How to Calculate Expected Value (EV)

How to Calculate Expected Value in Prediction Markets

Calculating expected value (EV) is the single most important skill you can develop for trading Kalshi and Polymarket contracts, and it starts with one formula you should be running before every position you open. Expected value tells you what a bet is worth on average across a large number of repetitions, stripped of emotion, recency bias, and narrative. On prediction markets, where every contract resolves to either $1 or $0, EV math is unusually clean compared to sports betting or options trading. The problem isn't the arithmetic. It's that most traders skip it entirely and size positions off gut feel or a headline. This guide walks through the exact formula, where your probability estimate should come from, and how to avoid the calculation errors that quietly erode an otherwise sound trading process.

The Core Expected Value Formula for Kalshi and Polymarket Contracts

Every prediction-market contract pays out $1 if the event resolves YES and $0 if it resolves NO (or vice versa for the NO side). That binary structure makes the EV formula simple:

EV = (Probability of Win × Amount Won) − (Probability of Loss × Amount Lost)

Say a contract is priced at $0.40 for YES. If it resolves YES, you collect $1.00, a profit of $0.60. If it resolves NO, you lose your $0.40 stake. Your own estimate of the true probability determines whether that trade is positive EV. If you believe the real probability of YES is 55%, the math looks like this:

• EV = (0.55 × $0.60) − (0.45 × $0.40)
• EV = $0.33 − $0.18
• EV = +$0.15 per contract

That $0.15 isn't a promise, it's a long-run average. Positive EV means that if you made the identical trade hundreds of times under the same odds and the same true probability, your average return per contract would land near $0.15. Any single instance can still lose. That's the part traders new to this market structure struggle with most, and it's why bankroll management matters as much as the calculation itself.

Converting Market Price to Implied Probability Before You Trade

Before you can judge whether a contract is mispriced, you need to convert its price into the market's implied probability. On Kalshi and Polymarket, this conversion is direct because prices are already denominated in cents on a dollar: a $0.40 YES price implies the market thinks there's a 40% chance of that outcome. There's no vig or juice baked into the quote the way there is with a sportsbook moneyline, though the bid-ask spread functions similarly, effectively taxing you on entry and exit.

If you're used to American odds format from sports betting, converting between the two systems takes practice. For a full walkthrough of implied probability, vig, and how to read quotes across formats, see How to Read Prediction Market Odds. Once you can convert price to probability instantly, in your head, without a calculator, you can screen dozens of markets in a sitting and only stop to build a full model on the ones where your gut says the market price looks off from your own estimate.

Building an Honest Probability Estimate: The Hardest Part of the Math

The formula for EV is trivial. The estimate you plug into it is not. Overconfidence is the single biggest source of negative-EV trades that look positive-EV on paper, because traders anchor their probability estimate to what they want to be true rather than what the evidence supports.

A disciplined estimate should draw on multiple independent inputs rather than a single source: polling or forecasting data where available, historical base rates for similar events, current market pricing on Kalshi versus Polymarket (divergence between the two often signals where the mispricing sits), and any structural or timing factors specific to the contract. When you're deciding which platform's liquidity and pricing to trust more for a given category, the comparison in Kalshi vs Polymarket 2026 is worth reviewing, since spread and depth differ meaningfully by market type across the two venues.

A useful discipline: write your probability estimate down before you check the market price. If you estimate first, then look, you avoid the common failure mode of reverse-engineering a probability that happens to justify the trade you already wanted to make.

Why EV Math Alone Doesn't Guarantee a Profitable Trading Strategy

Positive expected value on a single contract tells you nothing about whether your overall approach to prediction market strategy is sound. Variance is real, and a string of positive-EV trades can still produce a losing month, especially in low-sample categories like single-game sports outcomes or one-off political events with no repeatable structure.

This is where position sizing enters the picture. The Kelly criterion, a formula for sizing bets as a fraction of bankroll proportional to your edge, is the standard tool traders use to convert a positive EV estimate into a stake size that survives variance rather than getting wiped out by it. Sizing too aggressively on a genuinely positive-EV trade is one of the fastest ways to blow up an account, because a string of even modestly unlikely losses compounds fast against an oversized position.

The other failure mode is correlation. If you hold five positive-EV positions that all depend on the same underlying event, like multiple contracts tied to one Fed decision, you don't have five independent edges. You have one large, concentrated bet wearing five different tickets.

Applying EV Calculations Across Sports, Politics, and Economic Markets

The formula doesn't change by category, but the reliability of your probability inputs does. Sports markets on Kalshi and Polymarket often move fast around injury news, lineup changes, and live win-probability shifts, which means your EV calculation can go stale within minutes. If you're building a systematic approach to sports contracts specifically, the platform comparison in Best AI for Sports Betting covers which tools handle real-time line movement well.

Political and economic markets, by contrast, tend to move on a slower cadence tied to polling releases, data prints, and scheduled events, which gives you more time to build a careful estimate but also more time for the market to have already absorbed the obvious information. In both cases, the discipline is the same: don't trade a contract until you can state your probability estimate as a specific number, not a vague feeling that "this seems likely."

How PillarLab AI Fits Into This

Manually building a probability estimate for every contract you're considering doesn't scale, especially when you're screening across dozens of Kalshi and Polymarket markets in a single session. PillarLab AI is built around a structured 9-pillar analysis framework that breaks each market down into the components that actually drive an accurate probability estimate: news and event catalysts, historical base rates, market microstructure and liquidity, cross-platform pricing divergence, sentiment signals, and more, pulling from real-time Kalshi and Polymarket data rather than stale snapshots.

Instead of eyeballing a contract price and guessing whether it's mispriced, you get a structured breakdown that surfaces where the market's implied probability and the underlying fundamentals disagree, the raw input every EV calculation depends on. The edge-detection layer flags contracts where that gap is wide enough to be worth a closer look, so you spend your time building conviction on a shortlist instead of manually screening every open market on both platforms. For traders running the EV formula seriously across a large market universe, that screening layer is the difference between doing this analysis on ten contracts a week versus ten contracts a day.

Frequently Asked Questions

What is a good expected value for a prediction market trade?

There's no universal threshold. What matters is that EV is positive after accounting for spread and fees, and that your probability estimate is well-supported rather than optimistic guesswork.

Can a positive EV trade still lose money?

Yes. EV describes the average outcome across many repetitions, not any single trade. A positive-EV position can lose, just as a negative-EV one can occasionally win.

How do I estimate true probability for an EV calculation?

Combine independent sources: historical base rates, polling or forecasting data, cross-platform pricing, and event-specific factors. Write your estimate before checking the market price.

Does the Kelly criterion relate to expected value?

Yes. Kelly uses your EV and edge size to calculate a bankroll-proportional stake, helping you size positive-EV trades so variance doesn't wipe out the account.

Is EV calculation different on Kalshi versus Polymarket?

The formula is identical since both use $0-$1 binary contracts. What differs is liquidity, spread, and sometimes pricing itself, which can create cross-platform EV discrepancies.

Running this math manually on every contract you're considering is slow and easy to skip under time pressure. Start free with 10 credits and let the 9-pillar framework surface the probability gaps worth calculating by hand.