Pricing & Edge · EV

What is expected value (EV) in sports betting?

Expected value is the average amount a bet wins or loses per dollar staked if you could place the exact same bet an infinite number of times. It combines two inputs: how likely the bet is to win, and what the price pays when it does. A bet is positive expected value (+EV) when the probability you assign to winning is higher than the probability the odds imply — and it is −EV when the reverse is true.

The formula is straightforward. Multiply the probability of winning by the profit the bet pays, then subtract the probability of losing times the stake. Divide by the stake and you have EV as a percentage: the return you expect, on average, every time you make that bet.

Why EV is the only number that matters

Any single bet is mostly noise. A 60% favorite loses four times in ten, and a +300 longshot cashes often enough to feel encouraging. Over hundreds of bets, though, results converge toward the sum of each bet’s EV. Bettors who consistently take prices better than the true probability make money; bettors who take worse prices lose it, no matter how their last week felt.

This is why professional bettors talk about process rather than results. You cannot control whether tonight’s over hits. You can control whether the price you accepted was better than the fair price. EV is the scoreboard for that decision.

Where the fair probability comes from

EV is only as good as the probability you feed it. The two common sources are a devigged market — removing the book’s margin from both sides of a liquid market to recover the prices underneath — or a model that estimates the outcome from data. OddsGuy’s board does both: it devigs the market across books and compares that fair number against the best price available, then ranks every prop by the resulting EV%.

One honest caveat: a computed +EV is an estimate, not a promise. Small edges sit inside wide error bars, which is why staking discipline and volume matter as much as finding the edge in the first place.

Worked example

Worked example — fair 55% at −110

Market price
−110
Implied probability of −110
110 ÷ (110 + 100) = 52.4%
Your devigged fair probability
55.0%
EV per $110 staked
0.55 × $100 − 0.45 × $110 = $55.00 − $49.50 = +$5.50
EV per unit staked
$5.50 ÷ $110 = +5.0%
fair 55.0%mkt 52.4%

You expect to keep +5.0% of every dollar bet at this price, on average, over the long run. That gap between fair and implied is the entire ballgame.

See it live

Related terms

Frequently asked questions

Does a +EV bet guarantee a win?

No. A +EV bet is a good price, not a locked outcome. A bet with a 55% chance to win still loses 45% of the time. EV tells you what happens on average across many repetitions of the same decision, which is why volume and bankroll discipline matter more than any single result.

What EV% is considered good?

Anything consistently above 0% beats the market, but in practice edges of +2% to +6% are what disciplined bettors build around. Treat anything advertising double-digit EV with suspicion — liquid markets rarely misprice by that much, and the error is usually in the estimate, not the odds.

How do I find +EV bets in practice?

Devig a liquid market to get a fair probability, then compare that number to the best price available across books. When the price implies a lower probability than fair, the bet is +EV. The OddsGuy scanner automates this for MLB player props across every tracked sportsbook.

Why does the scanner rank by EV% instead of win probability?

Because win probability ignores price. A 75% favorite at −400 is a worse bet than a 50% shot at +120 if the fair prices say so. EV% folds probability and price into one number, which is the only honest way to rank bets against each other.

The morning board, in your inbox

Top +EV rows, biggest line movers, and hot streaks — one email before first pitch. No spam, unsubscribe anytime.