A contract on Kalshi is trading at 67 cents. It pays out one dollar if a particular candidate wins their primary, nothing if they lose. A reader scrolling past sees the number and moves on. A trader sees something more specific: the market is saying there is a 67% chance the candidate wins.
That translation, from a price in cents to a probability in percent, is the single most useful trick in reading prediction markets. It is also the thing most casual readers get wrong, or do not realise they are doing at all. Once you can do it in your head, every market price becomes a forecast you can argue with. Without it, the numbers are just numbers.
The 67-cent trade, unpacked
Start with the mechanic. A binary prediction market contract pays exactly $1 if the event happens and $0 if it does not. If you can buy that contract for 67 cents, you are risking 67 cents to win 33 cents of profit. For that trade to break even in the long run, the event needs to happen roughly 67% of the time.
Flip it around. If you think the event is genuinely 80% likely, then 67 cents looks cheap and you buy. If you think it is closer to 50%, you sell, or you stay away. The price is the market's collective best guess at the probability, expressed in cents on the dollar.
That is implied probability in one paragraph. Take the price, divide by the maximum payout, and you have the percentage chance the market is assigning. 67 cents on a $1 contract means 67%. 12 cents means 12%. 94 cents means 94%. The arithmetic is deliberately friendly.
Why this is different from betting odds
Anyone who has used a sportsbook will recognise the underlying idea but not the format. UK bookmakers quote fractional odds (5/2, 4/6, 11/10). American sportsbooks quote moneyline (+250, -150). European books quote decimal (3.50, 1.67). Each format encodes a probability, but you have to do the conversion.
Prediction markets skip all of that. The price is the probability. A 30-cent contract is a 30% chance. There is no margin to back out, no overround to subtract, no fractional-to-decimal mental gymnastics. The format is closer to a thermometer than a betting slip.
The deeper difference is that the price is set by traders, not by a bookmaker setting a line. Our guide to how prediction market odds work walks through the order-book mechanics, but the headline is this: the 67 cents reflects the equilibrium between buyers and sellers right now, not a bookmaker's view shaded by their book.
Where the probability actually comes from
Markets do not magically know the future. They aggregate the views of everyone willing to put money behind a position. The 67% on the primary contract is the weighted average of every trader's belief, weighted by how much capital each was willing to commit at that price.
That aggregation is the entire reason prediction markets exist as a forecasting tool. A polling firm samples a thousand people and asks them what they think. A market lets a thousand people place real bets and reads the price. The two methods often agree. When they disagree, the argument over which is closer to reality gets interesting, which is the subject of our piece on how prediction markets stack up against traditional polling.
Markets are not infallible. They can be thin (few traders, easy to push around), biased (one side is harder to short than the other), or just wrong (the crowd is sometimes the crowd). But the price is a real signal, and implied probability is how you read it.
The 1% gap that is not a 1% gap
Worth flagging a quirk. The arithmetic distance between 50% and 51% is one percentage point. The arithmetic distance between 98% and 99% is also one percentage point. The informational distance is wildly different.
Moving from 50% to 51% means the market has shifted from a coin flip to a coin flip with a feather on one side. Moving from 98% to 99% means the market has roughly halved the implied chance of the long-shot outcome (from 2% to 1%). The same one-point move tells you something very different about how confident the market has become.
Traders watch the tails for this reason. A favourite drifting from 92 cents to 88 cents looks like a small move on the surface. In probability terms, the long-shot outcome just went from 8% to 12%, a 50% relative increase. The price moved a little. The forecast moved a lot.
What can break the translation
The price-equals-probability rule has edge cases. Fees are the obvious one. If a platform charges a 2% fee on winnings, a contract trading at 67 cents implies a probability slightly above 67%, because part of your $1 payout disappears to the house. On most regulated US venues like Kalshi the fee is small enough to ignore for casual reading; our explainer on Kalshi event contracts covers the exact mechanics.
Time value is another. Money tied up in a contract that resolves in two years has an opportunity cost. A 67-cent contract resolving tomorrow is a cleaner probability read than a 67-cent contract resolving in 2028. The longer the wait, the more the price is shaded by what else the trader could have done with the capital.
And then there is liquidity. A market with a wide spread (say, 65 bid, 70 offered) does not have a single implied probability. It has a range, and the midpoint is a polite fiction. Reading 67% off a thinly traded market is reading more precision than the market actually provides.
