PolyMathBlogExpected Value in Prediction Markets

Expected Value in Prediction Markets — How to Calculate Your Edge

April 13, 2026 · PolyMath Team · 9 min read

The single most important question before entering any prediction market trade is not “will this happen?” — it's “is this mispriced?”

That's the difference between gambling and trading. And expected value (EV) is the math that answers it. This guide will show you exactly how to calculate EV on any prediction market, how to find mispriced contracts, and what separates traders who profit consistently from those who don't.


What Is Expected Value?

Expected value is the average outcome of a bet if you could repeat it infinitely.

EV = (P(win) × Profit) − (P(lose) × Stake)

Simple example:

You buy a YES share for 60¢

If YES: collect $1.00 (profit = 40¢)

If NO: lose your 60¢

You estimate the true probability at 70%:

EV = (0.70 × $0.40) − (0.30 × $0.60)

EV = $0.28 − $0.18

EV = +$0.10 per share → positive EV, good trade

If the true probability is only 50%:

EV = (0.50 × $0.40) − (0.50 × $0.60)

EV = $0.20 − $0.30

EV = −$0.10 per share → negative EV, bad trade


The Critical Distinction: Probability vs. Expected Value

This is where most beginners go wrong. The question isn't whether the event will happen — it's whether the market price is correct.

True probability = 70%, price = 70¢ → EV ≈ 0, no edge

True probability = 80%, price = 70¢ → positive EV, buy

True probability = 60%, price = 70¢ → negative EV, sell or skip

Key principle

Edge = Your probability estimate − Market price. If your estimate equals the market price, you're paying fees to stand still.


Where Does Prediction Market Edge Come From?

Before calculating EV, you need to know why you might have better information than the crowd. Genuine edges in prediction markets come from a few sources:

Domain Expertise

You understand a specific field deeply — medicine, policy, sports analytics, financial markets. You have priors the average trader doesn't.

Information Timing

You read the primary source. You tracked the regulatory filing. You saw the press release before it was widely covered. Markets lag news by seconds to hours.

Analytical Edge

You model the base rates. You track calibration across similar markets. You understand that "unprecedented event" markets almost always overprice the unlikely outcome.

Market Inefficiency

Some markets are thin. Small traders dominate. The "wisdom of crowds" requires a crowd — a market with 50 traders is not efficient.

If you don't have at least one of these, assume you have no edge and the trade is neutral EV at best.


How to Calculate EV in Practice: Step by Step

Scenario:A Polymarket market is priced at 58¢ on “Fed cuts rates in June.” You've read the Fed minutes and analyzed the inflation data. You estimate the true probability at 72%.

Step 1: Market price = 58¢ → implied probability of 58%
Step 2: Your estimate = 72%
Step 3: Edge = 72% − 58% = +14 percentage points

Step 4: Calculate EV per dollar invested

EV = (0.72 × $0.42) − (0.28 × $0.58)

EV = $0.3024 − $0.1624

EV = +$0.14 per share

For $1,000 invested (~1,724 shares): expected profit ≈ $241

Step 5: EV% = $0.14 / $0.58 ≈ 24% — strong positive EV.

Use the EV Calculator at PolyMath to do this in seconds for any market.


EV Across Multiple Platforms

One of the most powerful EV opportunities comes from the same event being priced differently across Polymarket, Kalshi, and Metaculus. The expected value is different on each platform for the same probability estimate.

Example with your estimate at 70% (platforms price at 62¢, 57¢, 68¢):

Polymarket (62¢)

EV = (0.70 × $0.38) − (0.30 × $0.62) = +$0.080 per share

Kalshi (57¢)

EV = (0.70 × $0.43) − (0.30 × $0.57) = +$0.130 per share

Metaculus (68¢)

EV = (0.70 × $0.32) − (0.30 × $0.68) = +$0.020 per share

Kalshi offers 60% more EV per dollar than Polymarket on this trade. This is why multi-platform analysis matters. The Arbitrage Scanner at PolyMath finds these spreads automatically.


The EV Threshold: When Is a Trade Worth Taking?

Not all positive EV trades are worth the transaction costs and cognitive load of managing them.

EV% per tradeAction
Below 3%Skip — doesn't cover costs and margin of error
3%–7%Consider if size is meaningful and edge is confident
7%–15%Strong trade — size appropriately with Kelly
Above 15%High conviction — verify your analysis twice, then size up

Common EV Calculation Mistakes

Mistake: Confusing Win Rate with Edge

"I win 65% of my trades" means nothing without knowing the EV of those trades. A trader winning 65% of trades at low stakes and losing 35% at high stakes can still be deeply negative EV. Track your expected value per dollar, not just win rate.

Mistake: Anchoring to the Market Price

When a market has been at 72% for a week, it starts to feel "right." Form your estimate before looking at the current price. Write it down. Then compare to the market.

Mistake: Ignoring Time Value

A trade with +10% EV that locks up capital for 8 months is very different from a +10% EV trade resolving in 2 weeks. Annualize your EV when comparing across different time horizons.

Mistake: Double-Counting Correlated Information

You have three data sources all pointing to the same conclusion. If all three derive from the same underlying information, you have one data point, not three. Overconfidence from correlated evidence is one of the most common analytical errors.


EV and the Long Game

Expected value is a long-run concept. Any single trade will either win or lose — EV is the average across thousands of similar trades.

  • A single loss doesn't mean your EV calculation was wrong
  • A single win doesn't validate a bad EV trade
  • You need volume to see if your EV estimates are accurate

After 50+ resolved markets, check whether your +10% EV trades actually returned ~10% on average. Use the Calibration Tracker to measure this automatically.


The Bottom Line

Before every trade, ask: What is my probability estimate — and why is it different from the market's? What is the EV per dollar at current prices? Does the EV exceed my threshold given the uncertainty in my estimate?

If you can't answer question 1, you have no edge. Don't trade. If the EV threshold isn't met, pass. Discipline is how EV compounds into profit.

Once you've identified a positive EV trade, you still need to size it correctly. Read our guide on prediction market bankroll management and the Kelly Criterion position sizing guide to complete your trading system.

📬 Get weekly market analysis & new calculator alerts

Free newsletter — no spam, unsubscribe anytime.

Sponsored

📢 Advertisement

Carbon Ads placeholder — set NEXT_PUBLIC_CARBON_SERVE_ID in Vercel to activate.