What is Expected Value (EV) in Prediction Markets?

Expected Value (EV) is the single most important concept for anyone trading prediction markets seriously. It's the difference between "I have a hunch" and "I have a measurable edge". This guide breaks down what EV is, how to compute it for Polymarket and Kalshi positions, and why traders who internalize it consistently outperform those who don't.

## What Expected Value Actually Means

Expected Value is the probability-weighted average outcome of a decision. Imagine you could repeat the same trade thousands of times. The average profit (or loss) per trade is the Expected Value.

The formula for a binary market (YES/NO outcome):

EV per share = (your_probability × payoff_if_right) − ((1 − your_probability) × cost_if_wrong)

For Polymarket, if YES is trading at $0.40 and you believe the true probability is 55%: - Payoff if right: $0.60 (you bought at $0.40, contract resolves at $1.00) - Cost if wrong: $0.40 (you bought at $0.40, contract resolves at $0.00) - EV per share: (0.55 × $0.60) − (0.45 × $0.40) = $0.33 − $0.18 = +$0.15

A 15-cent positive expected value per share. If you bought 100 shares for $40 total, expected profit is $15. Played consistently, this is how money compounds.

The key insight: EV is positive only when YOUR probability estimate is more accurate than the market price implies. If everyone agrees the chance is 40%, the market is at 40%, and no one has edge. Trading is about finding markets where your estimate diverges from consensus — and being right often enough to compound the edges.

## Why Most Traders Lose Money

The math of EV is simple. The execution is brutal. Most retail traders lose money in prediction markets because they:

1. **Don't actually compute EV**. They trade on gut feeling, narrative, or recent news. 2. **Overestimate their probability accuracy**. They think they're 70% right; they're actually 50% right. 3. **Underestimate fees and friction**. A 5pp edge gets eaten by gas, slippage, and bad fills. 4. **Skip the math on losing trades**. They focus on the 3 wins, ignore the 7 losses.

Positive EV is real but requires honesty about your own probability estimates. Most people overestimate themselves systematically.

For more on this calibration problem, our [+EV trading guide](/blog/what-is-positive-ev-trading) covers how to test if your estimates are actually well-calibrated against reality.

## EV in Liquid vs Illiquid Markets

Two markets with the same "edge" can have very different practical EV after execution costs.

**Example A**: Liquid market trading at $0.40 with $50k daily volume. You buy 1000 YES shares at $0.40 average. Total cost: $400. Execution gives you exactly the displayed price.

**Example B**: Illiquid market trading at $0.40 but only $2k daily volume. You buy 1000 YES shares. Order book is thin. Your fill: 200 at $0.40, 300 at $0.42, 500 at $0.45. Average: $0.43.

You paid 3 cents per share more in Example B due to slippage. That's a 7.5% drag on your position before the market even moves. If your edge was 5pp, slippage just ate most of it.

This is why our [+EV markets guide](/blog/how-to-find-ev-markets-polymarket) emphasizes liquidity filtering before trading. Edge that disappears to execution costs isn't edge — it's a learning opportunity.

## Computing EV for Multi-Outcome Markets

Many Polymarket contracts aren't simple YES/NO. They might ask "Which of 5 candidates wins?" Each candidate has a separate YES contract.

To compute EV for such a market, sum the individual EVs across all positions you'd take:

Position 1: Candidate A YES at $0.30, you estimate 35%. EV/share = (0.35 × $0.70) − (0.65 × $0.30) = $0.245 − $0.195 = +$0.05

Position 2: Candidate B YES at $0.25, you estimate 20%. EV/share = (0.20 × $0.75) − (0.80 × $0.25) = $0.15 − $0.20 = −$0.05

Trading both positions cancels out unless you only take the positive-EV one (Candidate A).

Multi-outcome markets reward specialists who can identify the underpriced candidate. Generalists who bet "the favorite" without computing each option's specific EV usually lose money.

