Comments on: Put your money where your mouth is

Put your money where your mouth is

DynamiteToast — Thu, 25 Oct 2012 08:33:01 -0800

Intrade is a Prediction Market, where you make predictions by buying and selling shares on the outcome of real-world events. These events are always defined on Intrade as a YES/NO proposition. Shares are bought at some point between $0.00 and $10.00, based on whether the buyer believes the event will or won't occur (which correspond to $10.00 and $0.00 respectively). Most popular propositions at the moment are election related, though this week the market for the Best Picture opened.

Generally, Intrade garners attention for it's mix of clear prediction values (a multiplication of 10 by any given "price" easily gets you to an option's percentage chance of happening) and market driven prices which allow it to react more quickly than polls do to current events such as debates. However, earlier this week something odd happened in it's presidential market.

At several points on Tuesday and early Wednesday, what appeared to be a single trader bought a large number of Mr. Romney's shares at Intrade, at one point boosting Mr. Romney's chances to about 49 percent from 41 percent over the span of a few minutes. The betting patterns (Previously) echo similar ones in the pricing of John McCain and Mr. Obama's stock at Intrade late in the 2008 cycle.

By: grobstein

grobstein — Thu, 25 Oct 2012 08:34:55 -0800

And then the prices re-equilibrated rapidly as speculators rushed to take the free money from the would-be "manipulator," who was quickly out thousands of dollars for no benefit.

By: RolandOfEld

RolandOfEld — Thu, 25 Oct 2012 08:36:50 -0800

If at first you don't succeed, try, try again.

By: DynamiteToast

DynamiteToast — Thu, 25 Oct 2012 08:37:16 -0800

By: jaduncan

jaduncan — Thu, 25 Oct 2012 08:38:37 -0800

And then the prices re-equilibrated rapidly as speculators rushed to take the free money from the would-be "manipulator," who was quickly out thousands of dollars for no benefit. Save headlines, of course. Not even adverts. Editorial about 'hey, that Romney has a chance'. The cynic might take the view it was cheap at the price.

By: sallybrown

sallybrown — Thu, 25 Oct 2012 08:39:53 -0800

Re: the link to Ezra Klein--he posted an update:

Update: Over at the Atlantic, Carl Wolfenden, the exchange operations manager for Intrade, says, "We checked this out for potential manipulation—it certainly fit the pattern at first glance." But, he says, that doesn't seem to be the case this time around. The Romney blip was apparently driven by a number of traders during an early-morning period when the market was very thin, rather than just one person.

Carl's full comment reads:

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.

Could still have been coordinated, just by multiple people. What I don't understand is with the massive amounts of money floating around, why no one has tried to game the system on a more massive scale. I guess Intrade isn't that influential in the larger picture.

By: chavenet

chavenet — Thu, 25 Oct 2012 08:43:31 -0800

Ah, I remember the good old days when prediction market outrage was about terrorism. That idea got stepped on but it's back. On InTrade. After all, it was always A Good Idea With Bad Press.

By: jquinby

jquinby — Thu, 25 Oct 2012 08:44:56 -0800

Huh. I heard a tiny bit on prediction markets in the car yesterday (on Bloomberg). I caught it mid-stream and had to get out before it had finished, but it was interesting stuff nonetheless. The segment is here.

By: dirtdirt

dirtdirt — Thu, 25 Oct 2012 08:45:51 -0800

Really, it'd seem to me to be a better use of Obama money to try and game Intrade to look like Romney money was trying to game Intrade, if you follow me. Romney doesn't have much to gain because of the nature of the thing, but Obama could gain a bit by "Romney camp so desperate and dishonest that it is trying to buy/cheat Intrade" stories gaining traction and adding to the both true and cultivated sense of the Romney campaign and candidate being less than forthright. Interesting stuff.

By: (Arsenio) Hall and (Warren) Oates

(Arsenio) Hall and (Warren) Oates — Thu, 25 Oct 2012 08:46:17 -0800

How much (or rather, how little) volume must there be for $18,000 to move the needle so much? Yeesh.

By: windykites

windykites — Thu, 25 Oct 2012 08:49:15 -0800

So "prediction market", that's just a bunch of strangers placing bets on what's gonna happen? Why is this a big deal?

By: spaltavian

spaltavian — Thu, 25 Oct 2012 08:50:08 -0800

Nate Silver discusses the Romney blip and the value of betting markets for political predictions.

