By the numbers

>> Tuesday, October 30, 2012

Digby points to an opinion piece by Dean Chambers in which he argues that Nate Silver's Five Thirty Eight is wrong because Nate Silver is a girly man.  No, seriously; Chambers writes:

Nate Silver is a man of very small stature, a thin and effeminate man with a soft-sounding voice that sounds almost exactly like the "Mr. New Castrati" voice used by Rush Limbaugh on his program. In fact, Silver could easily be the poster child for the New Castrati in both image and sound. Nate Silver, like most liberal and leftist celebrities and favorites, might be of average intelligence but is surely not the genius he's made out to be. His political analyses are average at best and his projections, at least this year, are extremely biased in favor of the Democrats. [internal links omitted]
In case you were wondering if fag-baiting was still around: apparently.
Presumably, Dean Chambers is a reliable bloviator because, I imagine, he chain-smokes unfiltered cigarettes, drinks bourbon straight with no chaser, wears his cowboy boots to bed and drives a Hummer that's been specially engineered to use baby seals for fuel.  Not only does Dean Chambers not eat quiche, he has never even seen one and wouldn't know how to spell the word (he thinks it begins with a manly "K" and the effete lisp is spelled out, and this is why no manchild of his would ever be named "Keith", though he has considered it a possible name for a daughter).  When Dean Chambers wants a steak, he goes out and throttles a fresh cow with his bare hands, cuts off the piece he wants and chars it with his cigarette lighter before eating it; when he wants dessert, he just throttles himself a new cow.  And he does not, repeat, not have any time for your queer math with its anal zeroes, phallic ones, limp-wristed twos, bent-over sevens and don't even get him started on how far an eight will go to make your daughter into a professional tennis player when she grows up.


Tim Stanley at the Telegraph makes essentially the same argument with more cogent arguments and a hundred-percent-less sashaying innuendo.  You can look through his points and decide whether you agree with any of them or not.  Me, I actually find the whole thing interesting on another level.

Which is the fact that Nate Silver points to an area in which the majority of liberals and the majority of conservative differ, I think.
We should start with what ought to be obvious.  Silver isn't at the New York Times because he's a lefty and the Times is a lefty paper and they're supporting each other's lefty agenda.  Nate Silver is at the Times because he an a political blog that had a seemingly good track record predicting the 2008 Democratic primaries and election results based on running poll numbers through a simulator Silver devised based on his experience as a sabermetrician, and this blog got an enormous online readership because Silver seemed to be a reliable predictor and his model and number-crunching appeared to be legitimately mathy and rigorous and not merely based on some kind of gut wisdom, and so the Times saw this and the New York Times is a business that's trying to figure out how to transition from a late-19th Century business model to a post-Internet model and Nate Silver's blog was exactly the kind of "new media" the Times is trying to copy / create / invest in / buy up / figure out / grok / compete with / destroy / coexist with; so the Times made Nate Silver an offer he couldn't refuse (apparently), and he didn't and here we are.
So, Silver is at the Times because he had a good run in 2008 and 2010, and the Times wanted to be a part of that.  But does that mean he's unbiased?  Well, no, one supposes not, though he seems to be doing his level-headed best on that score.  If he is unbiased and is trying to be rigorously mathy, does that mean he'll be right?  And again, no, he may be way off.
Certain conservatives are angry and upset, though, because right now as of this writing, for instance, Nate Silver is predicting President Obama has a  74.6% chance of being reelected.  This is crushing to these folks, because they evidently feel that Mitt Romney somehow has a one hundred percent chance of beating the President even though the Electoral College math is challenging for him even when you start throwing close-to-call battlefield states his way.  (These folks also don't seem to understand what a probability is in any case: 74.6% may be comforting to me, but I'm not overlooking the fact--excuse me while I dork out for a moment--that I've sent beloved Dungeons And Dragons characters into the maw of death knowing they had approximately a 25.4% chance of rolling the fifteen or better they needed to survive on a twenty-sided die (that would actually be a 30% chance of rolling well, kiddos).  People bet on football games and buy lottery tickets on worse odds than a 1-in-4 chance of winning.  That some people--right or left--are getting hung up on a 25% chance spelling doom is more a sign of general innumeracy than anything else.)
But this gets us to the thing.  The thing here is that certain conservatives assume or want to argue that Nate Silver is wrong because he's biased and a tool of the liberal elite and whatever else; and if he's wrong they'll crow about how reality beat leftist dewy-eyed dreams (or are those "Dewey-eyed dreams"?) yet again, and if he's right... well, they'll say the election was rigged or stolen and perhaps accuse Nate Silver of creating self-fulfilling prophecies by propagandizing and misleading the public with his deliberate distortions and intentional misrepresentations.

Versus what liberals are likely to do if Nate Silver is wrong: however they may feel about the election results and whether the numbers were the product of things like voter ID laws or just the President losing out to a rough economy, liberals are likely to look at Silver's numbers and ask what's wrong with Nate Silver's model, not what's wrong with reality.

