Was Intrade being manipulated over the last month?

http://www.overcomingbias.com/2012/11/was-intrade-being-manipulated-over-the-last-month.html

by randomwalker • 14 years ago

51 29 14 years ago

gojomo• 14 years ago

Over betting markets as a whole, it's hard to distinguish 'manipulation' from simple 'overconfidence' by a certain side.

The difference between InTrade and Betfair over long periods is the large mystery... though given the lags/fees in getting money in, and the fact the markets aren't that large -- so maximum profits aren't very high -- seem likely to be part of the answer.

There's a reasonable Robin-Hansonian-case to be made that attempts at prediction-market manipulation improve market accuracy over the long term, by increasing the rewards to those taking sides against the manipulator.

hooande• 14 years ago

The author brings up some good points about the potential for arbitrage between prediction markets. But in my experience, intrade has a strong conservative bias. This is more likely a case of people betting on their (incorrect) political convictions as opposed to intentional manipulation.

I remember observing a similar thing around the betting on the supreme court healthcare decision. The odds were 3-4 points more against the legislation on intrade than they were on other sites. The comments were overwhelmingly partisan and conservative. "The Supreme Court will find this unconstitutional because Obama is a socialist."

The idea behind prediction markets is that people will only bet when they know something that the rest of us don't. But we shouldn't underestimate the human ability to convince ourselves that everyone else feels as strongly about an issue as we do.

CamperBob22• 14 years ago

In addition to what you're saying, it could have made sense to buy Obama calls, so to speak, if you think your taxes are going to go up significantly under his second term. That behavior would severely distort Intrade's purpose as a prediction market.

javert3• 14 years ago

No, it wouldn't. The "purpose" of a prediction market is to let people buy and sell shares in a certain outcome, not to provide information.

That's just a nice side effect of a sufficiently liquid market.

yummyfajitas• 14 years ago

Hedgers are giving money away (on average) in order to hedge against undesirable outcomes.

If speculators respond to incentives and take their money, markets will be pushed back to normal. I think Robin Hanson even has a paper or two proving this is the equilibrium strategy. As the article indicates, he even put his money where his mouth is.

modarts• 14 years ago

Markets are by in large not NEARLY as rational as conservative economists make them out to be. Your point about the political biases informing the positions that traders were taking further supports that.

001sky2• 14 years ago

Your use of the word 'conservative' is ill-advised, both technically and rhetorically. Free-market Economics is rooted in [l]iberal traditions in the first instance, not conservative (see: repeal of the corn laws)[1]; and in the second: Modern [L]iberals (ie, interventionists) require near-perfect rationality as an operating behavioural assumption, ex-ante. Not to be pedantic, but this is an important point. [L]iberal economists are part of the Orthodoxy. Non-orthodox behavioural assumptions[1] are (arguably) quite problematic to an Orthodox interventionist. This is glossed over because it threatens the 'myth of expertise' that is required to politically implement a such an ideology. This is not a partisian point, but one of history and terminology.

______________

[1] http://en.wikipedia.org/wiki/Corn_Laws.

[2] In particular, bounded rationality and opportunism.

edits: clarity

pg• 14 years ago

I've been wondering about this for week. Is there a statistical smoking gun anywhere?

jsdalton• 14 years ago

One could look at the state-by-state markets in the weeks leading up and see if the probability predicted differed greatly from the main market. It would be far trickier to manipulate the markets state by state in such a way that their probable outcomes aligned with the probable outcome of the national markets.

I just looked at the InTrade state numbers from 11/2 (which I have on a spreadsheet) and running a few dozen simulations I was getting Obama winning about 80%. That's a pretty significant difference from where the national numbers were on 11/2 (I think about 66% maybe?)

Of course, if they differ it might just be because the Intrade market is imperfect, in which case it would represent a nice arbitrage strategy for the next election...

andrewcooke2• 14 years ago

there's http://www.washingtonpost.com/blogs/ezra-klein/wp/2012/10/23... which was mentioned at http://fivethirtyeight.blogs.nytimes.com/2012/10/24/oct-23-t...

chart in first link shows large romney buyer. not sure if that's what you're looking for with "statistical smoking gun" though. a little more analysis at second link.

sire404• 14 years ago

No it's most likely just overconfidence in the market by a few individuals.

Intrade actually checked for manipulation themselves:

"We checked this out for potential manipulation - it certainly fit the pattern at first glance. However, during the period you reference above 40 individuals bought Romney shares and no individual bought more than 15% of the shares traded during that period. A look at those 40 accounts shows nothing suspicious. So we don't believe this was manipulation but more a run on Romney when the market was unusually thin."

http://www.theatlantic.com/business/archive/2012/10/should-p...

