Despite a robust track record dating back to the 1988 U.S. Presidential election, prediction markets still faced hordes of skeptics leading up to the 2024 election.
What professional pundits, and Nobel Prize winners like Paul Krugman, fail to understand about prediction markets is that if you think Trump is overpriced, then you should buy Harris shares. It’s really as simple as buying an undervalued asset.
Professor Robin Hanson describes prediction markets as being difficult to manipulate because they consist of “wolves” and “sheep”:
The wolves know something and go to the market to trade on what they know. The sheep don’t know anything and they’re there to trade for some other reason. The wolves want to trade against the sheep. When a wolf trades against a wolf, on average, neither of them can make money. But when a wolf trades against a sheep they can make money, so the wolves want to trade against sheep. Wolves are attracted to where the sheep go. So this means the more sheep there are, the more wolves are attracted, and therefore the price gets more accurate with more wolves. The key thing to know is that a manipulator, someone who wants to push the price somewhere, is basically a sheep. Manipulators don’t know something, they just have a goal they want to achieve by pushing the price. Expecting there to be manipulators attracts the wolves and makes the prices more accurate. [Edited for clarity]
In the campaign’s final month, and several weeks after, the media fixated on “Théo” the French Whale buying tens of millions in Trump shares on Polymarket as if this invalidated the market prices. If Théo was brazenly wrong, traders should have gobbled up Harris shares. (They didn’t.) Below we can see that prediction markets were generally in agreement on Trump’s chances, while both Seer prediction markets, the Smarkets betting exchange and the bookmaker Bovada were all within 1 cent of the Polymarket Trump price on Election Day:
How did Seer markets perform on Election night?
Below are four takeaways about the accuracy of Seer traders in the 2024 election:
1) In the final weeks before Election Day, Seer traders correctly saw Trump winning the Presidential election in markets on both the Ethereum and Gnosis chains.
2) The Congressional Control market correctly forecasted a Republican U.S. Senate (80%) but missed the narrow Republican majority in the U.S. House (as of this writing, Republicans have a one-seat majority).
3) Using conditional markets to trade on the price of Bitcoin in the event of either a Trump or Harris victory, Seer traders foresaw a BTC price spike in the event of a Trump victory relative to Harris.
4) Traders didn’t see Trump winning the popular vote, only giving him a 28.4% chance on Seer. This was roughly in line with Trump’s chances of winning the popular vote on other prediction markets like Polymarket, where he traded at 27.1%, and Nate Silver’s model at 23.8%. Seer’s Multi Scalar market for the Presidential election popular vote totals showed Harris receiving the most votes, but the forecasted percentages for both Trump and Harris were close to the final results (<2%).
Overall, traders on Seer illustrated the wisdom of crowds by traders ‘betting their beliefs’ in the presidential election despite sharp criticism from the media and pundits. While not perfect, these markets showed the value of aggregating human knowledge. Ethereum founder Vitalik Buterin has described this as “Info Finance” where market traders try to profit from predicting future events and information seekers rely on the probabilities created by traders.
Seer was intentionally built to bring these diverse actors together in one decentralized prediction marketplace:
- Traders: Buy and sell positions in markets.
- Market creators: Create markets and provide liquidity to facilitate trading.
- Information seekers: Pay to gather information about the possibility of an event and/or its impact.
The two Bitcoin price markets conditional on the Presidential election winner also illustrate the untapped potential for futarchy: given a single decision, like the Presidential election, what do the markets think about potential outcomes for each option? Seer conditional markets showed competing Bitcoin prices in the event of either a Trump or Harris victory, with traders indicating that Trump winning would be more favorable to Bitcoin than Harris winning, which we’re now watching play out in reality.
About Seer
Seer is a non-custodial and decentralized prediction market protocol existing on the Ethereum and Gnosis blockchains. Traders use DAI tokens to buy and sell outcome tokens on whether an event will happen. The prices for these outcome tokens can reflect the probabilities of that event happening. Seer allows for permissionless market creation and users to provide liquidity to markets.
