The Bookmaker's Margin: A UK Punter's Maths Guide to Overround, Implied Probability and Value (2026)

The most expensive number in betting is the one nobody puts on the slip.

When a UK bookmaker prices Manchester City to beat Brentford at 1.36, they are not telling you what they think will happen. They are telling you what they need to charge for you to take that view. The difference between those two prices — between what they think and what they charge — is the bookmaker's margin. It is the only number that matters for whether your betting is a recreational expense or a slow leak in your bank account.

Most punters never learn how to read it. Most never have to. The maths is mid-GCSE at worst, and a working understanding of it is the cleanest dividing line between someone who bets for entertainment and someone who bets to win.

What you'll learn

• How any decimal price encodes a probability the bookmaker is willing to bet against — and how to extract it
• How to calculate the overround on any betting market in three lines of arithmetic
• Why a 5-fold accumulator at a "fair" Premier League price is up to 28% worse than the sticker suggests
• Where the bookmaker's margin is widest (horse racing, golf outright) and where it's tightest (Asian handicap, two-way tennis)
• How value betting, arbitrage, laying and Best-Odds-Guaranteed all reduce to one underlying question: is your probability better than theirs?

Table of contents

1. The single most important number on every betting slip
2. How decimal odds encode probability
3. Calculating overround in three lines
4. How overround scales with market structure
5. Average bookmaker margins in the UK, by sport
6. Stripping the margin out — finding true odds
7. Value betting — the only durable edge
8. How overround compounds inside accumulators
9. The exchange — laying overround instead of paying it
10. Arbitrage and Best-Odds-Guaranteed
11. Free bets and promotional boosts
12. Dead heat reduction — overround's quiet cousin
13. FAQ
14. Glossary
15. How this guide is maintained

1. The single most important number on every betting slip

There is one number missing from every betting slip you have ever placed, and it is the one number that tells you whether you have just made a sensible bet or a stupid one.

That number is the bookmaker's margin — known by industry shorthand as the overround, occasionally as the vig, the juice, or the book percentage. They are all the same thing under different names: the amount by which the bookmaker has priced the market above the underlying probabilities the prices imply. It is the bookmaker's edge, expressed in percentage points.

A market priced at 100% has no margin. The bookmaker stands to win nothing and lose nothing in the long run if their prices match reality. No bookmaker has ever offered such a market. Most football match-result markets in the UK sit between 103% and 108%. Most horse racing books sit between 112% and 130%. The Aintree Grand National morning book has been known to clear 150% on bigger-field years.

Every percentage point above 100% is, on average and across many bets, money you are paying the bookmaker to take the bet. Not to lose the bet — to take it. Win or lose, the margin is already priced in.

Why most punters never learn it

There is no malicious conspiracy here. The bookmakers don't hide the overround; you can derive it from any market in under a minute if you know the formula. It just isn't on the slip, isn't on the homepage, and isn't mentioned in the marketing copy. The first time most punters encounter the concept is when they wonder why their accumulators keep ending up half a percent on the wrong side of profit. Some of them work it out. Most of them don't.

A first worked example

Take a real-shaped Premier League match-result market. Manchester City home, Brighton away. The bookmaker prices it:

The market's implied probabilities — what each price says about how likely the bookmaker believes that outcome to be — come from dividing 1 by the decimal odds. We will derive that properly in section 2. For now, take the result on trust:

• City win: 1 ÷ 1.40 = 0.7143, or 71.43%
• Draw: 1 ÷ 4.75 = 0.2105, or 21.05%
• Brighton win: 1 ÷ 8.50 = 0.1176, or 11.76%

Sum: 71.43% + 21.05% + 11.76% = 104.24%.

A perfectly fair market would total 100%. Three outcomes, no overlap, exhaustive. The fact that this market totals 104.24% is the margin: the overround is 4.24%. That is the bookmaker's edge on this market, baked in regardless of what happens at full time.

