How Does Predict Work?
Last updated
Last updated
Predict is a prediction market: a platform where you can buy and sell shares whose value is based on the outcome of a future event.
In may ways, it's similar to the classic trading venues you know. In others, not so much.
Let's dive right in with an example of a market:
Will Candidate McCandidateFace win the election on December 21, 2025?There are only two answers to the question above: yes and no. Fittingly, the market has two tradable assets (we call them shares): YES and NO. One other thing you need to know, each market is time-bound, meaning it has an expiry date.
The above market expires on December 21, 2025, which is when we'll know for certain whether Mr. McCandidateFace has won the election. Once we reach this stage, the correct share can be redeemed for $1 — while the other becomes valueless.
Up until that point, though? These shares will trade freely, but the market will price them commensurate with their probability of occurring.
For instance, if YES trades at $0.95 while NO trades at $0.05, it would suggest a high likelihood (95%) that Candidate McCandidateFace will win, based on currently available information. Naturally, as more information becomes available, the market will trade to reflect an updated price.
Assuming Mr. McCandidateFace wins and YES resolves to $1.00, anyone that purchased a share at $0.95 will make a $0.05 profit on it. However, should he drop out of the race or lose, YES resolves to $0.00 while NO resolves to $1.00 — meaning that anyone having purchased NO at $0.05 would have made a per-share profit of $0.95.
Why does price represent probability?
In short, because people like making money. With an incentive to be correct, rational participants will examine all available information before taking a position — bidding the price to the probability they assign to that event.
In a case where the wider market believes probability to be higher than the share’s current trading price, it’ll bid it up until the share is no longer ‘undervalued.’
By aggregating these ‘educated guesses’ across hundreds, thousands, or millions of people, we arrive at an approximate probability of an event taking place.
Note, however, the inherent limitations of these systems: participants have limited liquidity and may be risk-averse, meaning they won’t necessarily bet with full conviction on a single position.
Above, we demonstrated how an event with only two possible outcomes works. But Predict also supports multi-outcome markets, where you can place bets on more than two options. For instance:
Which candidate will win the 2025 December election?Candidate McCandidateFace
Runner McRunnerFace
Politician McPoliticianFace
The same rules apply as in binary markets, but this YES/NO mechanism is used on a per-candidate basis (such that each 'answer' will resolve to either YES or NO). Again, only one of these outcomes is possible — so the winning candidate's YES share will resolve to $1 upon expiry, while the rest becomes valueless.
Remaining true to the platform’s decentralized ethos, results are obtained via an on-chain oracle: an entity that allows blockchain-based dApps to parse data from the real world.
To get the correct outcome for a given event, we use UMA’s optimistic oracle, a decentralized truth machine whose incentives make it easy (and rewarding) to act honestly, but difficult (and expensive) to cheat.
It’s known as ‘optimistic’ because statements are assumed to be true unless they’re disputed by UMA’s participants. In rare instances, ambiguous market parameters may cause some disagreement, in which case an emergency resolution can be enforced.
Will Elon Musk twerk on the surface of Mars in 2024? Will DOGE flip BTC in market cap by the end of the month? Will the Cookie Monster be the first Sesame Street character to become US President?
Ask 99% of people, and the answer to these questions will be a resounding NO. Many would even consider these events not happening to be a certainty, and, therefore, any NO share trading at less than $1 to be free money.
As mentioned, the vast, vast majority of participants accept these events to be virtually impossible — but some outliers will buy very cheap YES shares in the same way they might purchase a lottery ticket.
We call these markets bond markets, as they provide an opportunity for ‘rational’ actors to earn a yield by parking their USDB in them until market expiry. For example, buying NO shares at $0.98 in a bond market would result in a $0.02 profit (effectively, a 2% yield) upon market resolution.
However, it’s worth noting that there’s no such thing as a sure thing. No matter how certain an outcome looks, there’s always a chance that it doesn’t pan out.