By: DynamiteToast

DynamiteToast — Thu, 25 Oct 2012 08:50:28 -0800

How much (or rather, how little) volume must there be for $18,000 to move the needle so much? Yeesh. This is probably rhetorical, but you can check the number of shares for any proposition, so between Romney and Obama there are a little over 2.5 million. Not sure how to check for all the 3rd party guys too.

By: grobstein

grobstein — Thu, 25 Oct 2012 08:51:30 -0800

How much (or rather, how little) volume must there be for $18,000 to move the needle so much? Yeesh. There does seem to be very little volume, which is disappointing. Part of the problem is barriers to entry: US anti-gambling policy makes it sort of tricky to put your money into an Intrade account.

By: Eyebeams

Eyebeams — Thu, 25 Oct 2012 08:54:35 -0800

This appears to be consistent with Romney's election strategy.

By: Tell Me No Lies

Tell Me No Lies — Thu, 25 Oct 2012 08:57:15 -0800

So "prediction market", that's just a bunch of strangers placing bets on what's gonna happen? Why is this a big deal? The anonymity aspect encourages people with inside information to participate, in theory making prediction markets more accurate than random polling. I don't believe intrade publishes a list of all of its closed markets with results, which is unfortunate. It would be cool to see an analysis.

By: Kadin2048

Kadin2048 — Thu, 25 Oct 2012 08:57:25 -0800

US anti-gambling policy makes it sort of tricky to put your money into an Intrade account. Not really, you just mail them a check, it gets posted to your account in a week or so. I have an account with them but I'm thinking of closing it because of the monthly maintenance fee. (They don't have any per-trade commissions, but there's a flat fee of $10/mo or something that gets annoying over time.)

By: grobstein

grobstein — Thu, 25 Oct 2012 09:02:54 -0800

I guess that's not too cumbersome, but why can't I just pay by credit card?

By: bepe

bepe — Thu, 25 Oct 2012 09:12:55 -0800

You can also wire them money, which is pretty easy with online banking. That's what I did, from a Chase account. Just couldn't pass up the nearly 50-50 odds after that fake-seeming run on Romney. Also, while I usually don't like gambling because I get too upset when I lose, I figured in this case, I'll be so upset if Romney wins that I'll hardly notice the lost money. And even now I really think they're undervaluing Obama's chances.

By: grobstein

grobstein — Thu, 25 Oct 2012 09:16:02 -0800

Thanks Kadin 2048 and bepe, this changes my whole outlook!

By: vuron

vuron — Thu, 25 Oct 2012 09:24:07 -0800

The fact of the matter is that on many of the smaller intrade predictions it doesn't take much of a coordinated strategy to influence a significant shift in share prices at least for a short period of time. Even for the presidential market the rapid shifts in trade prices that can often happen during the late night in the US can make a significant daily shift in the market price of a share. This seems like it would undermine the assumption that the invisible hand could determine the likely outcome because it's simply too easy for someone to push their hands down on one side of the market or the other.

By: IAmBroom

IAmBroom — Thu, 25 Oct 2012 09:24:19 -0800

Tell Me No Lies: The anonymity aspect encourages people with inside information to participate, in theory making prediction markets more accurate than random polling.

That's only true if the insiders constitute a large portion of the trading. Otherwise, it's once again a survey, with a particular sampling bias, and a notably different weight on the answers (it's not just words).

By: cortex

cortex — Thu, 25 Oct 2012 09:29:18 -0800

Well, it's not precisely just a survey; the argument is not solely that insider information surfaces through the system, but that the direct monetary stake incentivizes all participants to bet on the winner (or on the misvaluation spread) instead of on their personal preference. The big question there is how well that actually works out, of course, but much as there's a certain sane-making value in charging five bucks for a Metafilter account, there's a basic rough filtering effect on Intrade propositions that does not exist on e.g. a random forum poll widget.

By: k5.user

k5.user — Thu, 25 Oct 2012 09:30:18 -0800

So "prediction market", that's just a bunch of strangers placing bets on what's gonna happen? Why is this a big deal? The anonymity aspect encourages people with inside information to participate, in theory making prediction markets more accurate than random polling. I don't believe intrade publishes a list of all of its closed markets with results, which is unfortunate. It would be cool to see an analysis.