Convert any fractional, decimal or American price into implied probability with our odds converter →

2. How decimal odds encode probability

The decimal price is the most useful price in the world precisely because it makes the underlying probability trivial to extract. This is the single most useful piece of maths in this entire guide. If you read no further, learn this.

The formula

For any decimal price, the implied probability is:

Implied probability = 1 ÷ decimal odds

That is it. There is no caveat, no asterisk, no version that differs by sport. A 2.50 horse implies a 1 ÷ 2.50 = 0.40 = 40% probability of winning, in the bookmaker's view. A 1.50 favourite implies 66.67%. A 21.00 longshot implies 4.76%.

What it actually means

The bookmaker's implied probability is not their true belief about how likely the outcome is. It is what they need to charge to take your bet given that they also have to pay other punters on the other side, cover their operational costs, return a margin to their shareholders, and stand a reasonable chance of breaking even on the market as a whole. The implied probability is the price-derived figure; the true probability is whatever actually happens across many repetitions of the same event under similar conditions.

The art and the maths of profitable betting both come down to the gap between those two numbers. Section 7 covers it properly.

A worked example with a real-shaped horse

Take a typical second-favourite in a Cheltenham Festival handicap chase. The horse is priced at 9/2 fractional, 5.50 decimal, +450 American. All three are saying the same thing.

• Fractional 9/2 → decimal (9 ÷ 2) + 1 = 5.50
• Decimal 5.50 → implied probability = 1 ÷ 5.50 = 0.1818 = 18.18%
• American +450 → implied probability = 100 ÷ (450 + 100) = 100 ÷ 550 = 18.18%

All three prices encode the same probability. The market is saying: in our pricing, this horse wins this race 18.18% of the time. If you think it wins more often than that, you have value. If you think it wins less often, you do not.

When the implied probability is wrong (and how you'd know)

The implied probability is wrong most of the time, in the strict sense that the actual outcome diverges from it on any given race. That is not a flaw — probabilities aren't certainties. What matters is whether the implied probability is wrong on average.

You cannot know for any single race. You can know across hundreds. If you bet only horses priced 9/2 to 5/1, and across 200 such bets your strike rate is 22%, the market was systematically too low on those prices. You would, on average, have value. We will show how to convert that into a strike-rate vs implied-probability formula in section 7.

Test fractional/decimal/American conversions on any odds in seconds →

3. Calculating overround in three lines

If you remember nothing else about the maths of betting, remember this:

Overround = (Σ (1 ÷ decimal odds for every outcome)) × 100 − 100

In English: add up 1 ÷ decimal odds for every outcome in the market, multiply by 100, subtract 100. What you have left is the bookmaker's margin in percentage points.

A worked example — full Premier League match book

Take a more competitive fixture: Arsenal at home to Liverpool. A typical book at a UK retail bookmaker on the morning of the match:

Overround: 102.26% − 100% = 2.26%.

That is among the tightest match-result books you will see on UK football. Premier League fixtures between the big six tend to be priced in the 2–4% range because the bookmakers expect heavy two-way action and cannot tilt the market without leaving themselves exposed. Lower-league fixtures, particularly mid-week League Two matches, sit at 6–9% because the bookmakers expect less volume and need a wider cushion.

Why a fair market would total 100% — and never does

Imagine a perfectly fair coin. The probability of heads is 50%; the probability of tails is 50%. The two probabilities sum to 100% because heads and tails are the only outcomes and they cannot occur simultaneously. A fair betting market on the coin would price each outcome at decimal 2.00. The book would total 1 ÷ 2.00 + 1 ÷ 2.00 = 100%. The bookmaker would, in the long run, win on half the coins and pay out on the other half, and end up flat.

