The Washington PostDemocracy Dies in Darkness

All of the election models are starting to converge. And they are all pointing to a Republican Senate.

Over the past few days, two of three big election models -- the Post's and the New York Times' -- have updated their predictions about the likelihood of Republicans taking the Senate majority in 69 days. And, the predictions are remarkably similar.

First, the Post's Election Lab, run by George Washington University professor John Sides, gives Republicans a 58 percent chance of winning the six seats they need to be in the majority come 2015. Writes Sides: "The Republicans are very likely to control at least 48 seats after the elections, including Georgia. With just three seats out of the [six] discussed above, they would take control of the Senate. That’s why we still see them as having a better than 50:50 chance to do it."

Leo, the New York Times' Senate model, shows Republicans with a 67 percent chance of taking the majority.  Writes the Upshot's Josh Katz:

(The third major modeler, 538's Nate Silver, hasn't issued a prediction from his model of late but as of early August he wrote that Republicans were "60-40" favorites to win the Senate majority.)

Now, it's worth noting that the Post and the Times have arrived at their conclusions from very different places. The Post's 63 percent likelihood of a Republican Senate takeover is down from an over 80 percent probability earlier this year. The Times model, on the other hand, is at its highest probability of a GOP takeover -- up from the low to mid 50s earlier this year.

The convergence of the models is to be expected.  The closer the election gets, the more settled the general election matchups are and the more polling exists about those matchups.  Plugging in specific candidates with specific polls attached to them makes it increasingly less likely that the models won't differ all that greatly.  Sides explains that trend this way:

A key reason is that polls are taking on greater weight in the forecast. Although both the House and Senate forecasts still incorporate our statistical model’s expectations — that is, expectations based on factors like presidential approval, state partisanship, incumbency, etc. — at this point our Senate forecast weight the polls more heavily in races with substantial polling. The polls will take on increasing weight in the forecast over the next month or so.

So, where do the Post and Times models disagree?  I looked at the nine most competitive races -- Alaska, Arkansas, Colorado, Georgia, Iowa, Kentucky, Louisiana, Michigan and North Carolina -- to find out.

In six of those nine races there is general  agreement between the models.

1. Alaska: WaPo 66 percent chance D victory/NYT 52 percent D victory (The Post model is more bullish on Sen. Mark Begich's chances but not by a ton.)

2.  Arkansas: WaPo 77 percent GOP win/NYT 78 percent GOP win

3. Colorado: WaPo 64 percent D win/NYT 57 percent D win

4. Georgia: WaPo 84 percent GOP win/NYT 82 percent GOP win

5. Kentucky: WaPo 94 percent GOP win/NYT 86 percent GOP win

6. Louisiana: WaPo 57 percent GOP win/NYT 60 percent GOP win

There are three races in which the Post and Times models differ.

1. Iowa: WaPo 72 percent GOP win/NYT 55 percent D win

2. Michigan: WaPo 99 percent D win/NYT 69 percent D win

3. North Carolina: WaPo 92 percent D win/NYT 51 percent GOP win

The difference in Michigan's open seat race is the least major of the three as both models favor a victory by Democratic Rep. Gary Peters -- just with differing levels of confidence. But, in Iowa and North Carolina, the models are predicting totally different outcomes. These differences are almost likely dependent on the weight the models give to prior experience in elected office, historical results in the state and available polling.  (Most high-end models like these are similar in terms of how they weight these various factors but not identical.)

Here's what's fascinating: If you assume -- as both models do -- that Democratic-held seats in Montana, South Dakota and West Virginia will be easily won by Republicans, then the GOP needs to pick up three more seats to win the majority. Both models agree that Arkansas and Louisiana look like GOP pickups while Republicans are likely to hold their own seats in Georgia and Kentucky.  According to the Post model, Iowa is the majority maker for Republicans while North Carolina makes the majority -- barely -- in the Times model.

Republicans, of course, will take those six seats any way they can get them.