Sunday, October 12, 2008

Intrade is Not An Opinion Poll, It's a Prediction Mkt.

Bobble comments: I don't put much stock (pun not intended) in Intrade. I find realclearpolitics.com is a far better source of how the candidates are doing. It computes averages of all the major polls and publishes the results in easy to comprehend format (see chart above).

MP: The average of the polls shows Obama ahead of McCain by 49.7% to 42.4%, while Intrade odds are 77.1% for Obama and 22.6% for McCain. Why the huge difference?

Because they are measuring something completely and totally different. The opinion polls are based on surveys of likely voters say who have stated which candidate THEY will likely vote for. The Intrade betting is not based on who traders will vote for themselves, and it is not based on who the traders WANT to win the election (or not win), the betting is on who is MOST LIKELY TO WIN THE ELECTION!

I could be a strong Obama supporter or a strong McCain supporter, or I could hate Obama or McCain, or be completely indifferent, and none of those positions about MY OWN voting preferences would influence what position I would take on Intrade. My position on Intrade is based on how I expect millions of other people to vote.

Bottom Line: Given that Obama has a 7.3% lead in the polls, the trading on Intrade suggests that Obama now has a 77.1% chance to win the election. That's not saying that Obama will get 77.1% of the vote, just that he has a 77.1% to win the election.

As some others have pointed out, if you think Intrade is flawed, inefficient or deficient in some way, that means you can make lots of money on Intrade by exploiting those inefficiencies by betting against the collective wisdom of the masses who have money at risk when they take a position on Intrade. Good luck.

6 Comments:

At 10/13/2008 4:49 AM, Blogger nigeleccles said...

Well put.

The bizarre thing with Intrade however has been the price difference between it and other exchanges such as Betfair. Over the past 3 weeks Intrade has been consistently 8-10% more in favor of McCain.

Currently we (and the UK exchanges) have Obama at 82-84%: http://www.hubdub.com/m17795/Who_will_win_the_2008_US_Presidential_Election

 
At 10/13/2008 11:30 AM, Anonymous Norman said...

From what I remember about 2004 you could have made good money betting against Intrade so I don't think the 'wisdom' of the markets works here because Intrade reflects the MSM and the polls; the latter is biased and the former has been very inaccurate and influenced by the MSM.

 
At 10/13/2008 11:50 AM, Blogger nigeleccles said...

@Norman

You can't bet against Intrade. It is a bet exchange and doesn't have a position. You could bet against the favorite however in 2004 the favorite won.

However, you also say you can't trust polls or the media either. What do you suggest? Star signs?

 
At 10/13/2008 12:37 PM, Blogger bobble said...

"Because they are measuring something completely and totally different"

agreed.

i do sometimes think about fading some of the economic bets available on intrade. for instance will the US go into a depression in 2009?

it's 6 bid 14 ask. if i could get it for 10, that's a possible 10 to one payoff. not a bad speculation, considering the trainwreck we are dealing with right now. LOL if i lose, i win, because i benefit from the better economy. if i win, i'll *really* need the money.

 
At 10/13/2008 4:53 PM, Anonymous qt said...

It seems not everyone is convinced . May the best man win!


In Canada, tomorrow is E-day and we will be free of election rhetoric :) Happy days! Whoopee!

 
At 10/15/2008 6:43 PM, Blogger OBloodyHell said...

Part of the question, though, is how accurate they can be when the information pool is poisoned by the MSMs, to say nothing of defective poll design (which certainly acts to reinforce notions as to the actual state of the situation, rather than to correct them)

AJ Strata analyzes the polls, and finds them wanting. His analysis: While Obama almost certainly has a lead, it's far more tenuous than the most touted polls suggest.

DJ Drummon also makes a case. He does a state by state analysis and suggests that GOP turnout will be crucial, especially in certain key states like Florida, Colorado, and a couple others. If the GOP turns out in its typical numbers, it means a McCain win.

...And the media almost certainly realize this, which is why they are doing their best to discourage GOP turnout.

Maybe they're wrong. I'm not claiming they aren't. But, if they are correct, it means you -- and every other McCain supporter you know -- are going to be kicking themselves if it turns out close and you DIDN'T go just because you believed them when they told you "all was lost".

So, vote, DAMMIT.

 

Post a Comment

<< Home