Expected value in betting is the average profit or loss a trade should produce if the same setup repeats many times. A positive EV trade pays more than its estimated probability justifies. A negative EV trade pays less. Traders still lose individual positive EV positions because outcomes vary. The calculation measures the decision, not the next result.
How do you calculate expected value in betting?
Multiply each possible result by its probability, then add the results. For a binary trade, the working formula is `EV = (win probability x profit if right) - (loss probability x amount lost if wrong)`. Smarkets explains the same calculation as the sum of every possible value multiplied by its probability.
Take a YES contract offered at $0.40. It pays $1.00 if the event occurs and $0.00 if it does not. A trader who estimates the true probability at 50% risks $0.40 to make $0.60. The EV is `(0.50 x $0.60) - (0.50 x $0.40)`, or $0.10 per share before fees and execution costs.
That 10-cent edge comes entirely from the trader's 50% estimate. If the event's true chance is 35%, the same contract has negative EV. The arithmetic is easy. Estimating the probability is the work.
What does positive EV mean in prediction markets?
Positive EV means your estimated probability exceeds the break-even probability implied by the price after costs. A 42-cent YES contract starts with a 42% break-even point before fees. If your research supports 49%, the raw edge is 7 cents per share. If your estimate is 38%, buying YES destroys value even when the contract later wins.
Prediction markets expose this comparison directly because contracts settle at $1.00 or $0.00. Polymarket's order book documentation describes bids, asks, midpoint prices, and last-trade prices. Traders should calculate against the price they can execute, not a stale last trade or a midpoint with no available size.
The prediction market trading strategies guide covers the research methods traders use to form those probability estimates. Domain knowledge, faster public data, and cross-venue price differences can each produce an edge. A confident opinion without a calibrated probability produces no usable EV number.
Why can a positive EV trade still lose?
A trader can lose a positive EV position because the calculation covers repeated decisions, not one outcome. A contract with a true 60% probability still settles against the trader 40% of the time. Ten good trades can produce a losing run. One winning trade can come from a bad price.
Traders test the estimate across a large record of forecasts. Group calls into probability bands, then compare each band with the observed resolution rate. Calls marked 70% should resolve true near 70% over a useful sample. That calibration check separates a measured edge from confidence after the fact.
Traders control how much variance a bankroll can survive through position size. Our prediction market position sizing guide shows how traders turn an edge estimate into risk limits. Event contracts can settle at zero. Traders can still lose an oversized positive EV position.
How do spreads, fees, and slippage change EV?
Traders receive less value after spreads, fees, and slippage. A model may value YES at $0.54 while the screen shows a $0.50 midpoint. If the best ask is $0.53 and the order moves the next level to $0.55, the trade has little or no edge at the actual fill.
Use the executable ask for a buy and the executable bid for a sale. Check depth for the full order size. Add venue fees where they apply. Estimate the exit price when the strategy depends on selling before resolution. Order book depth determines whether the displayed edge survives contact with the market.
Some traders also earn maker incentives. Polymarket says its liquidity rewards depend on competitive limit orders, qualifying distance from the midpoint, and market-specific requirements. Treat rewards as a separate cash flow. Never use an incentive to rescue a negative EV trade.
How should a trader use expected value?
Write down four numbers before entry: your probability, the executable price, the full cost, and the planned size. Calculate EV per share. Reject trades with no margin for estimation error. Record the forecast before the result can rewrite your memory.
- Price the event independently before checking the order book.
- Calculate with the executable bid or ask and the available depth.
- Track calibration across resolved markets and revise weak models.
Traders keep more of a modeled edge when they execute at the right price. Kairos puts Kalshi, Polymarket, and Predict.fun in one book with sub-second data, global best bid and ask, and low-latency execution. Compare the venues, price the contract, then open the Kairos terminal. See you in the order books.