Our concern about Silver’s predictions: Today, we’re going to admit it:
We’ve come to regard Nate Silver’s daily predictions as one more of the shiny objects which distract from the real campaign.
That said, his work is tightly reasoned! This morning, we read this in our hard-copy Times (Washington Edition):
SILVER (10/23/12): Still, it is important to emphasize that saying Mr. Romney is the underdog is not the same thing as saying he will lose. If we calculated the odds correctly, he has a substantial chance of winning. If the set of candidates you have listed as 67 percent favorites in fact win 95 percent of the time, or 100 percent of the time, you’ve done something wrong. Over the long run, such candidates should win two out of three times—no less and no more.If we’re following that correctly, a 67 percent favorite should win two-thirds of the time!
Whatever! With two weeks to go, can we share our concern about this year’s polling predictions?
According to Silver, Candidate Romney “still seems to be trailing, by perhaps two percentage points, in the states that are most vital in the Electoral College, the ‘tipping-point’ states.” That worries us, because of something we keep hearing on The One True Channel.
On the One True Channel, we’re sometimes struck by the juxtaposition of two warring claims. First, we’ll see a segment about the voter suppression efforts which are active in the land.
Then, we’ll see a triumphant segment about the way Obama is ahead in the polls in the key swing states.
Correct us if we’re wrong! But to the extent that those suppression efforts take hold, those polling results may be bogus. People who quit on an eight-hour line don’t end up voting at all, let alone in the way they told some pollster they would.
On The One True Channel, the children rarely seem to be struck by this problem. The one type of segment is designed to make us angry at the other side. The other type of segment is designed to make us all happy and glad.
The twain doesn’t seem to meet.
Day after day, Silver’s minutely feathered predictions change by tenths of a point. On our side, we attend to our shiny object.
Over on the other side, are they finalizing the plans by which they’ll suppress that (predicted) vote?
Next post: Looking back!
The brutal fascists who run Hawaii are going to make us show ID before we vote again this year, speaking of vote suppression.
ReplyDeleteLook who is counting votes in Ohio...
ReplyDeletehttp://www.salon.com/2012/10/23/romney_linked_voting_machine_company_to_count_votes_in_ohio/
I have great respect for Mr. Sommerby, but a small defense of Nate Silver is called for. Bob pokes fun at Nate for making the obvious point that a 67% favorite only wins 2/3 of the time. But Nate notes this because of the innumeracy of his readers. So many people assume that a 70% success rate means a sure win for their candidate, but the plain fact is: if your side has 10 congressmen, each with a 70% win percentage, instead of 10 nearly-sure wins, the expected result is that 3 of those guys will lose. In Obama's case, the future of our nation basically depends on the chances of a regular power-forward hitting a free throw.
ReplyDeleteBob's real point is that Nate's polls have become a shiny bauble, and I think that's obvious. People are obsessing over his numbers as though they were batting averages. But, in the case of voter suppression efforts, Mr. Silver actually does take those laws into account and his model includes their likely effect (you can root around on his site for discussion of this). Whether he has over- or underestimated this effect is a subject for debate, but unlike most shiny baubles, Nate's does address real issues like voter suppression.
Intrade has Obama down to a 57 - 58% chance of winning. If you go to the dive bars of the Democratic Party, places like Democratic Underground and, to a lesser extent, Daily Kos, everyone's all excited and sure Obama is going to win, and that Romney is some kind of joke -- people who say otherwise are immediately attacked as "concern trolls" and other such mock-clever slights. This late in the game, a joke shouldn't even be in the running. Speaking of jokes, your pal Kevin "Alfred E Neuman" Drum used to adopt a similar line about Obama's ivincibility; don't know if he's changed his tune, as I finally stopped reading him several months ago. But anyone who looks at these numbers, at Obama's inability to draw away from Romney, and isn't concerned, is a fool.
ReplyDeletenate silver started his ridiculous blog at the daily kos and the idea that the nytimes placed it on their pages as - fair and accurate - tells you all one needs to know about how dumbed down the gray lady has become.
ReplyDeleteIf we’re following that correctly, a 67 percent favorite should win two-thirds of the time!
ReplyDeleteSpeaking as a professional, I'm with Bob. In fact, that 67 percent figure means even less than the above quote. It means precisely that Nate Silver has a model in which Obama's chance of winning is 67%.
But, models are not reality. A model is based on a limited amount of data and a bunch of assumptions that may or may not be valid. Even if some of these assumptions were valid in the past, they may not still be valid. Note that other models predict a Romney victory. We can't judge the meaning of Silver's 67%, because we don't know how his model works.
I have no idea who's going to win nor do I know which model is better. However, I am dubious about putting percentage likelihood on the model result. It's hard enough to predict who's going to win and the approximate margin of victory, let alone putting a probability estimate on the results.
I think Democrats ought to be quite concerned, simply mocking Romney and Ryan will not do but that approach is typical. At the least the race looks to be close and my reading suggests Democrats will likely hold the Senate by less than last time and Republicans will likely gain in the House.
ReplyDeleteMy reading tells me the Democrats will actually gain in the Senate by picking up seats in Massachusetts and Arizona, and possibly even Tennessee and Indiana, holding onto seats that were vulnerable in Missouri, Connecticut, Montana, Wisconsin, Ohio, New Mexico, and Virginia, and welcoming Maine's independent Angus King to the Democratic caucus, while losing North Dakota and Nebraska.
DeleteAt church even in an affluent community by the way I still hear story after story of struggles for jobs for members or children or relatives. I also hear concerns about Social Security and Medicare. Obama just does not re-assure people on such matters.
ReplyDelete