🔮Polymarket Voodoo
How to not go insane watching election odds
In today’s Polymarket’s general election odds, Donald Trump is the 61.5% favorite to win the presidency and Trump is now ahead in all seven of the swing states.
The most notable move came over the weekend in Nevada where Republicans received some bullish early vote numbers.
Introducing … Polypoll
Donald Trump mentioned Polymarket in a rally last week, where he seemed to repeat a common error about prediction markets.
A similar mistake was printed in the New York Times. Can you spot it?
If both Trump and the NYT make the same mistake - confusing Polymarket odds with polls - in the same week, it seems like the education around prediction markets still has a way to go.
For this reason we’re departing from our normal Monday email to bring you a guide to some misconceptions about prediction markets that have surfaced as the election approaches.
What does the Polymarket number represent?
Odds on Polymarket are derived from the market price of a contract that pays $1 if an event occurs and expires worthless if it doesn’t. For example, someone paying 60 cents for a Trump “yes” contract, implies a 60% chance that he wins.
The logic is that a rational person will only pay the going rate, currently 60 cents for a Trump yes share, if he believes that the odds of Trump winning are higher than 60% (otherwise the trade would be negative expected value).
Odds change on Polymarket as shares are traded on an orderbook, in which the bid and ask prices are listed. Polymarket’s docs have more on how the orderbook works.
What’s the difference between polls and Polymarket?
The key difference between polls and Polymarket is their output: Polls estimate vote share, while prediction markets and election models estimate the odds of a certain outcome.
A big point of confusion is that both vote share and odds are expressed as a percent, but do not mean the same thing. In the context of a US presidential election, polls usually report the percent of the electorate (eg a state or national popular vote) a candidate will get, and odds reflect the probability that a candidate will win the election.
Imagine these hypothetical scenarios:
An FDR-redux candidate is the heavy favorite in polls, constantly polling in the popular vote at ~60% (a number unheard of in modern US politics). What would be his or her odds on a prediction market of winning the election? Probably well above 90%.
The Democratic and Republican presidential candidates are tied 49% to 49% in the national polls. Who is favored to win in the odds? The Republican in this scenario would be a favorite in the odds to win because of the slight edge that Republicans have in the electoral college.
Is Polymarket “skewed conservative”?
Simon Rosenberg has coined the term “Polymarket voodoo” for what he says is a systematic bias on the platform towards Republicans. Is this accurate? Here is a table shared on Twitter by Nic Carter comparing several forecast methods:
As Carter points out, the biggest gap is between market-based odds (Polymarket and bookmaker markets) and polling-driven models like Silver Bulletin. This makes sense because prediction markets react to news with a few clicks of a mouse, whereas election polling can take days to weeks to update.
And even among the market-based prediction methods, Polymarket is not a Trump outlier. European betting sites like PaddyPower (63.6% Trump odds) are quoting even higher for Trump.
“But you were wrong about Beyonce.”
Prediction markets are not perfect. Polymarket users gave a 90%+ chance for Beyonce to perform at the DNC based on what turned out to be false reporting. Intrade favored Blake Masters to win the Arizona senate race in 2022, which went the other way. There are plenty of examples you can find.
But prediction markets require you to balance two contradictory ideas: that the odds can be on average correct, but also can be temporarily wrong, just in the same way that a stock can become temporarily overvalued or undervalued due to market sentiment, forced buyers or sellers, or any number of idiosyncratic factors. These gaps create incentives for traders to profit by moving the price closer to the true probability: the bigger the mispricing, the bigger the incentive to correct.
Another way to think about it is in Scott Alexander’s must-read Prediction Market FAQ: “Either prediction markets will be accurate, or you can get rich quick,” ie, there is an opportunity to make money correcting the odds.
Whales are “distorting” the markets.
Several articles have appeared in the last week discussing the impact of large traders on Polymarket’s election odds.
As a policy, The Oracle does not comment on individual accounts or trades, but there are two points to make here. First, because Polymarket is built on top of public software networks, anyone on the internet can see what is happening on Polymarket at any time, including what accounts hold what shares. This information is free and public – a high degree of transparency generally and significantly higher than other trading platforms.
If any user has any reason to believe that trades are being made for anything other than financial reasons, it is easy to adjust the odds to account for this. In fact, there are already several researchers using Polymarket’s open data to calculate how excluding the largest traders betting on Trump would affect the odds.
Second, it is common in financial markets for there to be a power law, or 80-20 principle, i.e., a few whales and many minnows. Above, for example, are the top holders, with usernames redacted, in the market on whether Trump will appear on a podcast with Haliey Welch (aka Hawk Tuah Girl) before the election (🔮 2% odds).
Notice the large holders on both “yes” and “no” sides. Clicking “Top Holders” on most other markets reveals a similar whales and minnows dynamic.
Grab a Yuengling and Strap In
If you’re the kind of person who already has 50 Polymarket tabs open, alongside Nate Silver, RCP and whatever other esoteric political indicators you favor (Washington primary? Yard sign counts), then you probably already know everything covered today and are fully psychologically prepared for either candidate to win.
So the ask is this: can you grab a bottle of Pennsylvania’s favorite beer, and spare a moment tonight to write a postcard to your favorite journalists and remind them (politely of course) that Polymarket is not a poll?
Disclaimer
Nothing in The Oracle is financial, investment, legal or any other type of professional advice. Anything provided in any newsletter is for informational purposes only and is not meant to be an endorsement of any type of activity or any particular market or product. Terms of Service on polymarket.com prohibit US persons and persons from certain other jurisdictions from using Polymarket to trade, although data and information is viewable globally.
Who’s shorting this? is the more interesting question.
"In fact, there are already several researchers using Polymarket’s open data to calculate how excluding the largest traders betting on Trump would affect the odds."
I don't see this in the article you hyperlinked.