When the probability moves
The interesting part of implied probability is not the static reading. It is the movement. A contract that traded at 40 cents on Monday and 67 cents on Friday is telling you the market changed its mind. The question is why.
Sometimes it is news (a candidate dropped out, a court ruled, a poll landed). Sometimes it is flow (a large trader took a position and the order book absorbed it). Sometimes it is nothing identifiable, which is itself worth knowing, because mysterious moves often precede public news.
Reading the move is its own skill. The shape of the curve matters: a gradual drift suggests reweighting, a sharp step suggests a discrete event, a spike-and-fade suggests a misread. The numerical change from 40% to 67% is interesting. The path between the two is more interesting.
Why this matters if you never trade
Most people who read prediction market prices will never place a trade. That is fine. The price is still useful as a forecast, in the same way you can read a weather forecast without owning a weather station.
The 67% on the primary contract is roughly what a calibrated forecaster would charge you to take the other side. It is not a guarantee, not a guess, not a marketing number. It is a price someone is paying with real money, which makes it a more honest forecast than a lot of what gets published in the press. Reading it well, with the caveats about fees, liquidity, and tail sensitivity, is one of the most useful analytical skills you can pick up from spending time around these venues.
iPredicta tracks implied probabilities across UK and US prediction market venues and writes about how those numbers move when news lands. The goal of the /learn section is to give readers enough mechanical fluency that the market prices they encounter elsewhere stop being noise and start being signal.
Frequently asked questions
How do I calculate implied probability from a prediction market price?
Divide the contract price by the maximum payout, usually $1. A contract trading at 42 cents on a $1 payout implies a 42% probability that the event occurs. The arithmetic is deliberately simple because prediction market contracts are designed to settle at either $1 or $0. For markets that use a different scale (some venues quote in points or use multi-outcome shares), the same logic applies: price divided by maximum payout equals implied probability. The number you get is the market's current best guess, weighted by how much capital traders have committed at that price. It is a snapshot, not a prophecy, and it moves as new information enters the market.
Is implied probability the same as the true probability of an event?
No, implied probability is the market's collective estimate, not the underlying truth. The two can diverge for several reasons: thin trading, structural biases, missing information, or simply the crowd being wrong. Research on calibration suggests that liquid prediction markets are reasonably well-calibrated over large samples, meaning events priced at 70% tend to happen roughly 70% of the time. But any individual contract can be mispriced, sometimes badly. Treating implied probability as a forecast worth taking seriously is sensible. Treating it as the literal truth of the universe is not. The price tells you what informed money currently believes, which is a useful input, not a final answer.
Why do bookmaker odds and prediction market implied probabilities differ?
Bookmakers build a margin into their odds, while prediction markets generally do not. A sportsbook quoting a coin-flip event might price both sides at decimal 1.91, which implies roughly 52% on each side and adds up to 104%. The extra four points is the bookmaker's overround, their built-in edge. Prediction markets, by contrast, are peer-to-peer: traders match against each other, and the prices on both sides of a binary contract should sum to roughly 100 cents plus a small fee. The upshot is that prediction market implied probabilities are usually a cleaner read on consensus belief than bookmaker odds, which always have the house's cut baked in.
What does it mean when a market moves from 50% to 55%?
It means the market has shifted from a true coin flip to a modest favourite, with the implied probability of the event rising by five percentage points. In absolute terms, that is a meaningful but not dramatic move. In relative terms it depends where on the curve you are. A move from 50% to 55% is a 10% relative increase in the favoured outcome. The same five-point move from 90% to 95% halves the long-shot probability, which is a much bigger informational shift. Traders watch the tails for this reason. Small absolute moves near 0% or 100% carry more forecasting weight than equivalent moves near the middle of the range.
Can implied probabilities add up to more or less than 100%?
Yes, and the gap tells you something useful. In a clean binary market with no fees, the yes and no contracts should sum to exactly $1, or 100%. In practice they often sum to slightly more (because of fees and spreads) or slightly less (because of arbitrage opportunities that have not yet been closed). For multi-outcome markets like election winners, the sum of all implied probabilities across every candidate should also be close to 100%, but rarely lands exactly there. A persistent gap above 100% suggests structural friction; a persistent gap below 100% suggests an arbitrage that someone should be eating. Our guide to Polymarket arbitrage covers the mechanics.