Tuesday, October 23, 2012

2012 Election

FLG hasn't been writing much about the election, but here's the deal.  He's recently swung from the Neither column to the Romney column.  Given that FLG has voted for the winning candidate in every presidential election he's ever participated in, this bodes well for Mr. Romney's chances.

In fact, after the debates, FLG puts the probability of a Romney win at about 55% (if you really pressed him he might even say 60%), which means he's seriously considering putting some money in those prediction markets.   They are way overconfident about an Obama win.  Last time he looked, they've got Obama at 55 and Romney at 45.

10 comments:

The Maximum Leader said...

Wow! Wagering on a close election. I like your style. I am too unsure to go beyond my typical $1 bet... BTW, I still think that the President will be re-elected. Electoral math favors him (for the moment at least).

Withywindle said...

FLG: What decided you?

Anonymous said...

The trend is Romney. The momentum is Romney. The obtuse undecided will osmotically move to Romney. Obama will lose comfortably.

If I'm wrong I'll double down on: End of the country as we know it - 'Under Four Years' at 2 to 3.

George Pal

The Maximum Leader said...

I figured out my average for voting for the winner of the Presidential Elections in which I have been eligible to vote. My number: .500.

My votes since 1988: Bush, Bush, Dole, Bush, Bush, McCain.

It really upsets me that I've voted 4 times for Bushes... Variety!

The Ancient said...

In the last few days, I've finally felt that it was possible for Romney to win.

But the electoral odds still favor Obama.

Side bet: If Obama squeaks back in, he will behave as if he had received a huge mandate. He will do more and more things administratively (w/o legislation). Foreign policy will lurch even more sharply to the left. The economy will continue as before, with occasional upward sputtering, but no better. Islamic terrorism will continue to grow, and there will be at least one major attack against the U.S.

(This -- I would wager -- is also what the Clintons believe, which is why they are rooting for Romney.)

Andrew Stevens said...

Romney's back down to 40%. Now's the time to buy.

The Ancient said...

Andrew --

Paul wrote this just for you...

http://krugman.blogs.nytimes.com/2012/10/25/triumph-of-the-electoral-nerds/

(If I *really* thought Romney was going to win, I couldn't resist going in heavy. But I don't. At least, not yet.)

Andrew Stevens said...

Krugman is a moron who has no idea how to read the polls. This is also true of Nate Silver. That a poll says Obama is up 48-46 in Ohio just doesn't mean diddly beans. There is no poll where that isn't within the margin of error. That Krugman and Silver take very seriously tiny margins for Obama in a few states on polls which are reaching less than 10% of the people they try to contact (and just assume that the people they do contact are representative of everyone, with no evidence backing this assertion), all of them using questionable likely voter models, and radically different methodologies, and with no attempt made to assess the merit of the methodologies of the polls.... Well, it just goes to show the power numbers hold over people who have no real idea what they mean. I'm a mathematician. Numbers do not hold the magical power over me that they do over most non-mathematicians.

The election is simply too close to call. Obama never had much of an edge, not even when all the "experts" were calling it for him (when he was up by 3, half the margin Bush was beating Kerry by throughout 2004, when everybody talked about what a close election it was going to be), and that edge has now all but evaporated. I should say that I have placed no bets. 40% for Romney? Eh, that's close enough to right that I wouldn't bet much on it. Plus, apparently Romney's price spiked 7 points when somebody put as little as $17,800 into him the other day. It's impossible to take seriously such a thinly traded market nor would I bother to make any kind of significant bet on it. I'd just buy stocks I'd expect to shoot up if Romney wins and sell short the ones I'd expect to do poorly if Romney wins.

Nate Silver has Romney at 29% and I'd be happy to pile into that bet if anyone were to offer it to me. I'm not at all confident Romney is going to win, but he certainly has a better chance than that.

Andrew Stevens said...

I'm just going to give a quick example of a poll, which is being used by Real Clear Politics in their Ohio average. ARG released a poll today showing Obama winning Ohio 49-47.

Obama's winning Democrats 87-9, Romney's winning Republicans 88-10, and independents are going for Romney 57-36. So how is Obama winning this poll when even it shows he's getting slaughtered among independents? Literally nobody has won a major statewide race in Ohio in recent years while losing independents at all, never mind that badly. Well, Obama can win because the poll is projecting 43% of the Ohio electorate to be Democratic, 34% Republican, and 23% independent. Well, maybe that will happen.

But in 2008, it was 39% Democratic and 31% Republican. In 2004, it was 40% Republican and 35% Democratic. In 2010, it was 37% Republican and 36% Democratic. So ARG is expecting 2012 turnout to be even more heavily Democratic than in 2008, so much so that Obama can get crushed among independents and still win. Who knows? That could happen, but it's important to understand that the polls which show Obama winning Ohio expect that to happen. Obama is losing badly among independents in virtually all the major polls so he needs massive turnout from Democrats to win, better than he had in 2008. If the turnout looks more like 2004 or 2010, he doesn't even have a chance in Ohio. Romney will win it in an absolute walk (based on current data, that is).

Anyway, it's important to know what the polls are saying. If you were trying to predict the election and had unlimited resources, would you really get dozens of polling organizations, have them all use their own wildly divergent methodologies to attempt to get a random sample and predict who will actually vote, and then average them all together in a big stew? Of course you wouldn't. It's crazy. And yet we take seriously people whose models are doing exactly that.

The Ancient said...

"apparently Romney's price spiked 7 points when somebody put as little as $17,800 into him the other day."

Andrew --

Are you saying that, in principle, someone with a few million to burn, could completely control that intrade result?

Because if he could, the psy-ops value alone (among the commentariat and elsewhere) would be worth a lot more more than the same money spent on ads...

 
Creative Commons License
This work is licensed under a Creative Commons Attribution-No Derivative Works 3.0 United States License.