Slog: News & Arts

RSS icon Comments on Intrade: Obama Wins

1

So in using this information, you're implying that efficient, free markets work, right? That seems slightly hypocritical, given the daily trashing of capitalism and free markets that occurs on Slog.

Posted by john cocktosin | June 11, 2008 11:13 AM
2

Yeah Al Gore and Kerry were predicted to win also and look what happened. At least they have more experience and substance than that rhetoric-filled non substantive Odrama.

Posted by Odrama | June 11, 2008 11:15 AM
3

There was a period of several weeks when a small group of highly motivated Ron Paul fans had him trading at over $15, which is obviously laughable. Eventually, saner traders came along and took their money, but it didn't happen right away. I think you need to factor in trading volume before you can claim Intrade is making a worthwhile prediction.

Posted by elenchos | June 11, 2008 11:16 AM
4
Because, presumably, each investor is integrating a lot of information when deciding to buy a contract or not ... Even if the individual investors are each biased, the overall error price should be normally distributed, if the central limit theorem holds.

Wrong.

The problem here is that you're basically assuming that each person's bias and error is independent and identically distributed. Neither of these assumptions are necessarily true - the potential for groupthink in the population breaks the independence assumption, and the i.d. assumption is broken by the fact that certain inherent cognitive biases in humans will cause people to often fail in similar ways.

People are not rational, even on average.

Posted by tsm | June 11, 2008 11:24 AM
5

tsm--

You are totally right! I think some of the assumptions about these future markets are faulty and that the central limit theorem probably doesn't hold.

In the least, I think the independence of the investors can be called into serious question, as they are probably mostly drawing from the same biased information sources when making their decisions.

Still, a pretty interesting way of "predicting" the 2008 general election.

Posted by Jonathan Golob | June 11, 2008 11:29 AM
6

Not only that, but you're basing this not on whether you're going to vote for a candidate, but whether you think other people are going to vote for that candidate.

It seems like liberals like us specialize in coming up with fancy and complicated ways of convincing ourselves that certain things are true.

Posted by demo kid | June 11, 2008 11:30 AM
7

People still read slashdot?

Posted by w7ngman | June 11, 2008 11:31 AM
8

efficient market hypothesis is bunk as an actual state of affairs. But the more participants in a market the more efficient the market is.

Posted by Bellevue Ave | June 11, 2008 11:32 AM
9

Whoops - "similar ways" in @4 should be "different ways". Anyways ... yeah.

Posted by tsm | June 11, 2008 11:37 AM
10

@1: Markets work great for some things and not so great for others. Sure, the concept sounds simple, but it's embarrassing how few libertarians are able to grasp it.

Posted by shub-negrorath | June 11, 2008 11:41 AM
11

You can get the same information from polls. See http://www.electoral-vote.com/

Posted by F | June 11, 2008 11:42 AM
12

Let's hope the red and green game and the red green show don't blow out the olympics as warned about with the coming of George Bushes Clown detecter and the Chinese tariff wars leave the next Senate paige "out on the tiles....."

Any body remember that song from the led zepplin british hey day?

Posted by anonymus | June 11, 2008 11:45 AM
13

@12 "when I walk down the highway all I do is sing this song"

Posted by LZ III | June 11, 2008 12:46 PM
14

Intrade, IEM et al are great ways to get the current CW but they absolutely blow for anything else.

Take a look at the markets for the primaries this year. The post-Iowa/post-NH seesaw should dispel any notions of them being able to predict what happens.

Better yet: look at how they performed in 2004.

Posted by lol | June 11, 2008 12:57 PM
15

I don't buy it.

The African-American turnout is going to be huge, which is going to be a big factor in Georgia, Louisiana, Mississippi and South Carolina; all states with populations that are over 30% African-American. People that have never voted before, esp. the young, are going to turn out in droves for him.

Add in the fact that the Christian Right is lukewarm on McCain means that the South is not necessarily a done deal for the Republicans this year, at least the parts with large black populations.

Posted by michael strangeways | June 11, 2008 1:11 PM
16

I am really cynical on most things but honestly this map is missing A LOT of blue on it....

Unless Obama is found in a sex tape fucking a 10 year old cousin he has a landslide coming to him. Hell, him fucking a 10 year old cousin will probably lock him in to victory in the south.

Posted by Cato the Younger Younger | June 11, 2008 1:28 PM
17

Don't forget to factor in the October Surprise Bombing of Iran ...

Posted by Will in Seattle | June 11, 2008 1:32 PM
18

I think the point of these market based predictions isn't that they are always right, but that they tend to be more accurate than what any one person thinks. The vegas betting lines are much more predicitive of sporting events than what some ESPN columnist thinks, and the same probably goes for politics. Saying "they got Kerry-Bush wrong" is kinda like dismissing global warming because of the crappy weather we had for the past week.

Posted by Josh | June 11, 2008 1:34 PM
19

I wonder how much of that lean towards blue CO and NM are based on a market forecast of Richardson for Obama's VP.

Posted by K | June 11, 2008 1:40 PM
20

Well, Richardson is the top choice in the Stranger poll, and my second choice, so that's not a bad prediction ...

Posted by Will in Seattle | June 11, 2008 1:51 PM
21

@17: No, the October Surprise is Bin Laden is found dressed as a woman in a Kandahar marketplace.

You know, Bin Laden? That guy? With the videos and the terrorists and the beard?

An aside; What's with the blue Iowa? Really? Is this banking on Obama's Illinois history?

Posted by K | June 11, 2008 1:55 PM
22

I'm following this one:

http://www.demconwatchblog.com/2008/06/this-is-demconwatch-general-election.html

Get ready for President McCain, when he takes Michigan and Ohio...

Posted by MarkyMark | June 11, 2008 2:00 PM
23

Exactly Josh. Well put.

Posted by Jonathan Golob | June 11, 2008 2:01 PM
24

Ach! The central limit theorem holds but SO FUCKING WHAT?! The mathematical statistics only tell us about the sampling distribution of the mean of what people are predicting. So, now we can avoid having to take a census of the US population to find out what the whole country thinks about the future. We still know zero zilch zip about the accuracy of the predictions. That comes from outside the mathematics. Ach.

Posted by umvue | June 11, 2008 2:27 PM
25

I think it's a bit dumb to put too much stock into a website/betting pool that is probably used by a predominately white, upper middle class, older skewing and male clientle.

This presidential race is going to be hugely impacted by non-whites, the young and women.

Note to K: Iowa, and much of the upper midwest is not as conservative as the southern half of the midwest. Iowa has a strong Democratic base and both Nebraska and Iowa frequently have Democrats in Congress. Obama beat both Clinton and McCain in North Dakota, a state with a minuscle minority population.

Posted by michael strangeways | June 11, 2008 2:28 PM
26

@21 Iowa went for Dukakis. It's not a Republican state.

Posted by Eric F | June 11, 2008 2:34 PM
27

Not only that but ... they didn't even use the Intrade results correctly

Posted by RonK, Seattle | June 11, 2008 5:14 PM
28

I want to see how much these numbers change - not just in Intrade - after the first Obama/McCain debate . . . Daily Kos (who predicted both his Senate and nomination wins wayyyy early) already shows just about everything moving slowly in Obama's general direction (pun intended).

Posted by Zelbinian | June 11, 2008 7:40 PM

Comments Closed

Comments are closed on this post.