Value betting means backing an outcome only when the price on offer is higher than its fair price, the price at which a bet on it would break even in the long run. The gap is your edge: edge = offered price / fair price - 1, so with illustrative prices, $2.60 against a fair $2.45 is 2.60 / 2.45 - 1 = 6.1%. Whether the bet wins is not what makes it value. The price is.
Expected value (EV) turns that edge into money: the average profit or loss per dollar staked if the bet could be placed many times over. A value bet has positive EV and still loses often. Any one result is mostly luck, an edge of a few per cent can take thousands of bets to show in profit, and the fair price is your own estimate. So value betting is three jobs: finding prices above fair, testing your fair prices, and staking to survive the losing runs.
What makes a bet a value bet
A decimal price is also a probability: implied probability = 1 / decimal odds, so $4.00 implies 25%. Across a bookmaker's market those chances add up to more than 100%, because the bookmaker margin is built into every price.
A value bet is a price long enough that its implied chance sits below the real one. The fair price comes from your own model or ratings, a market with the margin taken out, or a busy exchange market, and fair odds weighs the three.
Likelihood does not decide value: in one market a $1.30 favourite can be priced too short for its chance while a $15.00 outsider is a value bet. Your betting edge measures how far a price sits above fair.
Expected value betting: what a value bet earns on average
EV is the average result of a bet if it could be repeated many times: EV per $1 = win probability x decimal odds - 1.
Example: Illustrative numbers, not a real game. You rate a team a 24% chance to win and a bookmaker offers $4.50. EV per $1 = 0.24 x 4.50 - 1 = 0.08, so the bet is worth 8 cents per dollar staked on average, or 25 x 0.08 = $2.00 on a $25 stake.
Your fair price for that team is 1 / 0.24 = 4.1667, about $4.17, and the edge is 4.50 / 4.1667 - 1 = 0.08. When your win chance is 1 / your fair price, edge and EV per $1 are the same number, which is why value betting and positive EV (+EV) betting name one idea. The expected value calculator works out EV from any stake, price and chance. At $4.00 the same 24% chance is worth 0.24 x 4.00 - 1 = -0.04: same judgement, a bet that loses 4 cents per dollar on average.
Risk: Betting involves risk. A +EV bet can still lose, and often does, and an EV figure built on a wrong estimate of the chance is not value at all. There is no guarantee of profit; see responsible gambling for limits and support.
How often a value bet loses at different prices
An edge changes the average, not the result of any one bet. Here are three bets with the same 8% edge, at illustrative chances:
| Price | Your chance | EV per $1 | Loses | Chance of five losses in a row |
|---|---|---|---|---|
| $1.80 | 60% | +8c | 40 times in 100 | 1.0% |
| $4.50 | 24% | +8c | 76 times in 100 | 25.4% |
| $13.50 | 8% | +8c | 92 times in 100 | 65.9% |
Each EV is chance x price - 1. Each streak is the chance of a loss to the power 5: for the $4.50 bet, 0.76 x 0.76 x 0.76 x 0.76 x 0.76 = 0.254. The longer the price, the more of a bet's value sits in rare winners, so losing streaks grow longer and results swing harder for the same edge. Variance in betting measures how hard.
EV is an average over many bets, not a result
A run of bets rarely lands on its expected value.
Example: Illustrative: 50 bets of $10, each at $4.50 on an outcome with a true 24% chance, so each carries the 8% edge above. A winner returns $45, so profit after 50 bets = winners x $45 - $500.
| Winners out of 50 | Profit | Chance of a result in this band |
|---|---|---|
| 8 or fewer | -$140 or worse | 12.1% |
| 9 to 11 | -$95 to -$5 | 32.5% |
| 12 to 14 | +$40 to +$130 | 35.3% |
| 15 to 17 | +$175 to +$265 | 16.3% |
| 18 or more | +$310 or more | 3.8% |
The chances come from the binomial distribution, which counts every way 50 independent bets can land. Expected profit = 50 x $10 x 0.08 = $40, the result of exactly 12 winners, which turns up only 13.1% of the time. The run finishes behind 12.1% + 32.5% = 44.6% of the time, though every bet in it was a value bet; 11 winners leave it $5 short.
So judge a value method by the prices it takes, not by a month of results. Sample size in betting shows how many bets profit needs before it says more than luck.
How to find value bets
The process is the same for any sport or race:
- Set your fair price before you look at bookmakers' prices, so theirs cannot pull your view.
- Compare as many bookmakers' prices on the outcome as you can; line shopping covers the habit.
- Keep only prices above fair by a minimum edge set in advance, big enough to cover the error in your fair price.
- Check the price is still on offer in your bookmaker account, because prices move.
- Log the price you got, and fill in the closing price once the event starts.
Example: Illustrative prices. A market with the margin taken out makes Team A a 45% chance, a fair price of 1 / 0.45 = 2.2222, about $2.22. Bookmaker A offers $2.20: edge = 2.20 / 2.2222 - 1 = -1.0%. Bookmaker B offers $2.35: edge = 2.35 / 2.2222 - 1 = +5.8%.
