Posts Tagged ‘Nate Silver’

Palin’s Still a Longshot, And Nate Silver is Overrated

Friday, January 8th, 2010

by Brien Jackson

Nate Silver writes:

I find this truly remarkable: in a National Journal survey of 109 Republican “party leaders, political professionals and pundits”, not a single one deemed Sarah Palin to be the most likely Republican nominee.

I’ve written extensive commentary about how I think Palin’s chances are in fact pretty decent. I’d probably call her the “favorite”, although “favorite” in this context might mean having a 25-30 percent chance of winning.[…]

But back to the point I made in November — there’s going to come a time, probably in July 2011 or so, where the knives are really drawn on Palin and Republican pundits, strategists and candidates start saying in public some of the things they’ve been thinking in private. And that in all likelihood will play very well for her. Although the Establishment’s concerns about Palin’s viability as a general election candidate are well grounded, mostly they’re just terrified of her because she doesn’t need them.

This largely misses the point. The problem with having no “establishment” support for your candidacy is that the establishment contains pretty much everyone with any idea how to run a campaign. Particularly the national campaign it takes to compete in a Presidential primary. We’ve spent a lot of time mocking Palin’s current staff, and some of the rather comical missteps they’ve made, and that’s just in booking speaking events. How well is she going to manage a year’s worth of on the ground campaigning in multiple states/regions if the same people she has arond her now are in charge of managing things? Moreover, the Clinton campaign proved that even professional operatives can be bumbling idiots, how well is a group of amateur bumbling idiots likely to do? And popularity alone, especially popularity within a narrow subset of the electorate, isn’t a substitute for a functioning campaign. Ask Fred Thompson.

But a larger point that needs to be made here, at the risk of sounding like a politco version of Mike Silva, is that Nate Silver actually doesn’t seem to understand that much about politics. He understands statisitcs and polling, to be sure, but that only takes you so far. His schtick, basically, is built on taking polling data and plugging it into a computer model that runs a lot of times, which tells you the most likely outcome. And that’s great, but it’s also a lot of work that really doesn’t add much value to the task of predicting election outcomes. I didn’t have a sophisticated compter program running thousands of scenarios, but I still managed to come pretty close to the ultimate outcome of the 2008 cycle; missing only Missouri in the Presidential contest, correctly predicting every Senate and Gubernatorial race, and missing by 6 seats in the House. Which isn’t (totally) to toot my own horn, so much as it is to point out that sometimes there’s a level of obviousness involved in last second pick ’ems, and you don’t really need computer simulations to figure this stuff out, particularly when the election isn’t very close. Perhaps Silver’s computers add value in an unusually close contest, but those are, well, unusual.

There’s nothing wrong with Silver’s blogging mind you, but he’s much better when he’s analyzing polling and numbers, as opposed to trying to analyze non-statistical aspects of politics. Which isn’t to say he can’t toss his two cents out by any means, but I wish more “A-Listers” would keep that in mind a bit when reading him. I mean, Silver’s models still give Democrats a better shot to retain Blanche Lincoln’s Senate seat in Arkansas┬áthan Joe Biden’s old seat in Delaware. Perception and conventional wisdom can often prove to be wrong, and polling can, over time, show where it’s wrong, but that has to pass the smell test too, especially this far out of an election, with little to no actual campaigning having occured, and with very little data to work with.