Over the last eight days there has been substantial movement in four of the key senatorial elections. The first is Kansas, where I feel my prediction is a little low. The reason is that the fundamental model, which still has some power, is not well identified on independent elections! But, you have to run with what the data says. Second, Colorado, Arkansas, and Louisiana have all trended a little bit towards the Republicans over the last few days. None of these are serious issues, as much as the release of new polling.
It is important to note two weird structural things about this election. First, the independent in Kansas has declared for the Democrats if they get 50 seats or more and the Republicans if they have 51 seats or more. There is a 20% likelihood of the senate being 49 Democrats and 50 Republicans, with Kansas not sure. Second, the Louisiana race goes to a December 6 runoff if neither candidate gets 50%. This is extremely likely, making the standard three-way polling very confusing, because the third candidate is a Republican. This gives the Republican an increased edge in any runoff election.