A bookmaker who runs flat goes bust. So the bookmaker prices each side of the coin at 1.95 instead of 2.00. The book now totals 1 ÷ 1.95 + 1 ÷ 1.95 = 102.56%. They have set the market at a 2.56% overround. Win or lose the next coin flip, across enough coin flips they keep that 2.56% of every pound staked. That is not a fix; that is how the entire industry stays solvent. Margin is the price of having a market to bet into.

4. How overround scales with market structure

A two-way market is the cleanest case. The bookmaker only has to manage two outcomes and can run thin margins because the action self-balances. A 14-runner handicap race is a different proposition: 14 outcomes, much harder to balance, larger price uncertainty on each horse, and a non-trivial chance of any of them winning. The bookmaker compensates by widening the book.

This is structural, not malicious. Wider outcome counts mean more exposure, more variance, less self-balancing action. Margins have to widen to compensate.

Two-way markets

A two-way market — tennis match-winner, Asian handicap, over/under goals — typically runs at 2–4% overround in the UK retail market. Pinnacle, the closest thing to an industry baseline, will run two-way markets at 1.5–2.5%. Most UK retail books sit higher because they do not pursue the same line-shopping audience.

Three-way markets

Football match-result is the canonical three-way: home, draw, away. The draw is the problem outcome — the public underbets it almost universally, so the bookmaker can price it generously without taking damage. A typical 3-way book runs 4–7% overround.

Outright markets with 12+ outcomes

Horse racing fields, tournament outright golf, election markets. Margins balloon. A typical 14-runner handicap race sits at 18–25% overround at the off, often higher in the morning book.

A worked Grand National book

The 2025 Grand National field had 34 runners on the day. A representative morning book had prices ranging from 6/1 on the favourite to 200/1 on the rank outsider. Adding up the implied probabilities of all 34 horses gave a book that totalled roughly 142% [VERIFY: 2025 Grand National morning book overround, multiple sources]. By the off, after late money tightened the favourite and the rags drifted, the book typically settles around 115–120%. That is still a 15–20% margin — the equivalent of buying a fairly priced market and paying a 15-20% commission to enter it.

You can see why ante-post horse racing is a punter's graveyard. The morning prices are priced for a bookmaker who has not yet seen the money. By the time the public has moved the market, the value, if it existed at all, is largely gone.

Build a multi-leg horse racing bet and see the compounded margin in real time →

5. Average bookmaker margins in the UK, by sport

The numbers vary by operator, by sport, by market depth and by time of day. The ranges below are what a punter using mainstream UK retail books (William Hill, Ladbrokes, Bet365, Sky Bet, Paddy Power) will typically encounter as of early 2026. Pinnacle and the betting exchanges sit considerably tighter; small operators on niche markets sit considerably wider.

[VERIFY: representative ranges across 5 UK operators, sampled across a single 2025 weekend — recompile quarterly.]

Why golf and racing books run so wide

Two reasons, both structural:

First, field size. Sixteen runners or 156 golfers is a lot of outcomes to price. Pricing the favourites tightly while pricing the back-markers generously balloons the book percentage even if every individual price is "reasonable" in isolation.

Second, public concentration on the front of the market. The serious money goes on the top 10 names. Bookmakers can price the bottom 100 in a golf field at whatever they like because virtually no money will land there. The aggregate book ends up wildly inflated; nobody actually bets the inflated section.

This matters less than it looks in the win market and more than it looks in the each-way market — the place-fraction returns are computed against the same prices, so a wide outright book translates directly to a wide each-way book.

6. Stripping the margin out — finding true odds

If the bookmaker's odds carry a 5% margin, what are the underlying "fair" prices the bookmaker would offer if they ran flat? You can recover them. There are two methods. The first is mechanical and approximate. The second is more accurate but requires more arithmetic.

Method one: pro-rata strip-out

Take each outcome's implied probability and divide it by the book total. The result is the fair probability, as the bookmaker sees it. Fair odds are then 1 ÷ fair probability.

