The GOP has been doing its best over the last five years to sabotage the economy and will continue to do so the next three. The problem is, that while they’ve slowed down the recovery, by doing so they may have actually lengthened the time it takes and thus delayed the inevitable slow down. This is my personal guesswork and not that of any economist I’ve read, but we certainly haven’t heard the term “overheated” to describe the economy over the last three or four years.
Well adahar so far so much for your prediction that Wang would be moving further to predicting GOP control by 10 to 15%. Today it moved back to 67% Dem+I.
Interesting take on Silver’s persistent attacking of Wang here. Some good points making explicable that which really has seemed a bit inexplicable to me, other than as a petulant Silver’s being butthurt that Wang’s simpler model has so far performed by almost every metric (yes, other than having too narrow of error bars in Nevada that time) at least as well as Silvers more complicated machinations.
Thing is though that if the Democrat’s maintain leadership these attacks will only have raised media awareness that Silver was with a pack consistently saying that the odds were (slightly) against it and that Wang, with his simple open-source model, as alone saying that was the way the odds were (slightly) pointing. Silver has his now media darling wizard status on the line; Wang has a day job and is only gaining attention from these attacks. Being “right” when you were alone in calling it that way always impresses more than being “right” in the middle of the consensus. Being “wrong” as part of the crowd really looks bad.
Kudos to Wang for responding in a mature manner and not attacking Silver in return btw.
If Wang’s back to 67% he’s just wrong. But we’ll find out very soon.
Braley continues to try really, really hard to lose in IA, saying he voted for airstrikes in Syria even though no such vote took place.
Hearing him lie about his positions and actions provokes disdain, hearing Ms Ernst tell the truth about hers provokes shock, horror, and dismay.
(Actually it now reads 64% … don’t know if I misread earlier or it changed.)
Once again, no we won’t. Not by who wins Senate leadership anyway. Democrats maintaining Senate leadership (hypothetically) will not prove that Wang’s model was correct or Silver’s wrong. Nor the converse. Of course most of the public is of the same statistical literacy that you evince … hence the potential big downside for Silver if the GOP loses, even though if his model is accurate the GOP should lose 42% of the time he makes a prediction that gives them the current 58% favored odds (and less often means his model is flawed). But if he says one side is “favored” and they lose the public falsely will see that his being “wrong” and no wizard after all.
As discussed earlier there are possible metrics to use. And the bulk of the public won’t use them.
Incidentally, I think I may have misread Sam Wang’s website: I worked off the Senate Current conditions (50%) and somehow missed the Election Day Probability of 50 or more Democratic+Independent seats (64%). I frankly am a little perplexed regarding the long range forecast being further from 50% than the short range forecast. If I’m reading Nate correctly, it seems that the long range forecast is a function of past polling, going back to June. That methodology strikes me as odd in an off year election with sporadic polling.
Reading over Wang’s reply, I was a little disappointed at his bragging of correctly predicting 10 out of 10 Senate races last time. He’s setting odds not making predictions: what I’d really want to see is a comparison between his predicted and the actual margin of victory. An comparison of the errors of various modeling frameworks would be the way to go.
Sam noted that while his is an amateur effort, he also gets a lot of feedback from commenters. That actually has some merit: his is more of an open source approach than that of his fellow modelers.
I don’t have any problem with the conduct of either Nate or Sam and continue to be disappointed by characterizations of their dispute.
His prediction peels off longer range information as the election gets closer (and has begun to do so) hence his current prediction of 50 D+I is 64% while his election today (which you need to hit the histogram to see) is 54% and the metamargin, also volatile short term, is D+ 0.2% (having gotten into +R territory for a few days). Yes, in fact pretty much everyone but WaPo says leadership is an anyone’s race.
Wang’s response to June bit:
DSeid: thanks.
Over at the Upshot, they put the various forecasters on a single table. For some reason they don’t list Sam Wang, which is odd. I’m guessing the WAPO model has narrower estimated standard errors than their competitors: races tend to be called further away from 50% than most models. They also have a different prediction in Kansas.
One venerable method of locating a robust central tendency is to drop the lowest and highest figure then take the average. So dropping WAPO and Sam Wang leaves us with NYT, 538, Daily Kos, HuffPo and Predictwise. They average to 58%. Too close to call.
(This is the first of heard of Predictwise.)
I will be flabbergasted if the Republicans win the Senate. Flabbergasted.
Yep. That one’s lost. BTW, the Post’s Election Lab, run by the Monkey Cage guys, says 84% chance of GOP control:
http://www.washingtonpost.com/wp-dre/politics/election-lab-2014
And in other news, Dana Milbank says “Get off my lawn!”
Care to predict the outcome of the election, adaher? How many seats will the Republicans have?
He started a thread for that.
THis is what happens when both parties are in the dumps. We may get four independents in the Senate. When was the last time that happened?
It doesn’t surprise me at all though. The Republicans are very unpopular with the public, for good reason. The only thing saving us from permanent minority status is the Democrats.
Democrats are dropping another million into South Dakota on ad buys and field operations, thinking that the race is potentially winnable at this point. I guess freeing up Michigan helps out there.
I doubt the Democrat can win. At this point it’s between the two Republicans. Their best hope is that they can lure Pressler to their side, but in South Dakota, with a former Republican? I doubt it unless they already have the majority.
Starting to feel a little anxiety, are you?
Nah. A Senate with only 46 Democrats, 50 Republicans, and 4 independents is a pretty nice Senate even if a Democrat is ostensibly the “majority” leader.