• Crypto prediction markets caught mainstream attention as (confirmatory) indicators of US presidential election outcomes deviated from official polls.
• Be careful of “resulting” when evaluating decisions and the efficacy of betting markets as signals.
The US presidential election between former President Donald Trump and Vice President Kamala Harris looked to be a “dead heat” – until it wasn’t. The polls got it wrong, again. What made it worse was the flashy, upcoming pollster that got it right: Polymarket.
Polymarket is a decentralized, crypto-based platform where participants can place bets on practically anything, ranging from Federal Reserve rate cuts this year, to if the US will confirm that aliens exist in 2024 (2% at the time of writing btw).1 Each event has a verifiable outcome with a binary payoff: $1 for Yes, $0 for No. Users can buy Yes/No tokens from the market-making system (a liquidity pool) at a given price (e.g., 60¢) until the event concludes at a specified time. Uniquely, participants can also sell these tokens at any time, profiting from price movements regardless of the event’s ultimate outcome.
Touting itself as a real-time gauge of public sentiment (the wisdom of crowds) via price discovery, how did Polymarket predict the results of the election better than polls? Well, it didn’t.
A crucial mistake in evaluating decisions is “resulting” – where we misconstrue judging the quality of the decision based on the end result, not the process by which it was made.
Paul DePodesta, a baseball executive and one of the protagonists in Michael Lewis’s Moneyball, tells about playing blackjack in Las Vegas when a guy to his right, sitting on a seventeen, asks for a hit. Everyone at the table stops, and even the dealer asks if he is sure. The player nods yes, and the dealer, of course, produces a four. What did the dealer say? “Nice hit.” Yeah, great hit. That’s just the way you want people to bet—if you work for a casino.2
There are quite a few issues with the Polymarket process with respect to its predictive efficacy:
- Understanding odds: These values are odds, prices influenced by money flows and speculation – not fundamental reflections of reality. Odds do not translate to chances of winning.
- Sampling: Critiques of polls often point to how the compositions of respondents do not truly reflect actual voters – and this is undoubtedly the case for Polymarket. For one, it is federally banned in the US. Additionally, the demographic of crypto users naturally leans pro-Trump based on his pro-crypto stance (along with Elon Musk being a well-known crypto supporter). A small, biased sample to say the least.
- Manipulation: Advocates will point to $3B+ in volume (not open interest) as validation of the product/process. However, studies of the transactions showed that at a certain point half of all purchase volume for the election market were wash sales trying to take advantage of airdrop rewards. This exacerbated the fact that there were just a handful of “whales” (large wallets) taking up a disproportionate amount of value and skewing prices/odds.3 Not much of a “crowd”.
Even if all polls were conducted on election day itself (no temporal error) and took an infinite sample size (no sampling error), the average poll would still miss the final margin in the race by about two points.4
While not perfect, rigorous polling remains a valuable tool that reflects fundamental voting patterns. Though prediction markets like Polymarket may grow more reliable with increased adoption, they currently warrant caution. Ironically, the French whale that profited from his Polymarket bet on Trump actually relied on polling data himself, using certain polls to rebut others.5 Notably, though, parts of his decision-making process, like the “shy Trump effect”, have previously been refuted.6 Dumb luck or deserved success?
Important Disclosures & Definitions
1 Polymarket. (n.d). Will the US Confirm That Aliens Exist in 2024? Retrieved 11/05/2024 from Polymarket online.
2 Mauboussin, Michael J. (2013). More Than You Know: Finding Financial Wisdom in Unconventional Places. New York, NY: Columbia University Press.
3 Tartar, A. and Fevrier, T. (10/25/2024). How 1% of Polymarket Bettors Are Boosting Trump’s Odds. Retrieved 11/05/2024 from Bloomberg online.
4 McPhail, S. (11/05/2024). How Accurate are the US Election Polls? Retrieved 11/05/2024 from The Spectator online.
5 Osipovich, A. (11/06/2024). How the Trump Whale Correctly Called the Election. Retrieved 11/06/2024 from Wall Street Journal online.
6 Keeter, S. and Kennedy, C. (08/28/2024). Key Things to Know About US Election Polling in 2024. Retrieved 11/06/2024 from Pew Research online.
AAI000786 11/12/2025