Take the Arsenal vs Liverpool book from section 3:

Every fair price is slightly longer than the book price — the difference is the bookmaker's cut, redistributed evenly. The pro-rata method assumes the margin is applied evenly across every outcome.

Method two: the Shin / power-margin model

In reality, bookmakers do not apply margin evenly. They load the favourites harder than the longshots, because the favourites attract the volume. A more accurate strip-out — used by quants at exchanges and large bookmakers — applies a power-based correction where each implied probability is reweighted by an exponent slightly above 1.

The maths is heavier than this guide will carry; an approximate version is:

Fair prob = (implied prob ^ k) ÷ Σ (implied prob ^ k) where k is chosen so the resulting Σ = 1

For most retail-bookmaker books, k lands between 1.05 and 1.15. The result is fair prices that look longer on the favourites (closer to true) and shorter on the longshots (where the public overbets) than the pro-rata method gives.

What neither method gives you

The bookmaker's actual internal probability. That is the model output the trading team works from, and it is not recoverable from the public prices because the public prices include a manual margin overlay, a public-volume adjustment, and a competitor-watching adjustment. The fair-price estimates above are useful, but they describe the market price stripped of overround, not the bookmaker's pricing model.

In practice that is fine. You are not trying to copy the bookmaker's view. You are trying to find prices where your own view is meaningfully different from theirs.

7. Value betting — the only durable edge

A "good price" and a "value price" are not the same thing. A 33/1 horse can be terrible value; an evens favourite can be a screaming bet. The price is irrelevant in isolation. The only question that matters is whether the price is too long for the underlying probability.

Definition

A value bet is one where:

Your estimated true probability × decimal odds offered > 1

If a horse you believe is a 25% chance is priced at 5.00 (implied probability 20%), then 0.25 × 5.00 = 1.25. That is a +25% edge per pound staked, before commission, in expectation. Across enough such bets you would make money. On any individual bet, anything can happen.

If the same horse is priced at 3.50 (implied probability 28.57%), then 0.25 × 3.50 = 0.875. That is a 12.5% expected loss per pound staked. A horse you think is a 25% chance is a terrible bet at 3.50, even though it's a "good price for a 25% chance" only in the most casual sense.

A worked example

A horse at the Cheltenham Festival Champion Hurdle, second favourite, priced at 5/1 with a UK retail book (decimal 6.00, implied probability 16.67%).

You have done the form. The horse won its last race convincingly. The favourite is doubtful in soft ground, which is forecast. The next-best in the market is unproven over the trip. Based on this assessment, you put the horse's true probability at 22% — a 1-in-4.5 chance.

Your probability × decimal odds = 0.22 × 6.00 = 1.32

You have a 32% edge in expectation. The bet is value. You should take it at this price.

If the same horse drifts to 7/1 (8.00 decimal) by the off, your edge increases: 0.22 × 8.00 = 1.76. If it shortens to 4/1 (5.00 decimal), your edge collapses: 0.22 × 5.00 = 1.10. If it shortens further to 3/1 (4.00 decimal), the bet becomes negative: 0.22 × 4.00 = 0.88. Same horse, same forecast, very different bets at different prices.

The Kelly Criterion in one paragraph

Once you know your edge, the question of how much to stake has a mathematically correct answer called the Kelly Criterion. The full Kelly stake is (probability × odds − 1) ÷ (odds − 1) of your bankroll. For the Champion Hurdle horse above at 6.00 with a 22% true probability, that's (0.22 × 6 − 1) ÷ (6 − 1) = 0.32 ÷ 5 = 6.4% of bankroll. Almost no professional bettor stakes full Kelly because the variance is brutal — most use quarter-Kelly or half-Kelly, accepting slower growth for considerably less risk of ruin. The mathematics are uncontested. The discipline of using them is rare.

Why "good odds" is the wrong question

A 20/1 horse can be wildly overpriced if its true probability is 6%. A 6/4 favourite can be value if its true probability is 50%. The price tells you nothing on its own. Your view of the probability tells you everything. The market price is a hypothesis you are testing your view against — not a quality marker for the horse, the team, or the bet.

