FiveThirtyEight Update: GOP still slight favorite to win Senate

[Free Bird]

I love how when Nate Silver predicts every state right in the 2012 election and forecasts an almost inevitable Democratic majority for decades to come he's an Atlas folk hero, but now that he's using the same methodologies to arrive at the conclusion of a GOP Senate come 2015 he's now unreliable or along the lines of any other Sabato or Rothenburg type.

Nice move, Atlas.  Nice move.

And I can't wait to see the waterworks here when the GOP gets to 51 on election night.

That's not true Del Tachi. He was quite inaccurate for the 2012 senate elections.

By two seats...

[windjammer]

I love how when Nate Silver predicts every state right in the 2012 election and forecasts an almost inevitable Democratic majority for decades to come he's an Atlas folk hero, but now that he's using the same methodologies to arrive at the conclusion of a GOP Senate come 2015 he's now unreliable or along the lines of any other Sabato or Rothenburg type.

Nice move, Atlas.  Nice move.

And I can't wait to see the waterworks here when the GOP gets to 51 on election night.

That's not true Del Tachi. He was quite inaccurate for the 2012 senate elections.

By two seats...

He predicted Berg would win: 95% of the chances.

[JerryArkansas]

I love how when Nate Silver predicts every state right in the 2012 election and forecasts an almost inevitable Democratic majority for decades to come he's an Atlas folk hero, but now that he's using the same methodologies to arrive at the conclusion of a GOP Senate come 2015 he's now unreliable or along the lines of any other Sabato or Rothenburg type.

Nice move, Atlas.  Nice move.

And I can't wait to see the waterworks here when the GOP gets to 51 on election night.

That's not true Del Tachi. He was quite inaccurate for the 2012 senate elections.

By two seats...

He predicted Berg would win: 95% of the chances.

When everyone, even veteran political watchers said the same thing.  She won under the radar.  All of the factors were against her.  You can't hit him for not calling an upset.

[windjammer]

I love how when Nate Silver predicts every state right in the 2012 election and forecasts an almost inevitable Democratic majority for decades to come he's an Atlas folk hero, but now that he's using the same methodologies to arrive at the conclusion of a GOP Senate come 2015 he's now unreliable or along the lines of any other Sabato or Rothenburg type.

Nice move, Atlas.  Nice move.

And I can't wait to see the waterworks here when the GOP gets to 51 on election night.

That's not true Del Tachi. He was quite inaccurate for the 2012 senate elections.

By two seats...

He predicted Berg would win: 95% of the chances.

When everyone, even veteran political watchers said the same thing.  She won under the radar.  All of the factors were against her.  You can't hit him for not calling an upset.

I can't hit him for having believed Berg would have won.

I can hit him however, for having considered this race as Safe rep.

[muon2]

I love how when Nate Silver predicts every state right in the 2012 election and forecasts an almost inevitable Democratic majority for decades to come he's an Atlas folk hero, but now that he's using the same methodologies to arrive at the conclusion of a GOP Senate come 2015 he's now unreliable or along the lines of any other Sabato or Rothenburg type.

Nice move, Atlas.  Nice move.

And I can't wait to see the waterworks here when the GOP gets to 51 on election night.

That's not true Del Tachi. He was quite inaccurate for the 2012 senate elections.

By two seats...

He predicted Berg would win: 95% of the chances.

When everyone, even veteran political watchers said the same thing.  She won under the radar.  All of the factors were against her.  You can't hit him for not calling an upset.

I would go a step further. If he calls a race at 95% then he should (and I believe does) mean that there's a 1/20 chance the race will go the other way. If there are 20 races that are rated 95% then the expectation is that one race will be an upset. He had far more than 20 races in his 2012 forecast, and even though they weren't all at 95%, one should expect some upsets. If there were no upsets, then he could be accused of cooking his books.

[windjammer]

Muon,
When you consider a race is: 95% chance to win. If finally, the candidate manages to lose by 1 point without any scandal. I'm sorry, this is inaccurate.

Of course, there is 1/20 to win. But that does mean that if a pollster predicts 99% of winning for a candidate, and he finally loses, he shouldn't be blamed.

Seriously, republicans of this forum, be honest, the last 538 predictions for the senate were heavily favoring the GOP. The predicted: 53-47 for the dem. And 2 seats, that's a lot, especially for the control of the senate.

[Antonio the Sixth]

In ND and MT, Silver's model gave too much weigh on "State fundamentals" at the expense of polls, which showed close races in both cases.

[JerryArkansas]

Muon,
When you consider a race is: 95% chance to win. If finally, the candidate manages to lose by 1 point without any scandal. I'm sorry, this is inaccurate.

He didn't and you aren't considering the most important factor, he didn't campaign hard for the race.  Everyone, including him thought he would win, so he didn't campaign hard.  And the 95%, he was just 95% sure of the win, he thought she had a chance to win.  Margin is no factor in that.

