See anything funny about this chart of candidate prices in DNOM at PredictIt?

Yeah, uh, why is Andrew Yang at 13c?  And if I told you he’d later hit 17c, how the hell did that happen?  What are his actual chances anyway?

Prediction Markets: How do they work?

The Yang pricing is so ridiculous that pretty much everyone is dunking on the market for it.  Nate Silver asks ironically “Not just Yang but now also Buttigieg ahead of Warren, Booker and Klobuchar on PredictIt, which definitely isn’t biased toward candidates that young white dudes who spend a lot of time on the Internet like, why would you ask a thing like that when markets can’t be wrong?”

Are the markets wrong?  If the price is so obviously bad, why isn’t it being corrected by smarter money?  What’s going on here?  The take-home is that basically Nate Silver is right.  But let’s explore a bit why that is.

Prediction markets are supposed to distill the wisdom of the crowd in numerical form.  Hundreds and thousands of traders each wager whether or not an event will happen.  Some will think it will, some will think it won’t.  One person says “I think this has a 25% chance of happening” but because they’re smart, they throw out an offer for 15c shares.  Another person says “No way this happens, but let’s see if I can get some 65c NO shares” and they put up an offer there.  Some more traders come in and, leapfrog those offers by just a bit, hoping someone will bite.  The process repeats until the market reaches equilibrium.  When events transpire that influence the odds of the contract resolving one way or the other, people will trade their shares and new people will enter the market.  Thus the movements in price represent how much the market thinks a given event influenced the odds.

Okay so that’s the theory.  The reality?  People bet on who they want to win, mostly, and then form an emotional attachment to those shares and hold them til the bitter end or glory.  And the online betting markets?  Yep, Nate is right.  Mostly a bunch of extremely online dudes, which means the “crowd” doesn’t represent the full range of relevant human input that it could.  And Andrew Yang…

The Power of Memes

Yang is the meme-candidate (see this subreddit).  And we’re just coming off a meme-candidate winning in 2016.  So why not again?  Same huge field of opponents, right?  He has a dedicated semi-ironic fanbase and that means new money and people willing to throw away a little bit in order to secure the bag.  Additionally, he was added to the market just as he started getting mainstream media traction.  A huge influx of new money all coming on to one candidate?  You’re going to get a bubble.

Is the Betting Limit Distorting his Price?

People like to argue that if there weren’t a max bet of $850, more smart money would be able to come in and quash these bubbles before they get started.  And there is a case to be made for that.  But let’s also not underestimate just how much demand there was for Yang.  And how much money is betting against him too!  The data:

Yang has attracted HUGE volume.  Overall, as of this writing, 820k shares of Yang have been traded at a price average of around 11c.  In only one week, he went from 0 active shares (that is, shares that are still held by traders) to nearly 480k, the most in the market.    By my estimate, around $50,000 is currently bet on him, and over $400,000 against him.  How much better could the market have performed without a limit?  Sure, there’d be more smart money to bet against (substantially more than $400k?), but don’t forget that there’d also be more stupid money to bet on him at those prices too.  Though I wasn’t a trader back then, people that were can attest that there were plenty of idiotic whales on the Intrade markets.

Eventually, the Bubble Bursts

Yang’s price peaked two days ago at 17c and the air has been slowly deflating from the balloon since (he now trades at 11c).  I don’t think he’ll go much under probably 7 or 8c as hopeful/fearful bettors will wait to see how he does in the first few debates, but I don’t think he’s revisiting 15-17c anytime soon without some sort of polling bump or something.

Then again, I didn’t think he’d make it over 10c and I’m holding 94c NO shares so what do I know?

So What Are Betting Markets Good For?

If transient influxes of biased new money can distort their pricing, if the population of bettors is skewed such that it favors certain candidates over others, what can we even learn from betting markets anyway?  This is the thrust of Silver et al.’s criticism of the markets.

And basically, the critics are right, but only narrowly.  Market pricing doesn’t tell you the actual odds of something happening.  It tells you what the crowd of extremely online dudes thinks the odds are, and that’s about it.  But!  The movements in price, the way bubbles come and go, when volume goes up or down – that stuff is informative.  It tells you what sorts of things the crowd thinks matter.  It tells you how much the crowd thinks those things matter.  And that information, if you study it, can help you make money off the crowd when those things happen again…