General information, not financial, investment, legal, tax or betting advice · Prediction markets carry risk of loss · 18+ or the legal age in your region
Prediction MarketIndex
Prediction Market Index/Learn/Expected value explained
Education pillarThe reference spine

Expected value explained, the maths behind a good or a bad bet.

Expected value is how you weigh a payoff against its probability. On an event contract it tells you whether a price is worth taking given your own honest estimate of the chance, and why the market price is the bar you have to beat.

By Fredrik FilipssonFounder and editor · Two decades in advisory, hospitality and mediaEditorial review by Morten Andersen · Last reviewed 27 June 2026
Last reviewed
27 June 2026
Reading time
About 11 minutes
Level
Beginner
In one screen
The direct answer

Expected value is the probability weighted average outcome of a decision. On a yes contract it is roughly your estimated chance minus the price you pay, before costs. Positive means worth considering, negative means not, at that price.

A key dated figure

Binary contracts settle at one dollar or zero, per Kalshi contract terms as of June 2026, so a 40¢ price implies a 40% chance and is positive expected value only if your own estimate is meaningfully higher, after fees.

The one honest thing

Positive expected value is not a profit and not a prediction. It describes an average over many decisions. Any single contract can lose, and your estimate of the chance can simply be wrong.

Last reviewed 27 June 2026 · Settlement and fee figures dated to each platform's published terms and verified on the review date.
Information, not advice. This page is general information, not financial, investment, legal, tax, or betting advice. Prediction markets carry a real risk of loss. You must be 18+ or the legal age in your region.
Quick answer

Expected value is the average result of a decision if you could repeat it many times, found by weighing each outcome by its probability. On a binary event contract that settles at one dollar or zero, buying yes at a price is positive expected value only when your own estimate of the chance is higher than that price, by enough to cover the fee and the spread. The market price is already the crowd's implied probability, so the whole task is having a better estimate than the price, which is genuinely hard. A positive number describes an average, not a single result, and the money is still at risk.

What it is

A way to weigh a payoff against its chance.

Expected value is the average result you would get if you could repeat the same decision many times. You find it by taking each possible outcome, multiplying it by the probability of that outcome, and adding the pieces together. The word average is doing the work here. Expected value does not tell you what will happen this once. It tells you what the decision is worth on average, which is exactly the right lens for a choice you will face again and again rather than only one time.

An event contract makes this unusually clean, because there are only two outcomes and they pay fixed amounts. On the regulated exchanges a binary contract settles at one dollar if the event happens and zero if it does not, per Kalshi contract terms as of June 2026. So if you buy a yes contract at a price you can call P, you have paid P and you will receive either one dollar or nothing. If your own honest estimate of the chance is a number you can call Q, the expected value of that contract, before any cost, is Q multiplied by one dollar minus the price P, which simplifies to Q minus P. The whole of the maths sits in that one line.

Read it slowly and it says something simple. A contract has positive expected value to you only when you believe the chance is higher than the price implies. Buy yes at 40¢ and you need to genuinely think the chance is above 40%. Buy at 70¢ and you need to think it is above 70%. The price is not a target to feel good about. It is a hurdle your estimate has to clear, and the gap between your estimate and the price, once costs are taken out, is the only thing that makes a trade worth doing.

Visual one

The two outcomes of one contract.

You buy one yes contract at 40¢. It will settle at one dollar or at zero.
If the event happens
Gain +60¢

Settles at one dollar, you paid 40¢.

If it does not
Lose −40¢

Settles at zero, you paid 40¢.

Putting the two together

If you think the chance is exactly 40%, expected value is (0.40 multiplied by +60¢) plus (0.60 multiplied by −40¢), which is zero. The trade is a wash. If instead you think the chance is 55%, expected value becomes (0.55 multiplied by +60¢) plus (0.45 multiplied by −40¢), which is +15¢ per contract before costs. Your edge is the 15 point gap between your estimate and the price.

A worked illustration using a 40¢ price and a one dollar settlement, per Kalshi contract terms as of June 2026. The numbers are an example of the method, not a quote or a prediction about any market.

Data table one

The same view, at different prices.