However, even without this spike the Intrade market was misaligned for more than 24 hours. A few of our arbitrage customers made huge amounts of money (outside our software) trading 5-10% arbs between Intrade and Matchbook. This is extremely uncommon.

Disclaimer: I'm one of the developers behind sports arbitrage software RebelBetting (http://www.rebelbetting.com)

Dn_Ab• 14 years ago

My friend and I also did some 'arbitrage', we almost didn't do it because we were sure we had something wrong, I mean, aren't markets supposed to be efficient? We even made a program that calculated optimal amounts and alerted us when the odds were aligned to a particular criteria taking into account exctraction fees, exchange rates, commission and stuff.

Our only explanation was that betfair (our other side) was mostly European while Intrade was mostly American so it is possible that lots of Intrade betters were buying into the media's attempt to try to make the race a lot closer than it was while the Europeans were not so biased. I don't know but it is a simpler explanation than someone purposely losing money because they think influencing a prediction market would affect the actual outcome?

javert3• 14 years ago

aren't markets supposed to be efficient

To a certain extent. Largely because of people doing arbitrage. So, if all arbitrageurs called it quits because they thought markets are automatically efficient, well, they wouldn't be.

Did you make much money?

zecho• 14 years ago

There's an economics joke as old as the hills:

An economist and a non-economist are walking down the street when the non-economist notices something on the ground and says, "hey there's a $10 bill!" The economist replies, "Impossible. If there was a $10 bill on the ground, someone would have picked it up already."

lincolnq• 14 years ago

I had a friend who made $10,000 (off a $200k bankroll) in 2008 doing this exact form of arbitrage...

bcoates• 14 years ago

One thing you have to watch out for on Intrade-Betfair arbitrage is the not insignificant risk of not doing the trades you think you're doing. I know Intrade has rules that various events will cause trades to be stopped and completed trades will be undone. Betfair might as well, but the two markets don't have any synchronization on when or why.

Since you'll constantly have many times what you expect to gain in a day at risk on the markets, a single rollback event could put you very deep in the red.

maxerickson• 14 years ago

I think a greater number of U.S. participants on Intrade is one of the better explanations.

I wouldn't be surprised if personal biases play a bigger factor than the media.

cypherpunks012• 14 years ago

Nate Silver wrote about this twice:

http://fivethirtyeight.blogs.nytimes.com/2012/11/06/live-blo...

http://fivethirtyeight.blogs.nytimes.com/2012/10/24/oct-23-t...

qq66• 14 years ago

If a prediction market is famous enough to affect an election, it will generally cost more than $1 million to manipulate. I doubt that voter turnout would have increased if Intrade had predicted a Romney win -- too few Americans know what Intrade even is.

bodyfour• 14 years ago

I'm sure all political reporters know what Intrade is. If you were going to spend money to move the market, those are the people you're most trying to influence.

I'm still personally skeptical of the whole manipulation theory, however. It's probably just a combination of too many conservatives buying with their heart instead of their head, and too few active arbitrageurs taking them to the cleaners.

colonel_panic• 14 years ago

One Intrade commenter from London claimed he was set up to earn about $20,000 in arbitrage against Betfair. If you expect the manipulation to continue over a period of time (which I certainly did, and wished I could register on other betting sites), it's worth it to deposit larger sums of money with bank transfers despite the wait.

At one point I wondered whether the manipulators were making large purchases once per day and whether this could explain why the different betting sites never came back in line with each other. In that case, a savvy arbitrageur might take advantage of prices while they're at their widest differential, and then wait to buy more the next day. I thought I saw a pattern like this going for a few days in a row but it was gone by the time I felt confident enough to trade around it.

cypherpunks012• 14 years ago

A commenter over there mentioned that a similar phenomenon was noticed during the 2008 presidential election, and quoted John Delaney (Intrade founder, now deceased):

Mr. Delaney conceded there had been erratic behavior - including spikes in the direction of Mr. McCain and away from Barack Obama "by up to 10 points." And he said "trading that caused the unusual price movements and discrepancies was principally due to a single 'institutional' member on Intrade."

But, after interviewing members of the institution involved and tracking its trades, Mr. Delaney wrote in an e-mail message on Saturday: "I do not believe based on our investigations and over seven years' experience in such matters that the trading on Intrade was designed or motivated to artificially inflate McCain’s chance of winning."

http://www.nytimes.com/2008/10/20/business/20predict.html

andrewcooke2• 14 years ago

could this be hedging against a romney win? for example, if you're an investor in green tech, then intrader shares in romney would be a way to protect yourself.

but i imagine the numbers are too small?

Was Intrade being manipulated over the last month?

Related Stories