This is the single hardest mental shift in betting. The casual punter wants to know "is this a good bet at these odds?" The disciplined punter asks "given my view of the probability, do these odds represent value?" Those are not the same question. The first cannot be answered. The second always can.

Test value bets across markets — convert any price to implied probability and check it against your own forecast →

8. How overround compounds inside accumulators

A 5% margin sounds small. Across a single bet on a single market, it is. Inside an accumulator of five legs, each carrying that same 5% margin, the compound effect becomes punishing.

The maths

Each leg's price is the fair price multiplied by 1 ÷ (1 + per-leg margin). Across N legs, the combined acca price is the product of leg prices. The bookmaker's combined edge on the acca is approximately (1 + per-leg margin) ^ N − 1.

For a 5-leg acca with 5% per-leg margin: 1.05 ^ 5 − 1 = 0.276 = 27.6% combined margin.

For a 5-leg acca with 7% per-leg margin (typical UK retail Premier League): 1.07 ^ 5 − 1 = 0.403 = 40.3%.

For a 10-leg acca with 5% per-leg margin: 1.05 ^ 10 − 1 = 0.629 = 62.9%.

A worked example

You pick five Premier League home favourites in a 5-fold accumulator. Each leg is priced fairly at 2.00 decimal but quoted by the bookmaker at 1.91 to reflect a 5% margin per leg.

• Fair acca price: 2.00 ^ 5 = 32.00
• Quoted acca price: 1.91 ^ 5 = 25.41
• Your £20 stake at the quoted price: £20 × 25.41 = £508.20 if it wins, £100 below the £640 a fair acca would have returned

The acca looks at face value to be priced normally. The compound margin has stripped 20.6% of the expected return on a winning bet without you ever seeing the number on the slip.

Why a 10-fold accumulator is a tax disguised as a bet

The same maths run on a 10-fold acca at 7% per-leg margin: 1.07 ^ 10 − 1 = 96.7%. The bookmaker is keeping nearly an entire stake's worth of expected value across the bet's life. Boost promotions ("get 30% extra on 4+ folds") undo some of this, but rarely all of it. Once you are out past 6 legs the maths get genuinely brutal even before you ask how often a 10-fold actually lands.

The 10-fold is the bookmaker's favourite bet. It has the highest combined margin, the worst payout-to-stake-spent ratio across many bets, and the cheapest marketing pitch ("£1 to win £20,000"). It is the slot machine of the betting menu.

Build any accumulator from 2 to 16 legs and see the implied combined margin alongside the payout →

9. The exchange — laying overround instead of paying it

A betting exchange — Betfair, Smarkets, Matchbook — is not a bookmaker. It is a marketplace where punters bet against each other. Half are "backers" (taking the same side a traditional bookmaker would offer) and half are "layers" (offering the bet, taking the opposite side). The exchange charges a commission on net winnings — typically 2–5% — and balances the book by matching the two sides at whatever price the market clears.

The structural consequence is that exchange prices have near-zero overround. The market clears at the price where supply meets demand. The exchange doesn't need to load a margin to make money — they take commission on what's already there.

What it means to lay a bet

When you back a horse at 6.0 with £20, you risk £20 to win £100 if the horse wins. When you lay a horse at 6.0 with £20 backer money, the maths inverts. Your liability is what you stand to lose:

Liability = (decimal odds − 1) × backer stake = (6.0 − 1) × £20 = £100

If the horse wins, you pay out £100 to the backer. If the horse loses (or doesn't place, or whatever the conditional is), the backer's £20 stake is yours — minus exchange commission.

Lay betting is how matched betting works. Lay betting is how a serious punter neutralises a bonus or a free bet exposure. Lay betting is how the trading desks at the bookmakers themselves manage their internal positions on the exchange — laying off liability on a horse where they have taken too much money.

