How to Read a Prediction Market Price
A 60¢ contract means the market puts the odds at roughly 60% — but spreads, depth, fees, and longshot bias all decide how much you should trust that number. Here's how to read a price like a forecaster, with worked examples.
A contract trading at 60 cents on Kalshi means the market’s collective estimate is that the event has about a 60% chance of happening. That’s the whole trick: each contract pays $1 if the event occurs and $0 if it doesn’t, so the price you’d pay for a dollar of potential payout is the implied probability. Pay 60¢ for something worth $1 if true, and you break even only if it comes true at least 60% of the time. The same logic runs on every prediction market — Kalshi, Polymarket, or anywhere else contracts settle to a fixed dollar.
But “price equals probability” is the first-day lesson, not the whole course. A price is only as trustworthy as the market behind it, and reading one well means checking the spread, the depth, and a couple of known biases before you treat 60¢ as 60%. Here’s how to do that, with numbers.
What does a 60¢ price actually tell you?
Run the arithmetic on one contract. You buy Yes at 60¢. If the event happens, the contract settles at $1 and you make 40¢. If it doesn’t, you lose your 60¢. Your expected value is zero exactly when the true probability is 60%: (0.60 × 40¢ gain) − (0.40 × 60¢ loss) = 0.
Now suppose you believe the true probability is 70%. Your expected value per contract becomes (0.70 × $1) − $0.60 = 10¢ — before fees. If you believe it’s 50%, the same contract has an expected value of −10¢. The price isn’t telling you what will happen; it’s telling you the probability at which buying and holding is a coin flip in expectation. Everything above that number is edge someone thinks they have; everything below it is edge someone on the other side thinks they have.
This also means a losing trade wasn’t necessarily a bad read, and a winning one wasn’t necessarily a good one. A 60% event fails four times in ten. A single resolution tells you almost nothing about whether the price was right.
A prediction market price isn’t a prediction of what will happen — it’s the probability at which the market can’t find anyone left to disagree.
Why don’t Yes and No prices sum to $1?
They do at settlement — one side gets $1, the other gets $0, always. What don’t sum to $1 are the prices you can actually trade at right now, and the gap is the bid-ask spread.
On an order book exchange, buying No is the mirror image of selling Yes. Say the best bid for Yes is 58¢ and the best offer is 60¢. If you want Yes immediately, you pay the offer: 60¢. If you want No immediately, you’re effectively selling Yes at the bid, which prices No at $1 − 58¢ = 42¢. Add them up: 60¢ + 42¢ = $1.02. That extra 2¢ isn’t a fee and isn’t an arbitrage you’re missing — it’s the spread, the cost of demanding immediacy from whoever is quoting both sides.
The market’s “real” estimate sits somewhere inside the spread. With a 58/60 market, the midpoint is 59¢, and quoting “the price” as 60¢ modestly overstates the probability while 58¢ understates it. In a tight, liquid market the distinction barely matters. In a thin market quoted 45/55, “the price” is doing a lot of hiding: the market is telling you it thinks the answer is somewhere around 50%, plus or minus five points, and nobody cares enough to narrow it down.
How do you read the order book?
The order book is the list of resting limit orders — what people are willing to pay (bids) and willing to accept (asks). A simplified Yes book might look like this:
| Side | Price | Contracts |
|---|---|---|
| Ask | 62¢ | 5,000 |
| Ask | 61¢ | 1,200 |
| Ask | 60¢ | 300 |
| Bid | 58¢ | 250 |
| Bid | 57¢ | 2,000 |
| Bid | 55¢ | 8,000 |
Three things to pull out of it. First, the touch: best bid 58¢, best ask 60¢, spread 2¢, midpoint 59¢. Second, the depth: only 300 contracts are offered at 60¢. If you tried to buy 1,500 Yes at market, you’d take all 300 at 60¢, then 1,200 at 61¢ — your average price is about 60.8¢, not 60¢. The quoted price is only the price for a small order. Third, the shape: big resting size stacked a few cents away on both sides suggests market makers are engaged and the price is being actively defended; a book that’s 300 contracts deep and then empty suggests the price is one motivated trader’s opinion.
When should you distrust a price?
A price deserves skepticism roughly in proportion to how cheap it is to move. Concrete warning signs:
- Wide spreads. A 45/55 market has an implied probability of “somewhere around 50%, we’re not sure.” Treat the midpoint as a rough estimate with a wide error bar, not a forecast.
