorrow.

Sam Wang// Nov 3, 2010 at 12:17 amI think you may be right, at least for the House. My preliminary reading is that Republican candidates ran a few points more strongly than polls indicated. Possible reasons are (1) district-level polls are sufficiently infrequent that they did not catch a last-minute swing, or (2) they are systematically biased.

In regard to the Senate, I am surprised by the size of the Manchin (D-WV) win. I wouldn’t score Nevada as a surprise yet because although Reid is currently ahead by 7%, the returns are not complete.

For Senate races, I should point out that I set a rather harsh standard by calculating the error bars as SEM. It probably would have been more appropriate to multiply the error bars by some factor. In 2004 SurveyUSA analyzed their own results suggesting that the factor would be 1.3 for their own polls.

Five election forecasters with better track records than 538 – and what they’re saying about the midterms « Casual Factors// Nov 2, 2010 at 9:04 am[...] 1. Princeton Election Consortium: Senate 51D-49R, House 205D-230R [...]