So stop being but hurt.  The man knows what he is doing.  Everyone gets things wrong.  Will he get a race wrong this year, yeah he will.  Everyone will.  That doesn't mean that his predictions weren't right and aren't good. 

You must be logged in to read this quote.

He still predicted there controll.  He got only two wrong.  I don't see the problem.  His numbers are good.  You are just mad because he is predicting rep controll now.   

Also, over the past three election cycles, they have only gotten 5% of senate races wrong.

[dmmidmi]

In ND and MT, Silver's model gave too much weigh on "State fundamentals" at the expense of polls, which showed close races in both cases.

Shouldn't this cast some doubt on what he considers "state fundamentals"? For instance:

-Democrats have held Montana's Class 2 Senate seat since 1913
-They have also held North Dakota's Class 1 Senate seat since 1960

The concern here is that "state fundamentals" is just a fancy term for "conventional wisdom."

[Keystone Phil]

I love how when Nate Silver predicts every state right in the 2012 election and forecasts an almost inevitable Democratic majority for decades to come he's an Atlas folk hero, but now that he's using the same methodologies to arrive at the conclusion of a GOP Senate come 2015 he's now unreliable or along the lines of any other Sabato or Rothenburg type.

Nice move, Atlas.  Nice move.

And I can't wait to see the waterworks here when the GOP gets to 51 on election night.

This place is flooded with Dem hacks? NO!

Literally nothing could be more hackish than predicting Rick Santorum would win in 2006, except possibly if there were Democrats predicting Blanche Lincoln would win in 2010. Are you really in a position to throw stones here?

Anyway, I was never a Nate Silver cultist. He's good at what he does and I respect his work, but he's not infallible.

Yawn. You're boring.

[Free Bird]

I'm starting to think the Democrats here are freaking out and in denial that Reps are gaining a bigger upper hand in this cycle.

[7,052,770]

Again people, Silver didn't get any states "wrong" in 2012. There are supposed to be occasional upsets, unless he gives a candidate exactly 100.00% chance of winning. The fact that Heitkamp is the only candidate with a single-digit probability on election day to win actually makes his model look better than if no one ever had.

[JRP1994]

It's obvious from reading this thread who understands statistical probability and who doesn't.

[JerryArkansas]

Again people, Silver didn't get any states "wrong" in 2012. There are supposed to be occasional upsets, unless he gives a candidate exactly 100.00% chance of winning. The fact that Heitkamp is the only candidate with a single-digit probability on election day to win actually makes his model look better than if no one ever had.

Thank you Harry, you said it better that I could have.

[Senator Cris]

The Tester victory in Montana was predictable.
http://en.wikipedia.org/wiki/United_States_Senate_election_in_Montana,_2012#Polling

[windjammer]

I m not a hack.

I indeed believe that républicans are right now favored to take the senate.

However, I believe that silver wasn t accurate for his senate prédictions in 2012 and indeed he wasn t accurate. Two seats that s a lot. And predicting democrats were going to keep the senate was nt really difficult at the end of 2012

[Amenhotep Bakari-Sellers]

I'm starting to think the Democrats here are freaking out and in denial that Reps are gaining a bigger upper hand in this cycle.

I dont get that feeling at all. www.electionprojection.com still projects that Begich and Landrieu are the stronger campaigners and narrowly winning. And Hagen providing that 51st seat.

In line with Nate silver.

[Senator Cris]

I'm starting to think the Democrats here are freaking out and in denial that Reps are gaining a bigger upper hand in this cycle.

I dont get that feeling at all. www.electionprojection.com still projects that Begich and Landrieu are the stronger campaigners and narrowly winning. And Hagen providing that 51st seat.

In line with Nate silver.

For ElectionProjection.com, Tillis is the favorite in NC. 50-50

[Senator Cris]

I asked the 538 guys to look at this over Twitter and actually got a response from Harry Enten, who writes political articles for them, so maybe they'll post an analysis of past 538 probs soon.

Did he say when they plan to release their real model? It would be nice if they did before November...

I asked him...

Kri America ‏@KriAmerica  6 h
@ForecasterEnten When you plan to release the real model for the Senate races? Thanks.

Harry Enten ‏@ForecasterEnten  44 min
@kriamerica shud be before sept. We will see tho

[IceSpear]

I love how when Nate Silver predicts every state right in the 2012 election and forecasts an almost inevitable Democratic majority for decades to come he's an Atlas folk hero, but now that he's using the same methodologies to arrive at the conclusion of a GOP Senate come 2015 he's now unreliable or along the lines of any other Sabato or Rothenburg type.

Nice move, Atlas.  Nice move.

And I can't wait to see the waterworks here when the GOP gets to 51 on election night.

This place is flooded with Dem hacks? NO!

Literally nothing could be more hackish than predicting Rick Santorum would win in 2006, except possibly if there were Democrats predicting Blanche Lincoln would win in 2010. Are you really in a position to throw stones here?

