Guide
Value betting explained
Last reviewed August 2026
What a value bet is
A bet has value when the price you are offered pays more than the outcome's real chance of happening justifies. That is the whole definition, and it contains no opinion about who wins.
Decimal odds and probability are the same number seen from two sides. A price of 2.00 implies a 50% chance; invert any price with 1 / odds and you have what the bookmaker is claiming. So if you believe an outcome is genuinely 46.38% likely, the price at which the bet is exactly neutral is 1 / 0.4638 = 2.156. Anything above that is value; anything below is not.
value exists when offered odds > 1 / true probability
// and its size is
edge = p × offered − 1 → 0.4638 × 2.30 − 1 = +6.67%
Note what this does not say. It does not say the bet will win — at 2.30 it loses about 54% of the time. It says that if you could place this same bet a thousand times, you would end up roughly 6.67% ahead of what you staked. Value betting is a claim about a long series, never about one match.
The probability problem
Everything above depends on p, the true probability. Nobody has it. The entire practical difficulty of value betting is that the formula is trivial and the input is unknowable.
The workable answer is to borrow it from someone with a strong financial incentive to be right. Pinnacle is the standard choice for three specific reasons:
- Low margin. It takes 2–3% where a recreational book takes 6–10%, so there is less distortion to remove.
- High limits. It accepts stakes large enough that a wrong price gets attacked immediately, and corrected.
- It does not restrict winners. This is the important one. A book that bans sharp customers can afford a lazy line; a book that welcomes them cannot.
But Pinnacle's quoted prices still contain its margin, so they are not probabilities either. Add up a typical 1X2 market and you get more than certainty:
| Outcome | Pinnacle | Implied | Fair prob. | Fair odds |
|---|---|---|---|---|
| Home | 2.10 | 47.62% | 46.38% | 2.156 |
| Draw | 3.40 | 29.41% | 28.65% | 3.491 |
| Away | 3.90 | 25.64% | 24.97% | 4.004 |
| Total | 102.67% | 100.00% |
That 2.67% excess is the vig. Removing it — dividing each implied probability by the total — leaves probabilities that sum to exactly 100% and the fair prices beside them. That is the reference line every value bet is measured against. The no-vig odds guide covers the methods in detail, including where proportional removal is the wrong choice.
Finding one, step by step
With a fair line in hand, a value bet is a comparison. Pinnacle's fair price for the home win is 2.156. Bet365 is offering 2.30 on the same outcome.
offered 2.30 at Bet365
edge = 0.4638 × 2.30 − 1 = +6.67%
The four steps, then, are:
- Take the sharp book's complete market for a fixture — every outcome, or the margin cannot be removed.
- De-vig it to get fair probabilities and fair odds (calculator).
- Compare each fair price against every softer book quoting the same outcome (calculator).
- Bet the ones where the soft price is longer, and only those.
By hand this takes a couple of minutes per market. The reason software exists is that edges are perishable — a soft book that is slow to move is precisely the one worth betting into, and it will not stay slow for long — and that the hit rate is low: scanning thousands of fixtures to find a few dozen prices worth taking is not work a person does with a spreadsheet. You can watch the raw inputs on the public odds comparison board.
Compare different books, never a book against itself
De-vigging Bet365 and comparing the result to Bet365 returns an edge of exactly zero, by construction. The value comes entirely from the disagreement between a sharp line and a soft one — and only if you have correctly decided which is which.
How much to stake
A 6% edge on a bet you have over-staked is a worse position than a 3% edge sized properly. The standard answer is the Kelly criterion: stake the fraction of your bankroll that maximises long-run growth, which works out as edge / (odds − 1).
On the example above that is 0.0667 / 1.30 = 5.13% of bankroll. Almost nobody should stake that. Full Kelly is optimal only if p is exactly right, and yours came from de-vigging a bookmaker — it is an estimate with error in it, and Kelly punishes optimistic errors hard. Staking a fraction k of full Kelly keeps k(2 − k) of the theoretical growth rate, so quarter Kelly gives up more than half of it — and is still the right choice, because that growth figure assumes a probability you do not actually have. The Kelly calculator shows both.
Knowing whether it worked
Profit is a terrible short-run measure. At the edges available here, a hundred bets tells you almost nothing — the variance swamps the signal, and both a winning and a losing run are consistent with a method that works and one that does not.
The measure that answers faster is closing line value: did the price you took beat the price the market settled on just before kickoff? The closing line is the market's most informed estimate, and consistently beating it is the recognised evidence that your edge is real. You can see a CLV signal in dozens of bets where profit needs hundreds. The CLV guide covers how to measure and read it.
The limits
Everything above is checkable arithmetic. This part is not, and it is what determines whether value betting is worth your time.
- Accounts get restricted. Soft bookmakers profile customers, and consistently beating the closing line is the profile they look for. Maximum stakes fall — sometimes to a few units — and no software prevents it. This is the real constraint on scale, not finding bets.
- The variance is larger than it feels. A 3% edge over 200 bets loses money a meaningful share of the time. Anyone whose bankroll cannot absorb that will stop before the edge shows up.
- Large edges are usually errors. A +20% edge is far more often a stale price, a market the soft book stopped maintaining, or a fixture matched to the wrong event than a gift. Being suspicious of your best-looking bets is a skill.
- It is work. Placing several hundred small bets across multiple books, recording each one, and doing it while losing for weeks is the actual job. The maths is the easy part.
Frequently asked questions
What is value betting?
Value betting is placing bets where the price offered is longer than the outcome's true probability justifies. If an outcome's fair price is 2.156 and a bookmaker offers 2.30, that bet returns more than it risks on average — regardless of whether it wins. Doing that repeatedly is the whole strategy.
Is value betting actually profitable?
Mathematically yes, and in practice it depends on execution. The edges available at soft bookmakers against a sharp line are typically 1–6%, which is enough to be profitable over hundreds of bets. The constraints are variance in the short run, and bookmakers restricting accounts that consistently take the best price.
How do I find value bets?
Take a sharp bookmaker's price for an outcome, remove the margin to get a fair price, and compare that against what softer books are offering on the same outcome. Anywhere the soft price is longer than the fair price is a value bet. Doing it by hand works; doing it across thousands of fixtures needs software.
How many value bets do I need to place?
Several hundred before results mean much. At a 3% edge, the variance over 100 bets is large enough that losing is entirely normal. This is the single most common reason people conclude value betting does not work — they stopped inside the noise.
Will I get limited for value betting?
Eventually, at soft bookmakers, yes. Books identify customers who consistently beat the closing line and reduce their maximum stakes, sometimes to trivial amounts. Spreading across books, avoiding obviously mechanical staking and betting some ordinary markets lengthens the runway. Pinnacle is the notable exception and does not restrict winners.
What is the difference between value betting and arbitrage?
A value bet has an edge that pays out on average across many bets, but any individual bet can lose. An arbitrage covers every outcome across different books so the profit is locked on that single fixture. Value betting has a higher ceiling and much higher variance; arbitrage is lower and steadier but needs both legs placed.
Do I need Pinnacle to value bet?
You need a sharp reference, and Pinnacle is the standard one because its low margin, high limits and tolerance of winning customers force its line to be accurate. You do not need an account there — the line is public — though having one is useful, since it does not restrict winners.
For research, not betting advice. Positive expected value is an edge across many bets, never a prediction about one. Bet only what you can afford to lose. 18+.