## The Time Component

Expected Value calculations assume you'll hold to resolution. But prediction markets have time risk:

- Capital tied up for months loses opportunity cost - Market sentiment shifts in unpredictable ways - New information arrives that changes the underlying probability - The market itself could be delisted or cancelled

A 5% expected return is great if it plays out in 2 weeks. Mediocre if it takes 6 months. Bad if your capital could earn 8% elsewhere over the same period.

When computing EV for actual trade decisions, factor in: - Expected time to resolution - Annualized return vs alternatives (treasuries, index funds, other markets) - Variance during the holding period (can you tolerate the swings?)

## EV with Imperfect Information

Real-world EV computations involve uncertainty about your own probability estimate. You think YES has 55% probability, but you might be wrong by 5-10pp in either direction.

Mathematically, this means your "true" EV is a distribution, not a point estimate. The right way to handle this:

1. Compute your best EV estimate 2. Compute the EV if you're 5pp too optimistic 3. Compute the EV if you're 5pp too pessimistic 4. If all three are still positive, trade with conviction 5. If the "too optimistic" scenario goes negative, size smaller

Most traders never do this exercise. They commit fully to their best estimate, then lose money when their estimate was off by a manageable amount. The trick is sizing positions so you survive being wrong by a modest margin.

For more on position sizing under uncertainty, see our [Kelly Criterion guide](/blog/kelly-criterion-position-sizing).

## Common Mistakes in EV Calculation

**Confusing probability with payoff**: people calculate "if it's 55% to happen and I bet $100, I make $55". That's not EV — that's confused. EV requires multiplying probabilities by the actual contract payoffs, which depend on entry price.

**Ignoring fees**: gas costs on non-Builder Polymarket trades, Kalshi commissions on profits. These reduce your effective EV.

**Anchoring on round numbers**: "I think it's about 50%" without rigor. Real EV calculations need real probability estimates, not round-number approximations.

**Forgetting that 0/100% don't exist**: "I'm 100% sure this is mispriced" is overconfidence. Real probabilities live in 5-95% range. Treating anything as certain is asking for catastrophe.

**Not accounting for resolution risk**: contracts can resolve in unexpected ways. A position that "should" pay $1.00 might pay $0.50 due to dispute or ambiguity.

## How Sophisticated Traders Use EV

For professional or semi-professional traders, EV becomes the central organizing principle:

- They reject trades with sub-3pp expected edge after costs - They size positions proportional to edge magnitude (Kelly Criterion math) - They track realized EV vs predicted EV monthly (calibration check) - They retire strategies whose realized EV consistently falls short of predictions - They focus on improving probability estimation, not finding bigger "edges"

The 5pp edge they take consistently beats the 15pp edge they take occasionally. Volume × consistent edge = real wealth growth.

## Tools

For computing EV on Polymarket, the Predite scanner does this automatically — comparing AI probability estimates against market prices and surfacing markets with significant edge. Helpful as a filtering tool, but final probability judgment is yours.

For manual EV computation: - Spreadsheet with formulas (Excel or Google Sheets) - Python calculator (more flexibility for complex multi-leg trades) - Mental math for quick checks (5pp × $0.40 cost = ~$0.02 per share expected profit)

## The Disciplined Approach

A useful daily exercise: pick 5 random Polymarket contracts. Without looking at the market price, estimate your own probability for each outcome. Then check the market price.

Where are you significantly different? Those are potential trades. Where do you agree with the market? Skip — no edge there for you.

This drill reveals: - How calibrated your estimates are - Which market types you have edge in (often narrow domains) - How market sentiment biases differ from your own

After 100 such drills, you'll know with confidence whether you have real edge — or whether you've been deceiving yourself.

## Bottom Line

Expected Value is the foundation of all profitable prediction market trading. Without it, you're gambling. With it, you have a framework for systematic decision-making.

The math is simple. The honesty about your own probability estimates is hard. Master both and you'll be in the top 10% of retail traders. Skip them and you'll be in the 80% who lose money slowly.

For broader strategy framework that builds on EV thinking, see our [+EV markets guide](/blog/how-to-find-ev-markets-polymarket), [risk management guide](/blog/risk-management-prediction-markets), and [common mistakes guide](/blog/common-mistakes-new-prediction-traders).