Why exchange prices are usually slightly longer than bookmaker prices

The bookmaker's overround is the bookmaker's margin. The exchange's overround is essentially zero. So a 6.0 horse on the exchange typically represents the bookmaker's fair price on the same horse, before the bookmaker's 5–10% retail margin is added. This is why the disciplined punter will often check the Betfair price first and use it as the true-probability anchor for whatever they are about to bet at a retail book. It is, in effect, the wisdom-of-crowds price after stripping retail margin out.

Calculate lay liability and exposure across any backer stake / odds combination →

10. Arbitrage and Best-Odds-Guaranteed

When two bookmakers price the same market and one of them prices the favourite tightly while the other prices the underdog tightly, the combined implied probabilities of the two prices can total less than 100%. That is a negative overround — an arbitrage opportunity.

The mechanics

A tennis match. Book A prices player X to win at 2.10 (implied probability 47.62%). Book B prices player Y to win at 2.10 (implied probability 47.62%). Combined book: 95.24%. Negative overround of 4.76%. Backing X at A and Y at B in the right stake ratio means one of them must win, returning a guaranteed profit.

The stake calculation

To return the same amount regardless of which player wins:

Stake A = (Total bankroll × (1 / odds A)) ÷ combined implied probability = (£100 × 0.4762) ÷ 0.9524 = £50

Stake the same £50 on B. Whichever player wins, you return £50 × 2.10 = £105. You staked £100. Profit: £5, or 5% of stake, locked in.

Why arbs disappear in seconds

Three reasons. First, the arb-spotting software the professional arbers run sees the same opportunity within milliseconds. Second, the bookmaker's own trading systems also see it, and adjust. Third, the bookmakers actively discourage arbing: stake limits get lowered ("gubbing"), accounts get restricted, and the maths become quickly impossible to act on at meaningful stake sizes.

In practice, arbitrage as a strategy in the UK retail market is dead at any serious volume. Arbs are still useful as a sanity check on prices — if you see a negative overround between Book A and Book B, one of those two prices is wrong and you can usually work out which by checking the exchange.

Best-Odds-Guaranteed: the bookmaker's softest concession

Best-Odds-Guaranteed is the closest thing to free value the UK bookmakers give a punter. The deal: you take an early price on a horse race; if the horse drifts and goes off at a longer Starting Price, the bookmaker pays you out at the SP instead.

It only fires on the upside. If your horse shortens, you keep the longer early price you took. Asymmetric. Free. Worth using on every horse race you bet.

Worked example: a 5/1 horse drifting to 7/1 at SP

You back a horse at 5/1 in the morning (decimal 6.0). The horse drifts during the day and goes off at 7/1 SP (decimal 8.0). BOG fires.

• Without BOG: £10 stake × 6.0 = £60 returned, £50 profit
• With BOG: £10 stake × 8.0 = £80 returned, £70 profit
• Difference: £20 of upside on a £10 stake, with zero downside

Across a season betting only horses 4/1 or shorter (which drift roughly 30–35% of the time [VERIFY: bookmaker industry data on drift rate by price range]), BOG returns somewhere between 1.5% and 2.5% in additional expected value on every staked pound, just from picking books that offer it.

The mathematical worth of BOG is highest on horses 4/1 to 7/2. Below 6/4 the drift is rare. Above 14/1 the price tends to either stay or shorten in big races. The favourite-band is where BOG quietly pays.

11. Free bets and promotional boosts

A "free bet" from a UK bookmaker is almost never literally free. It is a stake-not-returnable token, typically worth £10–£30, given out for signing up or as a periodic promotion. Understanding what it is actually worth requires one more piece of arithmetic.

Stake-not-returnable maths

A £20 stake-not-returnable free bet at decimal 3.00:

• If the bet wins: returns the winnings only, not the stake. £20 × 3.00 = £60, minus the £20 stake (which is "consumed" by the free bet token), = £40 net profit.
• If the bet loses: £0. The token is spent.