Anyway, I was never a Nate Silver cultist. He's good at what he does and I respect his work, but he's not infallible.

Yawn. You're boring.

LOL, come on dude, Santorum never led in a single poll and trailed by double digits practically the entire campaign. Insisting he would win is the epitome of hackishness, and I doubt any Democrat on this forum could match it.

[Keystone Phil]

I love how when Nate Silver predicts every state right in the 2012 election and forecasts an almost inevitable Democratic majority for decades to come he's an Atlas folk hero, but now that he's using the same methodologies to arrive at the conclusion of a GOP Senate come 2015 he's now unreliable or along the lines of any other Sabato or Rothenburg type.

Nice move, Atlas.  Nice move.

And I can't wait to see the waterworks here when the GOP gets to 51 on election night.

This place is flooded with Dem hacks? NO!

Literally nothing could be more hackish than predicting Rick Santorum would win in 2006, except possibly if there were Democrats predicting Blanche Lincoln would win in 2010. Are you really in a position to throw stones here?

Anyway, I was never a Nate Silver cultist. He's good at what he does and I respect his work, but he's not infallible.

Yawn. You're boring.

LOL, come on dude, Santorum never led in a single poll and trailed by double digits practically the entire campaign. Insisting he would win is the epitome of hackishness, and I doubt any Democrat on this forum could match it.

Ok and in the eight years since that election, I've explained my "prediction" probably a dozen times. You can keep harping on that prediction. Or even another one or two. It doesn't negate the fact that this place is a haven for Dem hackery applied to many races.

[Nichlemn]

If Nate Silver had to analyse Nate Silver's predictions with an unbiased eye, what he'd likely say is there's too small a sample size to get a good idea of exactly how his models are biased. It's clear he hasn't predicted 95%+ or whatever of races by chance alone, but if there's a pattern in his misses? People are making a huge deal over 2 close Senate races in the same year. I wouldn't go so far as to say that it means nothing, perhaps there is something to it, but there isn't enough data to say so with much confidence.

[TJ in Oregon]

If Nate Silver had to analyse Nate Silver's predictions with an unbiased eye, what he'd likely say is there's too small a sample size to get a good idea of exactly how his models are biased. It's clear he hasn't predicted 95%+ or whatever of races by chance alone, but if there's a pattern in his misses? People are making a huge deal over 2 close Senate races in the same year. I wouldn't go so far as to say that it means nothing, perhaps there is something to it, but there isn't enough data to say so with much confidence.

I would argue the biggest flaw in Silver's system is that it overestimates the likelihood of outcomes considered "Safe", especially early in the cycle. For instance, look at the earlier columns from the 2008 chart Harry posted. Notice the plethora of seats with really lopsided percentages (NC, OR, MN) that changed drastically as the race unfolded. That seems like too much variation to accept, say that as of July 20th the Democrats truly only had a 9% chance of taking the NC senate seat.

Of course, the model can't be much better than the polling it has to work with. I also find the margin of error reported by polling companies to be particularly useless because the reported margin is how far off we would expect the poll to be (at a given confidence level) only due to statistical variation. That assumes the poll was a sample that is actually representative of the electorate. Therein lies the greater problem.

[Nichlemn]

If Nate Silver had to analyse Nate Silver's predictions with an unbiased eye, what he'd likely say is there's too small a sample size to get a good idea of exactly how his models are biased. It's clear he hasn't predicted 95%+ or whatever of races by chance alone, but if there's a pattern in his misses? People are making a huge deal over 2 close Senate races in the same year. I wouldn't go so far as to say that it means nothing, perhaps there is something to it, but there isn't enough data to say so with much confidence.

I would argue the biggest flaw in Silver's system is that it overestimates the likelihood of outcomes considered "Safe", especially early in the cycle. For instance, look at the earlier columns from the 2008 chart Harry posted. Notice the plethora of seats with really lopsided percentages (NC, OR, MN) that changed drastically as the race unfolded. That seems like too much variation to accept, say that as of July 20th the Democrats truly only had a 9% chance of taking the NC senate seat.

Of course, the model can't be much better than the polling it has to work with. I also find the margin of error reported by polling companies to be particularly useless because the reported margin is how far off we would expect the poll to be (at a given confidence level) only due to statistical variation. That assumes the poll was a sample that is actually representative of the electorate. Therein lies the greater problem.

This may be the case, but it's worth noting that a number of the past few elections have been quite extraordinary. 2008 and 2010 were some of the biggest waves for their winning parties in decades. In almost every other election cycle, incumbent Senators with no obvious issues cruise  to re-election. The recent political turbulence may make it seem like there's more variation than there really is.

[Antonio the Sixth]

Harry Enten ‏@ForecasterEnten  44 min
@kriamerica shud be before sept. We will see tho

SEPTEMBER?!? "Should be"? Are you f**king kidding me?!?