Thursday, November 6, 2014

Comparing the 2014 US Senate Election Models

So how did all the modelers do? Pretty good overall. While all the models did pretty well, there were misses by all of them: specifically, none of the models predicted the outcome of North Carolina correctly.

Following Dr. Sam Wang from the Princeton Electoral Consortium, I have ranked the different models using a Brier score. It seems like my scores came out slightly different from his, probably because of the data I was able to find, but the results seem consistent.

As input the numbers provided by the New York Times Senate Model Comparison. I have converted each of the totals provided by them to the probability of a Republican win. Therefore, if they predicted a Democratic win by 75%, the number was converted to a Republican win by 1-D% resulting in a 25% chance of a Democratic win.

After calculating the Brier score for each of the models, here are the outcomes in order of best to worst (lower Brier score is better):

Brier Score
Daily Kos
Washington Post

As you can see, all the various models did really well.

Princeton Electoral Consortium –
Wikipedia: Brier Score –
New York Times: 2014 Senate Model Comparison –

Tuesday, November 4, 2014

US Senate Final Update

These are the final predictions that will be released from "El" on this election cycle. It looks like it is going to be a Republican blowout unless there is something systematically wrong with the polling.

El is predicting 48 seats for the Democrats and Independents, and 52 seats for the Republicans. If Orman wins and caucuses with the majority party, that will make yield a total of 47 seats for the Democrats and Independents, and 53 seats for the Republicans and Independents.

The individual probabilities for each state as predicted by El are as follows:

And in non-graphical form:

Sullivan 56% chance of winning
Begich 44% chance of winning

Cotton 86% chance of winning
Pryor 14% chance of winning

Gardner 77% chance of winning
Udall 23% chance of winning

Perdue 87% chance of winning
Nunn 13% chance of winning

Ernst 68% chance of winning
Braley 32% chance of winning

Roberts 37% chance of winning
Orman 63% chance of winning

McConnell 95% chance of winning
Grimes 5% chance of winning

Cassidy 95% chance of winning
Landrieu 5% chance of winning

Land 5% chance of winning 
Peters 95% chance of winning

New Hampshire
Brown 30% chance of winning
Shaheen 70% chance of winning

North Carolina
Tillis 24% chance of winning
Hagan 76% chance of winning

Adding all these seats together gives us the following results for the elections. While there are some ties listed, in the prediction, any ties go to the one with any lead whatsoever. The easiest way to do this is to only look at the short term prediction. In the following chart, a "1" signifies that the state is predicted to be won by the Republicans. A "0" signifies that the state is predicted to be won by Democrats.

Finally, we want to look at the sums of all these numbers together with the states that either aren't up for election this year or are considered safe seats by both the Democrats and Republicans.