Compared to a £20 cash bet at the same odds:

• Win: £60 returned (£40 profit and £20 stake)
• Lose: −£20

The free bet retains £40 / £60 = 66.7% of face value at decimal 3.00. The retention rate climbs with longer odds: at 10.0, retention is £180 / £200 = 90%. At 21.0, 95%.

How to extract maximum value

Always stake free bets on longer-priced selections. The conventional retail-bookmaker advice is "stake your free bet on a horse you'd bet anyway at 4/1+." That advice is mathematically correct. At 4/1 (decimal 5.0), retention is £80 / £100 = 80%. At evens, retention is £20 / £40 = 50% — half the face value evaporates.

If you intend to use a free bet for matched betting, you lay off the back-bet on the exchange to lock in the difference between the back retention and the lay liability. Done correctly, a £20 free bet at decimal 6.0 can be converted into roughly £14–£15 of guaranteed real money, depending on exchange commission.

Acca boosts

A 30% acca boost on a 5+ leg acca, advertised by virtually every UK retail bookmaker. The mechanics differ subtly book-to-book; the most common version is "30% extra on your winnings" rather than "30% on total return". The distinction matters.

A £20 stake on a 5-fold acca at quoted odds 10.00:

• Without boost: £20 × 10 = £200 returned, £180 profit
• "30% on winnings": Profit × 1.30 = £180 × 1.30 = £234 profit, £254 returned
• "30% on total return": £200 × 1.30 = £260 returned, £240 profit

You will sometimes see the boost framed as a multiplier on odds: 30% × 10.00 = 13.00 effective price. That is the same as "30% on total return" under the hood. Read the T&Cs.

Whether the boost beats the underlying compound margin

From section 8, a 5-leg acca with 7% per-leg margin loses 40.3% of expected return to compounded margin. A 30% boost recovers about three-quarters of that. The bet is, in expectation, still negative — but materially less negative than without the boost. Acca boost promotions are not free money; they are partial refund of a tax the bookmaker imposed in the first place. Treat them as the floor, not the ceiling, of fair value on accumulators.

Stake any accumulator with or without a promotional boost and see the effective return →

12. Dead heat reduction — overround's quiet cousin

Dead heat rules sit alongside overround in the small category of betting maths that quietly empty a punter's pocket without ever being mentioned on the slip. They affect any market with multiple paid positions: top 5 finishers, top 10, top 20, top finisher from a country, group winners.

The mechanic

When more players tie at the boundary of a finishing market than there are paid places remaining, the stake gets divided. Specifically:

Adjusted stake = original stake × (paid places remaining ÷ players tied)

The reduced stake is settled at the original odds; the rest of the stake is lost.

A worked example — a top-5 golf market

You back a contender at 12/1 (decimal 13.00) in a Top 5 finish market at a major. You stake £20. Four players tie for fifth. The top 4 finishers settled cleanly. One Top 5 paid place remains. Four players share that one place.

• Adjusted stake = £20 × (1 ÷ 4) = £5
• Settled return = £5 × 13.00 = £65
• Unused stake = £15 (lost)
• Net return on £20 wagered = £65 (= £45 profit)

Versus an outright top-5 finisher (no tie): £20 × 13.00 = £260 returned (£240 profit). Dead heat reduction has cut your return by 75% on what looked like a winning bet.

Where dead heat hits hardest

Golf majors, where 144-156 player fields produce ties at every boundary. Horse racing place markets, particularly handicap chases with photo-finish weather. Tournament outrights in tennis where multiple ties at quarterfinals are common.

Every UK bookmaker handles dead heats slightly differently. Most use the formula above. A small minority — particularly on golf and racing — use "dead heat reduction factor" tables that can be fractionally different. Always check the book's small print.