Step 3 is the judgement call. If your 45% could be 2 points out either way, the fair price could be anywhere from 1 / 0.47 = $2.13 to 1 / 0.43 = $2.33. Bookmaker B's $2.35 clears the whole range. A $2.28 would be +2.6% against the central estimate (2.28 / 2.2222 - 1) but -2.0% if the true chance is 43% (2.28 / 2.3256 - 1).
Value tends to turn up where one bookmaker has not caught up with a move elsewhere, or disagrees with the rest of the market. For racing, the usual reference near the jump is the Betfair exchange, which value betting in horse racing works through.
When your account already holds a promotion
A promotion already on your account can lift a bet's EV. A bonus bet normally pays only its winnings, so a $20 bonus bet at a fair price of $4.00 pays 20 x (4.00 - 1) = $60 one time in four. Its EV is 0.25 x $60 = $15, three quarters of its face value. How to value a betting promotion applies the method to refunds and boosts. A bonus bet that loses pays nothing.
Risk: Betting involves risk. An edge against your own fair price is only as good as that price, a quoted price can be gone before your bet lands, and promotions carry each bookmaker's terms. See responsible gambling for limits and support.
How to test whether your value is real
A wrong fair price can take thousands of bets to show up in results. The faster test is closing line value (CLV): compare each price you took with the margin-free price at the start, as CLV = your odds / margin-free closing odds - 1. Take $2.35 on a team whose margin-free close is $2.24 (illustrative) and you beat it by 2.35 / 2.24 - 1 = 4.9%.
By the start, team lists, scratchings and late money are already in the price. If your prices keep beating that close over a few hundred bets, your fair prices are probably sound; if they keep falling short, the value was probably never there, whatever the results say.
How much to stake on a value bet
Size each bet as a small share of your bankroll, the money you keep apart for betting and can afford to lose. The Kelly criterion turns an edge into a stake: Kelly share of bankroll = edge / (decimal odds - 1). For the 8% edge at $4.50 that is 0.08 / 3.50 = 0.022857, about 2.3% of the bankroll: $2,000 x 0.022857 = $45.71 at full Kelly, and $45.71 / 4 = about $11.43 at a quarter of it.
Kelly assumes your edge is exact, and an estimated one never is, so a quarter or a half of it is common; level stakes are simpler still. Kelly is a formula, not a B337 feature. Keep a deposit limit with each bookmaker as a hard cap.
Mistakes that turn value into losses
With illustrative numbers, most from the examples above:
| Mistake | What it costs |
|---|---|
| Treating one bookmaker's price as fair | In a market that adds up to 106%, a price 5% above that bookmaker's own is 1.05 / 1.06 - 1 = -0.9% against the margin-free price (proportional method) |
| Betting a price that has already moved | $2.25 instead of Bookmaker B's $2.35 cuts the edge from 5.8% to 2.25 / 2.2222 - 1 = 1.25% |
| Judging the method on a month of results | 50 bets at a real 8% edge finish behind 44.6% of the time |
| Raising stakes after a losing run | The edge per dollar stays put while more money rides on each bet: that is chasing losses, not value |
| Ignoring the bookmaker's terms | A bet voided under the terms, for example over an obvious pricing error, loses its value with it |
When value betting is the wrong tool
Value betting needs a fair price you trust, stakes a bookmaker will take and a bankroll that can ride out losing runs. Skip it when:
- you have no credible fair price, as in a thinly traded or one-off novelty market;
- your stakes are cut so low the edge is not worth the time: expected profit = edge x turnover, so 8% on $5 bets is 40 cents a bet;
- a run like the 50-bet table, behind close to one time in two, would leave you short or tempted to chase;
- betting has stopped feeling like a choice. The signs of a gambling problem are worth reading, and the National Gambling Helpline, 1800 858 858, is free and confidential, 24 hours a day, 7 days a week.
For bets you would make anyway, comparing prices still cuts the margin you pay. Value is one part of a betting strategy that also covers staking, records and restrictions.
Risk: Betting involves risk. A better price lowers the margin you pay, not the chance a bet loses, and there is no guarantee of profit. See responsible gambling for limits and support.
Where B337's +EV screener fits
Across sports markets, B337's +EV screener ranks bookmaker prices that sit above a fair price by their edge, filtered to the sports, markets and bookmakers you pick. The fair price is built by stripping the margin from many prices on the same selection, exchanges and global market references included, and combining them. You choose which prices feed it and how much margin comes off each, and the result is an estimate. Positive EV betting sets out the method.
Terminal Pro and Full Automation include the screener rows, and a Screeners plan, set up with the team, offers them on their own. A free account shows the live counts, with the rows locked. Every bet is yours to place, in your own bookmaker accounts, and bets placed through B337 use credits. Prices are read on a repeating cycle, so check each one with your bookmaker first. Betfair and bookmaker names are trade marks of their owners, and B337 is not affiliated with them.
For free and confidential support call 1800 858 858 or visit gamblinghelponline.org.au.
Risk: Betting involves risk. A screener row points at a price above an estimated fair price, not at a winner; the price may be gone by the time you bet, and bookmakers can restrict accounts and void bets. See responsible gambling for limits and support.