Expected value per yes contract when you privately estimate the chance at 55%, across the price you would pay, before and after the entry fee.
Buy priceImplied chanceYour estimateEV before feeEntry fee (about)EV after fee
30¢30%55%+25¢1.5¢+23.5¢
40¢40%55%+15¢1.7¢+13.3¢
50¢50%55%+5¢1.75¢+3.25¢
55¢55%55%1.7¢−1.7¢
70¢70%55%−15¢1.5¢−16.5¢

Methodology: a binary contract settles at one dollar or zero, per Kalshi contract terms as of June 2026, so EV before fee is your estimate minus the price. The entry fee is the Kalshi taker fee of 0.07 multiplied by price multiplied by one minus price, per its fee schedule for February 2026, and the example assumes you hold to settlement so there is no exit fee. Notice the last two rows: once the price reaches your own estimate, the fee alone makes the trade negative. This is illustrative arithmetic, not a forecast.

Visual two

Expected value falls as the price rises.

+25¢0−15¢price 55¢ = your estimate30¢55¢70¢Buy priceEV per contract

Expected value per contract for a fixed private estimate of 55%, before fees, as the buy price moves. It is positive to the left, where you pay less than you think the chance is worth, and crosses zero where the price equals your estimate. Illustrative shape, as of June 2026. Not a prediction.

The benchmark

Why the market price is the bar to beat.

The reason expected value is hard to capture in practice is that the price is not arbitrary. It is the crowd's implied probability, the chance that all the buyers and sellers together have settled on. To have positive expected value you do not just need a view, you need a view that is better calibrated than that crowd, and you need the gap to be large enough to survive the fee and the spread. A confident feeling is not an edge. An edge is a probability estimate that is genuinely more accurate than the price, and most of the time it is not, which is the uncomfortable but useful starting point. Our explainer on reading prices as implied probability covers how to translate a price into the chance it represents.

This is why expected value and the market price belong together. The price hands you the benchmark for free. Your job is to decide, honestly, whether your own number is different and better, and by how much. If you cannot say why your estimate should beat a price that already reflects a lot of public information, the safe assumption is that it does not, and the expected value of acting is at best zero before costs and negative after them.

Why it matters for you

A tool for decisions, not a promise of profit.

The most common mistake with expected value is to treat a positive number as a result rather than an average. Positive expected value says that if you made the same kind of well judged decision many times, the outcomes would average out in your favour. It says nothing about the next contract, which can lose even when the decision was sound. The gap between the average and the single result is variance, and variance can run against you for a long stretch before the average has any chance to show up. Understanding that gap is what keeps a run of losses from being read as proof the method was wrong, and a run of wins from being read as proof you have an edge you do not.

Because expected value is an average over repetition, it only does its work if you also survive the variance long enough to reach the average. That is where it connects to position sizing. Staking too much on any one contract, even one with positive expected value, can wipe out a balance before the averages arrive, which is the practical reason careful participants size positions small relative to the money they can afford to lose. Expected value tells you whether a decision is worth making. How much to stake on it is a separate question, covered in our guide to position sizing and bankroll.

Costs belong inside the calculation, not beside it. Every fee, every crossed spread, and every withdrawal charge lowers expected value directly, and a trade that is barely positive before costs is frequently negative once they are counted. The table above showed this plainly, where a fee alone turned a fair price into a losing one. The disciplined version of the method is to compute expected value using the price you will actually pay, including the spread you will actually cross, so the number reflects the real decision rather than a frictionless version of it. Our pillar on fees and how they affect returns works through the full cost.

One frequently cited idea deserves a careful note. The favourite and longshot bias is a pattern, observed in some betting and prediction markets, where very unlikely outcomes can be priced a little high and heavy favourites a little low relative to how often they actually occur. If real and stable, it would mean small systematic edges in certain price ranges. The evidence is contested. The size of the effect, and even whether it appears at all, varies across markets, sports, and time periods, and it is easily swamped by fees. We mention it not as a strategy but as a reminder that prices are produced by people and are not perfect, and that any apparent pattern has to clear the same cost hurdle as everything else before it is worth acting on.

What we did

We built the worked examples on this page from the published contract specifications and fee schedules for regulated US exchanges, read directly on 27 June 2026, in particular the one dollar or zero binary settlement and the per contract taker fee. The arithmetic is ours and is shown in full so you can follow it. We have not represented any figure as the outcome of a live funded trade, and we name no contract to buy. Where a claim is contested, such as the favourite and longshot bias, we say so rather than present it as settled.

The long run

Why expected value needs repetition to mean anything.

Expected value is a statement about the long run, and the long run is longer than most people expect. Imagine a decision that is genuinely worth a positive amount on average. Made once, it is mostly luck. Made a handful of times, luck still dominates and a losing streak is entirely ordinary. It is only across a large number of independent, well judged decisions that the average begins to assert itself and the positive expectation has room to show up in the results. This is the law of large numbers in plain clothes, and it has a hard practical consequence. A method can be sound and still lose for a long time, and a method can be unsound and still win for a long time, so the outcome of any short run tells you very little about whether the underlying decisions had positive expected value.

