Calculating Your Edge in Sports Betting — The Formula and 3 Worked Examples

'Edge' is the most important and most abused word in sports betting. Tipsters advertise 'my edge is research'. Forum pros say 'you need edge'. Few actually calculate it. Edge isn't a property, it's a number — and the formula is no harder than a fraction.

Definition: Edge is the expected profit per £1 staked, expressed as a percentage. Formula: Edge = (decimal odds × true probability) − 1. Positive = value bet. Zero = fair (rare). Negative = losing bet (the market default).

Why 'minus one'? Decimal odds include the stake: odds 2.00 pays out £2 per £1 staked, which is £1 of profit. Multiplying odds × probability gives the expected payout per £1. Subtracting the stake (1) leaves expected net profit — your edge.

Example 1 — Premier League home win: Odds 2.20, you estimate true probability at 50%. Edge = 2.20 × 0.50 − 1 = 0.10 = 10%. On a £100 stake, expected value is +£10. You may lose this particular bet — across 1,000 bets of this kind you'd average £10,000 profit (with perfect estimation).

Example 2 — tennis underdog: Odds 4.50 on the underdog, your model says 25%. Edge = 4.50 × 0.25 − 1 = 0.125 = 12.5%. High edge at high odds means high variance. You'll lose 75% of bets and only profit through rare wins. Psychologically hard, mathematically still +EV.

Example 3 — NBA moneyline: Odds 1.55, your estimate 60%. Edge = 1.55 × 0.60 − 1 = −0.07 = −7%. Despite a 'logical' bet on the favourite you lose an expected £7 per £100. The bookmaker priced the probability higher than you did — if your model is right, his odds are too short.

Implied probability from odds: Bookmakers communicate their estimate implicitly. Formula: implied probability = 1 ÷ odds. Odds 2.00 → 50%. Odds 1.50 → 66.7%. Odds 4.00 → 25%. Edge only exists when your own estimate deviates positively from the implied probability — and by enough to beat the bookmaker margin.

The bookmaker margin (overround or vig): In a 2-way market (e.g. a tennis match) implied probabilities sum to typically 102–106%. Example: odds 1.80 / 2.00 → 55.6% + 50% = 105.6%. The excess 5.6% is the margin. To estimate the fair market, divide each probability by the sum: 52.6% / 47.4%. Only these are what the bookmaker 'really believes'.

Edge vs. ROI: Edge is per bet, ROI is across many bets. At a constant 3% edge across 1,000 bets at £100 each, expected ROI is about 3% of total turnover = £3,000. In practice ROI varies around this value due to variance. Realistic pro bettors achieve 1–3% ROI over thousands of bets — anyone promising '+20% ROI' is either lying or working with a sample of 50.

From edge to stake — the Kelly link: Once you've calculated edge, the Kelly criterion answers the stake question: f* = (b·p − q) / b, where b = odds − 1, p = your probability, q = 1 − p. At 5% edge on odds 2.00, full Kelly is 5% of bankroll. Most practitioners stake Half-Kelly (2.5%) or Quarter-Kelly (1.25%) to reduce variance — see 'Half-Kelly vs. Full Kelly'.

Common edge mistakes: (1) Probability estimated by gut feeling rather than model — your 'edge' is wishful thinking. (2) Edge not validated by CLV — if you're betting 5% edge but systematically getting beaten by the closing line, you have no edge. (3) Ignoring margin — a '0.5% edge' bet vanishes the moment the bookmaker moves the line by one tick.

Tools and validation: Our Kelly criterion calculator computes edge, implied value and optimal stake in a single input. The Dutching calculator shows edge across multiple selections at once. For historical tracking: a spreadsheet with date, match, odds, your probability, closing odds — the 'your odds vs. closing odds' comparison is the most honest edge test over 200+ bets.

Related articles: 'Value betting explained', 'Half-Kelly vs. Full Kelly', 'Understanding sports betting odds' (in the sports betting hub). External source: definition of expected value at Wikipedia. Glossary: 'Edge', 'Probability', 'Expected value'.

Bottom line: Edge isn't an idea, it's a calculation. Odds × probability − 1. Anyone who doesn't run this formula for every bet is betting blind. Anyone who runs it but estimates probability from the gut is fooling themselves. Real edge comes from a calibrated model, validated through closing line value, deployed with Half-Kelly. Everything else is gambling with better vocabulary.