- Shallow depth. If a few hundred dollars can move the price five points, the price reflects whoever traded last, not aggregated opinion. This is common in brand-new markets and obscure questions.
- Stale quotes. Some markets go hours without a trade. A “last price” from before relevant news broke is a fossil, not a forecast. Check the timestamp and the current bid/ask, not the last trade.
- Long-dated, low-attention questions. Markets are best where informed traders have a reason to show up. A market on next week’s CPI print attracts people who model CPI. A market on an event three years out mostly attracts noise and time-value-of-money discounting.
- Resolution ambiguity. Sometimes a weird price is the market correctly pricing uncertainty about how the contract’s fine print will be settled, not about the event itself. Read the resolution rules before concluding the market is wrong.
Why do prices move?
A contract has no cash flows, no dividends, no earnings — nothing to reprice except the probability itself. So prices move for exactly two reasons: new information, and new money.
Information moves are the ones you want. A Fed governor gives a hawkish speech and the September rate-cut contract drops from 72¢ to 61¢ within minutes — that’s the market re-estimating in real time, and it’s the core reason these prices are useful as forecasts.
Money moves are the ones to be careful with. A large buyer hitting a thin book can push a price several points without any news at all; the price then usually drifts back as other traders fade the move. When you see a sharp move, the first question is always: what happened? If you can’t find the news, consider that the answer might be “someone with a big account had an opinion.”
Implied odds vs. true odds: fees and the longshot problem
Two systematic wedges sit between the printed price and the true probability.
Fees. As of August 2026, Kalshi charges takers a trading fee of roughly 7% of expected profit, by the formula fee = round up(0.07 × contracts × price × (1 − price)), per its published fee schedule (July 2026 revision) — resting maker orders pay nothing on standard markets. On 100 contracts at 60¢, that’s 0.07 × 100 × 0.60 × 0.40 = $1.68. You paid $60 plus $1.68, so your true breakeven probability isn’t 60% — it’s about 61.7%. Polymarket runs a similar structure: as of August 2026 its taker fees use the same price × (1 − price) formula with rates from 0.04 to 0.07 depending on category, geopolitics markets are fee-free, and makers pay nothing. Fees on both platforms peak at 50¢ and shrink toward the extremes — which matters most exactly where the next wedge lives.
Longshot bias. Across betting and prediction markets, low-priced contracts have historically been overpriced relative to how often they actually happen: research on the favorite-longshot bias (see Snowberg and Wolfers, NBER) finds that longshots win less often than their odds imply, while heavy favorites win slightly more often than their prices suggest. A contract at 3¢ is not reliably a 3% event; the true frequency for that bucket has tended to run lower. The structural reasons are mundane: some traders enjoy cheap lottery tickets, and the people who know the longshot is hopeless face a lousy trade — risking 97¢ to win 3¢, with fees and capital tied up until resolution — so the correction never fully happens. Practical translation: trust prices most in the 20–80¢ range, and mentally haircut the extremes.
Quick answers
What does 60 cents mean on Kalshi? The market estimates a roughly 60% probability the event happens. The contract pays $1 if it does and $0 if it doesn’t, so 60¢ is the price at which buying breaks even, before fees, if the true odds are 60%.
Why don’t the Yes and No prices add up to $1? Because you’re quoted the ask on each side, and the gap between bid and ask is the spread. A 58-bid/60-ask Yes market prices No at 42¢ to buy; 60 + 42 = 102¢, and the extra 2¢ is the spread, not a fee. At settlement the two sides always sum to exactly $1.
Is the price the market’s actual probability estimate? The midpoint between bid and ask is the better estimate, and it’s only as good as the market is liquid. Tight spread and deep book: take it seriously. Wide spread and thin book: treat it as a rough guess.
Do fees change the odds I need to win? Yes. As of August 2026, Kalshi’s taker fee is about 7% of expected profit (0.07 × price × (1 − price) per contract), which pushes the breakeven on a 60¢ contract to roughly 61.7%. Polymarket’s taker fees run 0.04–0.07 by category with the same formula; makers on both platforms generally pay nothing.
Are cheap longshot contracts good value? Historically, no. The favorite-longshot bias means low-priced contracts have tended to pay off less often than their prices imply, so a 3¢ contract is usually a worse-than-3% proposition.
Why did a price move with no news? Probably a large order hitting a thin book. If you can’t find the information behind a move, the move itself is weak evidence — watch whether the price holds or drifts back as other traders react.