So, since it's blocked by my work proxy, I'll rely on what NPR reported earlier this week, about how wrong this market was about the Supreme Court's decision on obamacare.. (Not exactly something a Justice would bet on, but even with potential access to insider info, this thing can be spectacularly wrong)

By: chavenet

chavenet — Thu, 25 Oct 2012 09:38:14 -0800

but even with potential access to insider info, this thing can be spectacularly wrong That outcome indicates that there was insider information approaching zero for that particular item. The predictive abilities of these markets has to do with the level of information of those participating, above and beyond the 50-50 guess factor for yes/no type questions. If there is zero inside information, then the predictive value falls to the level of "educated guessing", which, yes, can be spectacularly wrong.

By: (Arsenio) Hall and (Warren) Oates

(Arsenio) Hall and (Warren) Oates — Thu, 25 Oct 2012 09:39:37 -0800

If you want Obama to win, then put your money on Romney. Obama wins, YAY, consider the bet your donation. Romney wins, BOO, but as a consolation prize, your investment has been doubled (or more).

By: dgaicun

dgaicun — Thu, 25 Oct 2012 09:39:39 -0800

Intrade currently has Obama at 59.7% chance of winning. Nate Silver in the FiveThirtyEight link above notes:

However, Intrade is also something of an outlier relative to other prediction markets. Another prediction market, Betfair, gave Mr. Obama about a 63 percent chance of winning as of early Wednesday morning. The consensus of sports books, similarly, implied that he has about a 65 percent chance. These values are not significantly different from the FiveThirtyEight forecast.

Silver has Obama's odds at 71%. Another model puts his current chances above 86%.

By: Kadin2048

Kadin2048 — Thu, 25 Oct 2012 09:57:56 -0800

It's worth noting that if you have accounts at multiple prediction markets -- and especially if you're in Europe and have access to Betfair, the other big one -- you can make money arbitraging the markets against each other. It's probably a lot easier to make money that way than by placing straight bets on the actual outcome of various events (unless you have inside information). E.g., if Intrade has Obama at 59.1 and Betfair is at 63%, you can buy Intrade shares and sell Betfair shares and when both markets close, you should end up making the spread. Of course, you have to be really careful that the markets are going to agree with each other at close. I think that there are some prediction markets that only care about the popular vote, not who actually becomes President. It's entirely possible that they might close to opposite values (actually a fairly high risk, as investments go). I've never done it because it's too time-consuming, but it strikes me as the sort of thing some college kids with a lot of time and analytics abilities could just clean up at, done carefully. The spread between Intrade and sports books are often fairly high, in particular.

By: nobody

nobody — Thu, 25 Oct 2012 10:17:01 -0800

about how wrong this market was about the Supreme Court's decision on obamacare.. (Not exactly something a Justice would bet on, but even with potential access to insider info, this thing can be spectacularly wrong) For what it's worth, that appears to be a situation in which a bunch of insider info turned out to be wrong, probably because of a last minute change in Robert's vote that may not have had much if any chance to filter out through the grapevine. (Though I'm using 'insider info' here to mean something more along the lines of people keeping track of which justices' clerks are keeping longer hours, not something like the clerks or justices themselves placing guaranteed-to-win bets at the last minute to make a quick buck.)

By: MuffinMan

MuffinMan — Thu, 25 Oct 2012 10:20:38 -0800

lso, while I usually don't like gambling because I get too upset when I lose Or you could do what I did absentmindedly in 2004 - put a cheeky bet on Kerry to win at the bookies, wake up after the election annoyed that you had lost your money and then find out that you had mistakenly backed George Bush and had won a small sum. Rest easy folks. I can't vote in this election.

By: Blue_Villain

Blue_Villain — Thu, 25 Oct 2012 10:23:04 -0800

I'm pretty sure the "cheeky" bet made that apparent. Well, to those of us who weren't really concerned about it anyway.

By: Mayor West

Mayor West — Thu, 25 Oct 2012 11:01:49 -0800

Which is what'd you expect. However, I think if he would've covered his tracks better and made it less clear it was just one guy, it could have had a tangible effect. Definitely worth $17,000 in the age of SuperPACs and unlimited ads. I still don't get it. Contemporary political ads are targeted at undecided voters, or (if we can drop the act for a moment) people who are willfully ignorant of politics and their elected leaders. People trading shares on InTrade aren't going to change their mind on the basis of obvious manipulation. Their only hope is to trigger a big enough shift that some bored political beat reporter in Sandusky decides to write something sensationalist about it, and that the AP picks it up. While that's not *totally* impossible, it seems like a pretty long shot, even for a PAC with unlimited Koch money.

By: fuse theorem

fuse theorem — Thu, 25 Oct 2012 11:23:52 -0800

Silver has Obama's odds at 71%. Another model puts his current chances above 86%. Inkling (bets are not made with actual money) has Obama at 84%.

