Expected Value Trading: The EV Math for Event Contracts
Buy a Kalshi contract at exactly the price you think is fair, and you've already locked in a loss. Not a big one — a few cents per 100 contracts — but a real one, because the taker fee is subtracted after your edge, not before. Expected value trading is the discipline of doing that subtraction before you click buy, not after the fill. On a $0.00–$1.00 contract that pays $1 or $0, the formula is short: EV = (your probability estimate × payoff) − (cost + fee). Everything else is just plugging in real numbers.
The formula, and why the payoff side is fixed
On Polymarket and Kalshi alike, a winning share redeems for $1 and a losing share for $0 — there's no variable payout to estimate, unlike a stock or a sportsbook parlay with odds attached. That simplifies EV down to one moving part: your probability estimate versus the fee-adjusted cost of entry. If you believe an event happens with probability p, and it costs you C per share (price plus fee) to buy 100 shares, EV = 100 × p − 100 × C. Anything above zero is expected-value positive at your stated edge; anything at or below zero isn't, no matter how confident you feel about the pick.
Worked example: Kalshi at 50 cents
Buy 100 YES contracts on Kalshi at $0.50. The taker fee is round_up(0.07 × 100 × 0.50 × 0.50) = round_up($1.75) = $1.75 — the maximum on Kalshi's fee curve, since fees peak at the 50-cent midpoint. Total cost: $50 + $1.75 = $51.75. Payoff if correct: $100.
Breakeven probability = cost ÷ payoff = 51.75 ÷ 100 = 51.75%. The market is pricing this at 50%. If your own estimate is also 50%, EV = (0.50 × $100) − $51.75 = −$1.75 — a guaranteed loss equal to the fee, because you have zero edge over the crowd. You need a probability estimate above 51.75% — at least 1.75 percentage points of real edge over the quoted price — before this trade is expected-value positive.
Worked example: Polymarket at 50 cents
Buy 100 YES shares on a Politics-category Polymarket market at $0.50. Polymarket's 2026 category taker fees are symmetric around 50% and peak there — for Politics, Finance, and Tech markets, the maximum is $1.00 per 100 shares. Total cost: $50 + $1.00 = $51.00. Breakeven probability = 51 ÷ 100 = 51% — you need just 1 percentage point of edge over the quoted price to clear this fee, versus 1.75 points on the equivalent Kalshi trade at the same price. That gap is entirely a function of each venue's fee schedule, not the odds themselves, and it's why comparing fee structures matters as much as comparing prices when the same event trades on both venues — see our Polymarket vs. Kalshi comparison for the full fee table.
Where the required edge shrinks: away from 50 cents
Both venues taper their fees toward the extremes. On Kalshi, buying 100 YES contracts at $0.90 costs a fee of 0.07 × 100 × 0.90 × 0.10 = $0.63, versus $1.75 at 50 cents. Total cost: $90.63. Breakeven probability = 90.63 ÷ 100 = 90.63% — you need only 0.63 percentage points of edge over the 90% quoted price, a much smaller bar than the 1.75-point requirement at the midpoint. This is the same shape described in Polymarket's fee design, which also tapers toward its 1-cent/99-cent extremes. The practical read: high-confidence, near-resolved contracts require less edge to clear their fee than 50/50 coin-flip markets do — but they also offer far less room for that edge to be wrong.
| Trade | Price | Fee (per 100) | Total cost | Breakeven probability |
|---|---|---|---|---|
| Kalshi YES @ $0.50 | $0.50 | $1.75 | $51.75 | 51.75% |
| Polymarket YES @ $0.50 (Politics) | $0.50 | $1.00 | $51.00 | 51.00% |
| Kalshi YES @ $0.90 | $0.90 | $0.63 | $90.63 | 90.63% |
When EV is actually positive
A trade clears the positive-EV bar in exactly one scenario: your genuine probability estimate, arrived at independently of the quoted price, sits above the breakeven probability the fee creates. That means EV trading isn't about finding contracts that "feel right" — it's about having a specific, defensible reason your estimate diverges from the crowd's by more than the fee requires. That reason might be model-based (a polling aggregate, a weather forecast, an economic release you can price faster than the book updates), or flow-based (reading who's trading, covered in our order-flow trading guide). Without a specific edge, the honest EV of any trade at the quoted price is negative by the fee amount — full stop.
Maker-side economics change this math entirely. Polymarket charges makers no taker fee at all and pays a rebate (25% of taker fees collected, 20% on crypto markets) back to resting orders, so a maker's breakeven probability sits at the raw share price with no fee drag — the entire calculus explored in market making explained. If you can get filled as a maker instead of paying to take, the edge requirement drops or disappears.
Sizing the trade once EV is positive
A positive EV number doesn't tell you how much to risk — it tells you the direction is favorable on average, across many repetitions, at your stated probability. A single 51.75%-edge trade at full account size is still a bad bet if you're wrong about your own edge, because variance on individual contracts is high even when the long-run math favors you. That's a bankroll question, not an EV question — see bankroll management for event trading for how to size positions so one wrong probability estimate doesn't wipe out ten right ones.
The same fee-aware math applies directly to cross-venue arbitrage, where the entire trade thesis is locking in a spread wide enough to survive both venues' fees — see our arbitrage fee math breakdown for the version of this calculation used when you're trading both sides at once instead of taking a single directional position.
Run the fee math before every fill, automatically
PolyMarketMaker's arbitrage scanner computes per-market fee math across Polymarket US, Polymarket global, Kalshi, and PredictIt in real time, so you see the fee-adjusted breakeven before you're in the trade, not after the confirmation. PolyMarketMaker surfaces the EV math this article walks through, live, on every quoted contract. Simulation $149/mo, Live Trading $299/mo.
FAQ
What does expected value mean in trading?
The average outcome of a trade if you could repeat it many times: your estimated probability of winning multiplied by the payoff, minus what the trade costs you to enter. Positive EV means the trade profits on average over many repetitions; negative EV means it loses on average, regardless of any single result.
How do you calculate EV on a prediction-market contract?
EV = (your probability estimate × $1 payoff per share) − (share price + trading fee). Because shares only pay $1 or $0, the payoff side of the formula is fixed — the entire calculation comes down to your probability estimate versus the fee-adjusted cost.
Do fees change whether a trade has positive EV?
Yes, directly. If your probability estimate exactly matches the quoted price, the trade's EV is negative by exactly the fee amount — you need to be more right than the market, not just as right as it, to clear the fee and turn a profit.
What's a realistic edge needed to overcome prediction-market fees?
It depends on price and venue. Near 50 cents, where taker fees peak on both Polymarket and Kalshi, the required edge is largest — roughly 1 to 1.75 percentage points per 100 shares on the 2026 fee schedules. Near the 1-cent/99-cent extremes, fees taper and the required edge shrinks.
Is a market maker's EV calculated differently than a taker's?
Yes — makers on Polymarket pay no taker fee and collect a rebate, so their breakeven probability sits closer to the raw share price with no fee drag, while takers on either venue must clear the taker fee before their edge translates into profit.
This article is for educational purposes only and is not financial advice. Trading involves risk of loss.