That is why expected value pairs so badly with impatience. If you need this trade to work, you are no longer making a decision that can afford to be averaged out, and the maths quietly stops applying to you. The participants who get the most from expected value are the ones who can treat any single contract as one of many, size each one small, and let a large number of decisions do the work. None of that turns a negative expectation positive. It simply lets a genuine positive expectation, if you really have one, survive the variance long enough to matter. Our guide to calibration and forecasting skill looks at how to tell whether your estimates are actually any good.

Where it shows up

The same maths across every category.

The expected value lens does not change from one kind of market to another, only the difficulty of forming a better estimate than the price does. On an economic indicator market the crowd is often pricing against published forecasts and data releases, so beating the price means knowing something the consensus has not already absorbed, which is rare. On election markets the price blends polls, models, and sentiment, and your estimate has to be better calibrated than that blend. On sports event contracts the lines move quickly and the cost of crossing the spread can be the deciding factor in whether a view is worth acting on at all.

In every one of these, the procedure is identical. Translate the price into the chance it implies, form your own honest estimate of the chance, subtract the price from your estimate to get the raw edge, then take out the fee and the spread you would really pay. What remains, if anything, is the expected value of acting. The categories differ only in how hard it is to make that raw edge large enough to survive the costs, and in most liquid markets, most of the time, it is not.

A note on risk,

Positive expected value does not make a trade safe. It describes an average over many decisions, not the result of the one in front of you, and any contract can lose. Prediction markets can lose you money. Stake only what you can afford to lose, never to chase a loss, and never on borrowed money. If it stops feeling like a free choice, step back. In the United States you can call or text the helpline on 1-800-GAMBLER or visit ncpgambling.org.

Where this matters

Carry this into prices, costs, and how much you stake.

Common questions

Answered plainly.

What is expected value in simple terms?

Expected value is the average result of a decision if you could repeat it many times. For a yes contract you multiply your estimated chance of the outcome by what you would gain if it happens, then subtract the chance it does not happen multiplied by what you would lose. If the total is positive the trade has positive expected value at that price, and if it is negative it does not. It is a way to weigh a payoff against its probability, not a promise about any single result.

How do I calculate expected value on a prediction market contract?

On a binary contract that settles at one dollar or zero, buying yes at a price of P costs P and pays one dollar if the outcome happens. If your own estimate of the chance is Q, the expected value per contract is Q minus P before costs. So a contract priced at 40¢ is positive expected value only if you genuinely believe the chance is above 40%, by enough to cover fees and the spread. Per Kalshi contract terms as of June 2026, settlement is one dollar or zero.

Why is the market price the benchmark?

The price is the crowd's implied probability, the chance the market collectively assigns to the outcome. To have positive expected value you need your own estimate to differ from that price and to be better calibrated than the crowd. Beating the price is hard precisely because it already reflects a lot of information, which is why a confident feeling is not the same as an edge.

Does positive expected value mean I will make money?

No. Positive expected value describes the average over many independent decisions, not the result of one trade. Any single contract can still lose, variance can run against you for a long time, and your probability estimate can simply be wrong. Expected value is a tool for making consistent decisions, not a guarantee of a profit.

How do fees change expected value?

Every cost lowers expected value and raises the probability you need before a trade is worth it. The trading fee, the spread you cross on entry and exit, and any withdrawal charge all come straight out of the result. A trade that is barely positive before costs is often negative once they are counted, so you should fold the full cost into the calculation rather than treat it as an afterthought.

What is the favourite and longshot bias?

It is a pattern, observed in some betting and prediction markets, where very unlikely outcomes tend to be priced a little too high and heavy favourites a little too low relative to how often they occur. The size and even the existence of the effect is contested and varies by market and period, so it is not a reliable rule. It is a reminder that prices are not perfect, not a recipe for a guaranteed edge.

Reviewed by Fredrik Filipsson, Editor, on 27 June 2026. Worked examples built from Kalshi contract terms and fee schedule for February 2026, verified against the source on the review date. The favourite and longshot bias is presented as contested.
The Forecast

Learn one useful thing a week.

The rules change fast. Get the changes that affect you, plain and current, not tips.

Independent. Every claim dated and sourced. No platform pays for its place.

No tips, no picks, no spam. Information, not advice. Unsubscribe anytime.