That is to say: I'm not sure I want to call what Nate Silver does "science", but there is a scientific approach in what he's trying to do.  And when you have a predictive model that doesn't work out, the scientific thing to do is to figure out what's wrong with the model--can it be salvaged or does it need to be abandoned and we start all over again?--instead of insisting the model works and making excuses for what is merely an "apparent" failure.

It seems to me there are ingrained patterns here, perhaps rooted in the difference between a conservative regard for tradition and a liberal regard for experimentation: that when a conservative fails, the answer is he should have done the same thing, only more of it; whereas when a liberal fails, the answer is to break the whole process down and figure out what went wrong and do things completely differently next time.  I tend to find the latter process superior, but then I guess I would.

This tangentially gets us to a problem with an argument Brother Seth recently made; he wrote: 

The best reason for liberals to throw the election to Romney, as I said Friday, is to hasten the apocalypse. Just throw everything to the Republicans and let them govern, unimpeded, until the country is a scalded ruin and the image of what conservatives would do with power is so branded on the American soul that nobody ever votes for them again. That’s a strategy. Hoping to fall back into the loser’s position so that liberals behave more like Matt Stoller wants them to is not. [internal link omitted]

Which would be more tempting argument but for the fact we've been through that several times now.  Regardless of what one thinks of conservatism as an ideology or governing philosophy, one might think George W. Bush, the 104th Congress, Richard Nixon, Joseph McCarthy, Herbert Hoover, the Teddy Roosevelt-led Progressive insurrection or even the Grant Administration could have been a death-knell for the Grand Old Party, but they've come home more times than Michael Myers and his Shatner mask.
Even if we confine ourselves to the postwar Republican Party, or even the post-Vietnam Republican Party, which are wholly different species of elephant from the parties of Hoover, Coolidge, Taft and Grant (which were different parties from the party of Lincoln), you're talking about the GOP recovering from the embarrassments of Nixon, Iran-Contra, Gingrich, De Lay, et al. when any similarly-scaled humiliations and/or scandals would put a comparable party in a parliamentary democracy into a permanent minority.  Not only that, but the wilderness periods, such as they were, have gotten shorter and shorter.  They rehabilitated Nixon, for fuck's sake, which should have been harder than populating Mars.

Well, it's because there's a mentality, held, apparently, by roughly half the population, that believes the real problem with Nixon was that he wasn't Nixonian enough, that if he'd really had the fullest latitude to smash down the greasy hippies, it would have been awesome.  Or if you want to bring us up to date, that the problem with George W. Bush was that he was insufficiently conservative, that bailing out the banks was socialism and No Child Left Behind a flawed set of programs because of underfunded mandates and a misplaced reliance on testing as a success metric, but because it violates the Tenth Amendment.

There is no apocalypse the Republican Party could cause that would shake the blind faith of these folks.  If President Obama, reelected, somehow caused the zombie apocalypse, liberals would start second-guessing everything they'd ever done and have a massive crisis of faith and probably schedule several circular firing squads.  If Romney triggers the Zombocalypse, however, conservatives will just blame Obama, or Nancy Pelosi, or maybe even Ted Kennedy; there's nothing wrong with the principles, so if everything got bunged up, it must be someone else's fault.


Randy Tuesday, October 30, 2012 at 7:08:00 AM EDT  

"Versus what liberals are likely to do if Nate Silver is wrong...."

It kind of looks like you forgot your point of just a few paragraphs prior at that point. Just what do you mean by "wrong" here? Because the obvious thing would simply be, Romney winning. But as you said before, there's about a 1/4 chance of that happening with Silver's model, which is hardly two sigmas or anything like that. It would just mean that the dice came up the less likely way, that time. But, if that's not what you mean by "wrong", then what?

Janiece Tuesday, October 30, 2012 at 7:46:00 AM EDT  

This exact lack of accountability (either internal or external), is one of the reasons I've abandoned the GOP permanently and with extreme prejudice. I wish the whole system would just die in a fire.

Eric Tuesday, October 30, 2012 at 12:05:00 PM EDT  

Randy, that's an excellent point and I phrased that badly. (Possibly because I'm an idiot, possibly because the piece was written in two chunks.) I'm leaving the poorly phrased version in place for the blogistorical record, but would ask readers to observe that point, which is important: probabilities aren't certainties.


Nick from the O.C.,  Tuesday, October 30, 2012 at 10:03:00 PM EDT  

"Probabilities aren't certainties." Yes, indeed. Which is how somebody can suck out and put a bad beat on you at poker, even though s/he had only a couple of outs. Thus, poker teaches us all that the most unlikely of probabilities occurs at approximately the probability of occurrence.

Or something like that. I did really poorly at stats in college. I figured the odds were against me ever needing to use that math-y stuff.

Post a Comment

Thank you for commenting! Because of the evils of spam, comments on posts that are more than ten days old will go into a moderation queue, but I do check the queue and your comment will (most likely) be posted if it isn't spam.

Another proud member of the UCF...

Another proud member of the UCF...
UCF logo ©2008 Michelle Klishis international gang of... international gang of...
смерть шпионам!

...Frank Gorshin-obsessed bikers.

...Frank Gorshin-obsessed bikers.
GorshOn! ©2009 Jeff Hentosz

  © Blogger template Werd by 2009

Back to TOP