« The End of TABOR? | Main | "Minutes upon minutes of soul-searching" »

November 09, 2006

Brave New World

With a Democratic Senate in it. Blogging will be light today -- unless some of the weekenders want to jump in with their incisive post-election thoughts -- as I'm crashing on a post-election article before the close. Elections really suck for monthly political magazines, particularly when they fall four days before the end of your close cycle. For the moment, though, go read Atrios on how the betting markets failed to predict the Democratic Senate. As he writes, "the idea that there's something magical about market aggregated preferences seems to have infected the minds of too many people. They provide a cute distillation of conventional wisdom, but that's all."

Also, a quick bleg: Anyone know of any articles detailing Rove's plan for an enduring Republican majority, possibly focusing on policies like tort reform and paycheck protection that weaken the Democratic side?

November 9, 2006 | Permalink


what was bizzare about tradesport was that it predicted each of the races that we won, but then didn't predict an overall take over of the Senate. Proving that the system isn't always rational or logical.

Posted by: akaison | Nov 9, 2006 11:41:15 AM

Atrios's dismissiveness is unwarranted. The markets didn't say the Republicans *would* keep control of the Senate, they said there was a 70% chance the Republicans would keep control of the Senate. Events with 30% chances of taking place happen really quite frequently -- 30% of the time, in fact. If you think the markets are way off in their probability assessments, put your money where your mouth is and invest in the alternative. If you can't consistently make money off the markets, then they're doing a better job predicting the outcomes than you are. Put up or shut up.

Posted by: Tarendipitous | Nov 9, 2006 11:43:22 AM

Anyone know of any articles detailing Rove's plan for an enduring Republican majority

John B. Judis and Ruy Teixeira did a treatment of Rove and his 1896 McKinley delusions in the Emerging Democratic Majority (2002)

There is a review of the book (initially published in Amercian Prospect in Summer of 2002 - I remember commenting on an aspect of it on Aug 7, 2002 in a letter to the Tapped editor - the book quotes Rove as saying that the 1896 ticket was erroneously McKinley/Roosevelt) which appears to be reprinted at this site:


Posted by: Andy | Nov 9, 2006 12:01:22 PM

Did you hire Yglesias' "editor" for this uncharacteristically poorly typed post, Ezra?

Posted by: Haggai | Nov 9, 2006 12:05:53 PM

I agree with Taren...

Two of the Senate races were virtually tied. If you say the Dems really had about a 50% chance of winning each of them, and they needed to win both of them to get control of the Senate, that translates into a 25% chance of winning. Changing the chances to some slightly higher number (55% or so? I don't feel like doing the arithmetic) will yield something like a 30% overall chance.

This gets to the logical fallacy in akaison's post. If the Dems have a 60% chance of winning each of two races, then tradesport should predict they will win each of them. As for a prediction of winning BOTH of the races (assuming independent probabilities), tradesport should predict they won't (since .6 * .6 = .36).

Posted by: Jim W | Nov 9, 2006 12:07:32 PM

I blame Neil for the OH-2 loss...

Posted by: Petey | Nov 9, 2006 12:13:38 PM

Someone ought to write an '06 article entitled: What's Not Wrong With Kansas...

Posted by: Petey | Nov 9, 2006 12:16:33 PM

I agree completely with Taren and Jim. If the Democrats had won those Senate races with thumping majorities, then you could complain about the "market" getting it wrong. A little more rain or another bad Kerry joke and Burns and Allen could easily have won. The chances of all factors coming together for the Dems to take six Senate seats was unlikely and the market reflected that. In fact, the market's prediction was based on the same kind of polling data you praised yesterday.

Posted by: Joseph Hovsep | Nov 9, 2006 2:02:49 PM

A pitch for a snarky analytic angle to your Rove-up: Rove, though claiming to have been working a "long con," was actually working a short con. In fact, the very claim to the long con (building the "enduring/permanent Republican majority") was essential to his short con.

I think Rove always knew that the tide of modernity was against him, and that he to get what he and George could while the getting was still good. (This dovetails not only with Rove's career-long investment in short cons, but also with George's business resume, which is littered with short cons masquerading as long cons; Josh Marshall has an excellent late-election-night post where he said he expected Bush to announce on Wednesday that he'd sold the US to a consortium of his father's buddies.)

Now, this is *not* to come up with a creative spin on the "No, Rove really IS still an evil genius, and it was all part of his plan!" Cassandra cry. It's merely to emphasize that Rove is and always has been a con man (as has George), and that his "30-year plan" was most likely nothing more than hot air.

Yes, perhaps Rove finally seduced himself into believing his own pitch (as many con men do when their cons become *too* successful through accidents of circumstance) -- but it was always a con.

Posted by: The Confidence Man | Nov 9, 2006 2:06:42 PM


the problem with that probability is that it still doesn't make much sense because it takes nothing into account other than multiplying two probabilities together, and is therefore not representative of any real sussing of the real chances involved. winning races isn't a roll of the dice- there is significantly more involved- hence the reason why politically it didnt make logical sense.

Posted by: akaison | Nov 9, 2006 2:12:01 PM

It doesn't matter how many factors contribute into the outcome of winning a race. Based on the knowledge available at the time (i.e., polls), it is reasonable to come up with an estimate that the Democrats had about a .5 or .6 chance of winning each race in Virginia and Montana. Since the outcomes in these races are largely indpendendent of each other, then it is reasonable to multiply these probabilities to get the probability of them winning both races.

Its easy to reason after the fact and say that since it ended up a particular way, that is the way it had to end up. But the information available at the time suggested that the races would be close, as they were.

Posted by: Jim W | Nov 9, 2006 3:19:22 PM

"Since the outcomes in these races are largely indpendendent of each other, then it is reasonable to multiply these probabilities to get the probability of them winning both races."

While I disagree with akaison's position, it's worth noting that the races really aren't fully independent. Much as they did in '00, '02, '04, and '06, it's not unusual for a bunch of Senate races to all break in the same direction.

I thought TradeSports was undervaluing Dem chances in the Senate market, though I didn't think the undervaluation was drastic. And this was probably due to not fully factoring in the 'all close races fall the same direction' factor.

I thought fair value for Dems winning the Senate on election eve was about 35% - 40%.

Posted by: Petey | Nov 10, 2006 12:43:03 AM

The comments to this entry are closed.