538: GOP slight favorite for Senate control

[RogueBeaver]

Basically a coin-flip.

[7,052,770]

I don't really think Michigan is as much of a tossup as Alaska, but I'm not going to doubt 538's models.

[Anti Democrat Democrat Club]

Yeah, I'd tweak MI and NH's ratings up 5-10 points.

[Amenhotep Bakari-Sellers]

Not nothing new here. Cook tossup ratings weren't correct and Larry Sabato crystal ball gives us more of a clear picture.

3 states that matter are NC, MI and AK

55 percent chance of Dems retain the senate due to weakness of GOP opponents in NC the tipping pt state.

[windjammer]

Remember 538 has always been bad for senate predictions!

[Amenhotep Bakari-Sellers]

They are simply going with the GOP because they dominate the House. The Dems won't let this slip away from them.

[Antonio the Sixth]

Is this an actual forecast model or just the editors giving their own rankings? I don't have the time to read the whole thing right now.

[Meeker]

Is this an actual forecast model or just the editors giving their own rankings? I don't have the time to read the whole thing right now.

It's a forecast model like the ones 538 has designed in the past.

Taking anything away from this beyond "control of the Senate is up for grabs" seems silly to me. It's still early to be focusing intensely on the odds in individual races.

[Joshgreen]

Silver wasn't great last cycle on Senate races in the upper Plains and Rockies: Montana and North Dakota. He relies too much on 'state fundamentals', as he calls it.

[Maxwell]

This looks accurate, though I'd give Begich more of a likelihood to win re-election, particularly because there is a 10% chance the GOP completely blows it and nominates Joe Miller.

[ElectionsGuy]

This looks accurate, though I'd give Begich more of a likelihood to win re-election, particularly because there is a 10% chance the GOP completely blows it and nominates Joe Miller.

Actually what will really blow it is if Miller runs as an independent. He has about a 5-10% following, so if he does that.... Begich can thank the gods of politics.

[International Brotherhood of Bernard]

55 percent chance of Dems retain the senate due to weakness of GOP opponents in NC the tipping pt state.

I believe this is what will keep the Senate Democratic in the end. GOP absolutely blew recruiting in NC, any credible candidate would easily defeat Hagan. Democratic control would be further sealed by a Broun nomination in GA (which I do believe will happen, because you can almost always count on the GOP to nominate the craziest possible candidate.)

[International Brotherhood of Bernard]

This looks accurate, though I'd give Begich more of a likelihood to win re-election, particularly because there is a 10% chance the GOP completely blows it and nominates Joe Miller.

Actually what will really blow it is if Miller runs as an independent. He has about a 5-10% following, so if he does that.... Begich can thank the gods of politics.

I would literally die of joy if that happened.

[publicunofficial]

Silver had the GOP as a 68% favorite to take the Senate last cycle up until the last few months, and even then had Rick Berg and Denny Rehberg with a 90% chance of victory. Silver's analysis is much more of an outside view than Sabato, Cook, ect. A lot of Silver's thinking is "Louisiana is a very red state, therefore Democrats can't win there, therefore Landrieu is doomed." For some reason he doesn't translate that logic to Michigan, but whatever.

[Amenhotep Bakari-Sellers]

Hagen or Landrieu will lose and then it will be up to Udall and Begich to come through.

[7,052,770]

Just because Silver said that Berg had a 92% chance on Election Day doesn't mean that Silver was wrong.  One out of every 12-13 times he says "92%" on Election Day, he should be "wrong."

[publicunofficial]

Just because Silver said that Berg had a 92% chance on Election Day doesn't mean that Silver was wrong.  One out of every 12-13 times he says "92%" on Election Day, he should be "wrong."

Not saying he's wrong, just that his chances are misleading.

[Maxwell]

Just because Silver said that Berg had a 92% chance on Election Day doesn't mean that Silver was wrong.  One out of every 12-13 times he says "92%" on Election Day, he should be "wrong."

Not saying he's wrong, just that his chances are misleading.

How many people really thought Berg was going to lose?

*besides me *

[windjammer]

Just because Silver said that Berg had a 92% chance on Election Day doesn't mean that Silver was wrong.  One out of every 12-13 times he says "92%" on Election Day, he should be "wrong."

Not saying he's wrong, just that his chances are misleading.

How many people really thought Berg was going to lose?

*besides me *

This race was more tilt rep/lean rep. Not safe rep.

[publicunofficial]

Also:

You must be logged in to read this quote.

Nice to know Silver's still being a petty douchebag over PPP.

[Maxwell]

Just because Silver said that Berg had a 92% chance on Election Day doesn't mean that Silver was wrong.  One out of every 12-13 times he says "92%" on Election Day, he should be "wrong."

Not saying he's wrong, just that his chances are misleading.

How many people really thought Berg was going to lose?

*besides me *

This race was more tilt rep/lean rep. Not safe rep.

That's fair, but for real, just about every poll had Berg ahead, and it was North Dakota.

[Joshgreen]

Just because Silver said that Berg had a 92% chance on Election Day doesn't mean that Silver was wrong.  One out of every 12-13 times he says "92%" on Election Day, he should be "wrong."

A model like Silver's shouldn't be judged on its overall hit rate (any schmuck with an interest in politics can call 80-90% of the races correctly). His model should be judged on its ability to call close races, which it has spectacularly failed at in several instances.

[windjammer]

Just because Silver said that Berg had a 92% chance on Election Day doesn't mean that Silver was wrong.  One out of every 12-13 times he says "92%" on Election Day, he should be "wrong."

Not saying he's wrong, just that his chances are misleading.

How many people really thought Berg was going to lose?

*besides me *

This race was more tilt rep/lean rep. Not safe rep.

That's fair, but for real, just about every poll had Berg ahead, and it was North Dakota.

http://en.wikipedia.org/wiki/United_States_Senate_election_in_North_Dakota,_2012
Not really. It was closer, but every pollster has ignored these polls!

[bballrox4717]

I think it's gonna come down to whether the Republicans blow it in Kentucky, Mississippi or Georgia. Otherwise the Senate control will come down to the run-off for Landrieu's seat, and she'll get hammered by Cassidy if all the press is on her winning for the Dems to keep the Senate.

[7,052,770]

Just because Silver said that Berg had a 92% chance on Election Day doesn't mean that Silver was wrong.  One out of every 12-13 times he says "92%" on Election Day, he should be "wrong."

A model like Silver's shouldn't be judged on its overall hit rate (any schmuck with an interest in politics can call 80-90% of the races correctly). His model should be judged on its ability to call close races, which it has spectacularly failed at in several instances.

Silver doesn't "call" races at all, he gives probabilities.

His model should be judged on things like:

• Do candidates he says have a 90% chance of winning on Election Day win 90% of the time, not more or less?
• Do candidates he says have a 50% chance of winning on Election Day win 50% of the time, not more or less?
• Do candidates he says have a 27% chance of winning on Election Day win 27% of the time, not more or less?
• etc.

He absolutely should NOT be judged on how often the candidates he gives a 51% to win actually do, vs. candidates he gives a 49% chance to win.  Well, at least until he gives enough 51-49 odds to have a large enough sample judge if 51% calls win slightly more often than the 49%.