Reading prices and expected value

A prediction-market price tells you what the market collectively thinks the probability of an outcome is. Once you can read the price as a probability, expected value (EV) is a useful number for describing the gross payoff implied by a probability estimate — and how it stacks up against the same trade on a different venue.

This guide walks through that path end-to-end:

1. Reading a Yes price as an implied probability.
2. Computing EV on a single trade.
3. Comparing the same market across venues.
4. The things EV does not tell you.

If you haven't read the structural overview yet, start with How prediction markets work.

Reading a Yes price as a probability

Each binary contract pays $1 if its named outcome happens and $0 otherwise. Because of that simple payoff, the price you pay today is also the implied probability the market is assigning:

• A YES contract trading at 40¢ implies the market thinks YES is roughly 40% likely.
• A YES contract at 62¢ implies ~62%.
• A NO contract at 38¢ implies the NO outcome is ~38% likely (i.e. YES at ~62%).

Two practical implications:

• No odds conversion. You don't have to translate American ("+150"), decimal ("2.50"), or fractional ("3/2") odds into probability. The price already is the probability.
• YES + NO ≈ $1. They're complementary instruments. Best asks usually sum a bit above $1 and best bids a bit below; the gap is the spread.

If the price doesn't match your own probability estimate, the difference is an edge under that estimate. EV is the tool for quantifying that gap before fees, slippage, and variance.

Expected value, in one formula

For a binary contract that pays $1 on the winning side and $0 otherwise, expected value per contract is:

// market price reads as the market's implied probability
// yourProbability is your own estimate, in [0, 1]
const evPerContract = (yourProbability - marketPrice) * 1;

Positive EV when your probability sits above the market's. Negative EV when it sits below. The size of the gap is your edge.

A worked single-venue example

Suppose Polymarket prices "Will it rain in NYC tomorrow?" YES at $0.40. You think the actual probability is 55% because three weather services agree. Mechanically:

Sum the two: EV = +$0.15 per contract. That's a 38% expected return on the $0.40 you put up — before fees and assuming your 55% estimate is well-calibrated.

The same math works for buying NO: substitute "yourProbability of NO" for the market's NO price. YES and NO are mirrored instruments, so you only ever need to compute EV for the side you'd actually buy.

Comparing the same market across venues

The same real-world question often shows up on multiple venues with different prices. Your belief is the same; your EV per dollar isn't.

For example, a 55% probability estimate produces different gross EV on a 47¢ YES quote than on a 49¢ YES quote. The lower quote has a larger gross edge, but net comparison still needs current fees, spread, slippage, funding costs, and rule matching.

Two things to compare:

• EV per contract. Strict math: (yourProbability − price) × $1.
• EV per $1 risked. EV per contract divided by the price you paid. This is the more useful comparator across venues, because contract sizes (and currencies, in the cross-platform case) can vary.

In that example, both venues are positive gross EV under the same belief. The 47¢ quote has the larger gross EV per dollar at risk because the same 55% belief sits farther above the entry price. Net EV may still differ after platform fees and execution costs.

Including fees in the EV math

The EV formula above is a gross EV — it ignores fees. Net EV subtracts the platform fees from the payoff side:

// netPayoutWin: $1 minus any payout-side fee, if one applies
// tradingFee: per-contract fee paid up front (always)
const netEv =
  yourProbability * (netPayoutWin - price)
  + (1 - yourProbability) * (0 - price)
  - tradingFee;

On thin edges, fees can flip a positive gross EV estimate to negative net EV. Compute both gross and net before comparing markets. The Polymarket fees guide and the Kalshi fees guide detail what the fee buckets look like on each venue.

What EV does not tell you

EV is the long-run average if you could repeat the same trade many times under the same conditions. It does not tell you:

• How any single trade will resolve. A +EV trade can lose, sometimes several in a row. Variance is real even when EV is positive.
• How much to size. EV is per-contract; sizing is about bankroll fraction and variance tolerance. See the Polymarket first-trade guide for the sizing framework.
• How calibrated your probability is. If you systematically overstate your probabilities by 10 points, the trades that look +EV will tend to be break-even or worse in practice.
• Liquidity / slippage. A quoted price is only good for top-of-book size. Larger orders walk into worse prices on a book, or eat AMM impact, both of which erode EV.

How EdgeLedger surfaces this

EdgeLedger doesn't quote you a "this trade is +EV" badge — that judgement rests on your probability estimate, which the system can't know. What it does surface:

• Realized P&L on closed trades, by market and category.
• A portfolio P&L curve showing whether your historical EV estimates actually played out.
• Cross-venue price data in the Pro arbitrage scanner so you can spot the gap between Polymarket and Kalshi prices for the same question.

Use these to audit your calibration over time. The math in this guide is only as useful as the probabilities you plug in — the audit is how you keep those probabilities honest.

Where to go next

• How prediction markets work — if you want the structural overview behind the price.
• Polymarket: Placing your first trade — the buy-side mechanics for putting an EV view into a position.
• Kalshi: Placing your first contract trade — the regulated US counterpart with cents pricing.