By: srboisvert

srboisvert — Thu, 25 Oct 2012 11:38:07 -0800

I'm gradually coming around to the belief that the Republican Party is actually Improv Everywhere and this is their best most comprehensive performance yet.

By: srboisvert

srboisvert — Thu, 25 Oct 2012 11:40:50 -0800

I still don't get it. Contemporary political ads are targeted at undecided voters, or (if we can drop the act for a moment) people who are willfully ignorant of politics and their elected leaders. People trading shares on InTrade aren't going to change their mind on the basis of obvious manipulation. Their only hope is to trigger a big enough shift that some bored political beat reporter in Sandusky decides to write something sensationalist about it, and that the AP picks it up. This isn't about manipulating voters. Like all attempts at poll manipulation it is about manipulating the highly gullible and lazy press which then manipulates voters for you. Basically by making their candidate look like a possible winner they avoid the "I don't want to vote for a sure loser" effect hurting them as much at the polls.

By: jhc

jhc — Thu, 25 Oct 2012 11:56:40 -0800

It's interesting that InTrade becomes less predictive when outsiders (e.g. reporters) start adjusting their behavior based on its predictions. It's kind of like those science fiction stories where the computer can predict the future, but can't predict the future taking into account the effects of its own predictions.

By: Wordshore

Wordshore — Thu, 25 Oct 2012 12:18:56 -0800

Well, there's other bets you can place if looking for variety. Paddy Power, an Irish bookie, offers you the chance to bet on Obama's response to Trump, when and where a live yeti will be found, how Julian Assange will leave the Ecuadorian embassy, the name of Lady Gaga's first baby, the next Facebook CEO, and which drugs will be legal by the end of 2020 in the UK.

By: atrazine

atrazine — Thu, 25 Oct 2012 14:59:08 -0800

And then the prices re-equilibrated rapidly as speculators rushed to take the free money from the would-be "manipulator," who was quickly out thousands of dollars for no benefit. A Romney partisan who doesn't understand economics? Surely not.

By: arcticseal

arcticseal — Thu, 25 Oct 2012 16:19:04 -0800

I'm gradually coming around to the belief that the Republican Party is actually Improv Everywhere and this is their best most comprehensive performance yet. The flash mob of Tea Party ID checkers on election day will be simply hilarious.

By: Tell Me No Lies

Tell Me No Lies — Thu, 25 Oct 2012 17:45:34 -0800

but even with potential access to insider info, this thing can be spectacularly wrong Definitely. In addition given how small these markets are there isnt much impetus for real players to get involved.

By: drezdn

drezdn — Thu, 25 Oct 2012 18:12:36 -0800

Obama is up to 64% last I checked. One possibility for manipulation mentioned in another thread is to create positive media stories for your candidate.

By: saber_taylor

saber_taylor — Thu, 25 Oct 2012 19:07:22 -0800

It's not just inside information that would make a prediction market better. When you poll voters on who they think will win versus who they say they'll vote for then the results are more predictive. I think the polls this season are attracting to 50% since it provides more room for discussion before pulling the lever.

By: honeybee413

honeybee413 — Thu, 25 Oct 2012 22:10:14 -0800

I'm planning to vote for Obama, but I considered betting a little bit against him. That way, if Romney wins, I at least have *something* to be happy about.

By: Kadin2048

Kadin2048 — Sat, 27 Oct 2012 15:34:44 -0800

Re the whole poll-manipulation thing, I think there's another possibility: rather than super-PACs manipulating polls in order to cause the media to run pro-candidate stories, it could instead be the media picking polls that make the race seem more even, because that's a more interesting story than the incumbent (or the clear favorite) winning, still winning, yep still winning, yep he won. The media has a vested interest in a horse race. A poll that either shows something close to 50/50 ("stay tuned!"), or that shows change from whatever's currently going on ("what an upset!"), are the best things to report. Polls that show no change are essentially a non-story. The US media is, fundamentally, entertainment. They are constantly trying to put together a compelling story in order to attract and retain eyeballs. Even if they were completely politically agnostic, they might still favor one candidate over another at various points in the race (much in the way that they arbitrarily pick various athletes to "profile" during the Olympics) because they think it makes more compelling viewing that way. Thus the media conspiracy isn't really a conspiracy, it's just the banality of giant entertainment companies run by avaricious producers constantly churning out whatever they think the American public wants fed to them. The super-PAC cash is really the icing on the cake.