Run any dead heat scenario — top 5, top 10, top 20 — through our dead heat calculator →

FAQ

Glossary

Accumulator (acca) — A single bet combining 2 or more selections where all must win for the bet to pay out. Returns multiply across legs; so does the compounded margin.

Asian handicap — A two-way football market that eliminates the draw by handicapping one team. Tightest overround in mainstream UK football betting, typically 2–3.5%.

Back — To bet on an outcome happening. The traditional retail-bookmaker side.

Best-Odds-Guaranteed (BOG) — A bookmaker promotion on horse racing where, if you back a horse at a longer price than the eventual Starting Price, you are paid out at the longer price; if SP is longer, you are paid at SP. Asymmetric, free.

Bookmaker margin — See overround.

Book percentage — See overround.

Decimal odds — A price format where the displayed number is the total return (stake + winnings) per £1 staked. A 2.50 horse returns £2.50 per £1, including stake.

Dead heat — Multiple competitors finishing in the same paid position. Triggers a stake reduction in any finishing market where the tie spans the boundary.

Fair odds — The price an outcome would carry if there were no margin. Often used loosely to mean the exchange price minus commission.

Fractional odds — A price format quoted as "winnings to stake" (e.g. 5/1 = £5 winnings on £1 stake, plus stake returned = £6 total). UK traditional format.

Gubbing — Industry slang for a bookmaker restricting or closing an account showing evidence of value-finding behaviour. Not illegal; entirely within the operator's licence terms.

Implied probability — The probability a price encodes. 1 ÷ decimal odds, expressed as a percentage.

Juice / vig — American terms for overround. Same concept.

Kelly Criterion — A mathematical formula for optimal bet sizing given a known edge. (probability × odds − 1) ÷ (odds − 1) gives the fraction of bankroll to stake. Almost no professional uses full Kelly because of variance; quarter-Kelly and half-Kelly are standard.

Lay — The opposite of back. To bet that an outcome will not happen. Available on exchanges; conceptually the bookmaker's role.

Liability — The amount a layer stands to lose if their lay bet wins (i.e., the backed outcome occurs). (odds − 1) × backer stake.

Negative overround — A book that totals less than 100%. Indicates an arbitrage opportunity. Rare on a single bookmaker; possible across multiple bookmakers' prices on the same market.

Overround — The bookmaker's margin on a market, calculated as the sum of all implied probabilities minus 100%. Always positive at a single bookmaker.

Place market — A market that pays on multiple finishing positions (top 5, top 10, etc.). Dead heat rules apply at the paid-place boundaries.

Shin model — A more accurate margin-stripping technique that accounts for favourite/longshot bias in implied probabilities. Used by professional traders and exchanges.

Starting Price (SP) — The official price of a horse at the moment a race begins, set by an industry panel. The reference price for BOG settlement.

Value bet — A bet where the punter's estimated true probability of the outcome is higher than the bookmaker's implied probability. Mathematically, the only durable source of edge.

How this guide is maintained

This page is reviewed quarterly. The maths do not change; the margin ranges in section 5, the worked examples drawn from real fixtures, and the BOG / drift-rate references in section 10 are revisited every three months to reflect current UK retail-bookmaker behaviour. Any number flagged [VERIFY:] in the body indicates a figure we have approximated from public sources and not independently audited — these are not estimates we would rely on for any individual bet without checking the underlying market.

This guide was written by the BetCalc365 editorial team. Corrections, market-by-market overround data points, or worked-example corrections should go to Hello@crofthughdigital.co.uk with the section number in the subject line. We update fastest on factual errors.

Last reviewed: May 2026. Next review: August 2026. The principle behind this page — that the cleanest dividing line in betting is between punters who can read a margin and punters who can't — was true when bookmaking was a chalk-and-bag trade in 1890, and is true now. The arithmetic has not aged, and is unlikely to.

Convert any odds to implied probability and check it against your own forecast — the start of every value-driven bet →