Sports betting explained: the whole knowledge base in one guide
Sports betting is big business: in Great Britain alone, online betting operators kept £2.6 billion of their customers' stakes after paying out winnings in the year to March 2025, £1.3 billion of it on football, according to the Gambling Commission. On top of that come hundreds of markets, bonuses, terms and "sure-fire systems". This guide gives you the essentials in condensed form, so you can quickly see how it all fits together: the conclusions of every article in our knowledge base, with numbers from tests on more than 100,000 matches. Each topic links to the full article when you want to go deeper, and the terms are explained in our betting glossary.
Key takeaways
- 1
A bet pays stake × odds, so at odds of 2.00 a tip has to win 50% of the time just to break even, and the margin means the real chance is usually a little lower.
- 2
At a 5% margin every bet costs about 4.8% of the stake on average, and an acca of five legs leaves £78.35 of every £100 against £95.24 for a single.
- 3
The market prices accurately, as favourites priced at 60–70% won 64.7% of matches, and none of the 89 betting rules of our model passed the test.
- 4
Value starts at any price above the fair odds, but at a 5% margin you have to beat the market's estimate by 5%, and both value bets and surebets end in stake limits sooner or later.
- 5
Stake 1–2% of your bankroll and judge your results only after several hundred bets, because after 100 even a bettor with a real 5% edge is down in about 41% of cases.
A bet is buying a chance at a price the bookmaker sets: odds of 2.00 mean about 50%, and the bookmaker's margin is already built into the price. In Great Britain the Gambling Commission states that customers' winnings are tax-free whatever the amount, so the margin is the whole cost of a bet: at a typical 5%, without an edge over the market, about £95 of every £100 staked comes back on average. Almost everything else follows from those two numbers: why the margin, not luck, decides your result in the long run, why an accumulator costs more than a single, why a bonus has to be worked out before you take it, and why results are judged over hundreds of bets, not over one weekend.
Odds, chances and payouts
- Chance from the odds = 1 / odds. Odds of 2.00 are 50%, odds of 4.00 are 25% and odds of 1.30 about 77%. That is the chance according to the bookmaker, with its margin included.
- Payout = stake × odds. £100 at odds of 2.00 pays £200, stake included. Nothing is deducted, and a push (a refund) gives you back the whole stake.
- Break-even = 1 / odds. At odds of 2.00 a tip has to win 50% of the time for you to break even, at 1.60 62.5% and at 3.40 29.4%. To make a profit, the real chance has to be higher than that.
- Fractional odds, the UK's traditional format, are the same price: divide and add 1, so evens (1/1) is 2.00 and 5/2 is 3.50.
- Odds are an accurate price, and long odds are the most expensive. In 125,000 matches from our database, events at Pinnacle closing odds of around 2.00 came in 48.9% of the time, and those at around 1.50 65.6%. Up to odds of 3.70, between £96.80 and £100.60 of every £100 staked was left; at odds of 8–12 it was £86.60, and above 20 only £61.20.
- The odds change until kick-off, and the closing odds predict results better than earlier ones. Your bet keeps the odds of the moment the bookmaker accepted it.
Read more: Betting odds explained, Betting tracker spreadsheet.
The margin: what every bet costs
The margin is the sum of the implied chances (1 / odds) minus 1, always worked out at one bookmaker on one market. Odds of 1.98 / 3.55 / 3.58 add up to 1.0661, a 6.61% margin. Without it the odds would be "fair", exactly 1 divided by the chances.
| Market | Typical margin |
|---|---|
| 1X2 in the top leagues | 3–6% |
| 1X2 in lower leagues | 6–10% |
| Over/under 2.5, main Asian handicap line | 3–7% |
| Both teams to score | 5–8% |
| Correct score | 15–30% |
| Goalscorer | 20–40% |
| Outright bets | 15–40% |
These are ranges for the world market. At three recreational bookmakers we track, the margins just before kick-off are higher: 6.6–8.4% on 1X2, 7.6–8.5% on over/under 2.5 goals and on both teams to score, and 8.0–8.7% on the main Asian handicap line (October 2026).
The margin is the whole cost of a bet. At 5%, every £100 staked brings back £95.24 on average (100 / 1.05), so each bet costs you about 4.8% of the stake. If the implied chances of a 1X2 market in a popular league add up to more than 1.08, that is a signal to look for another price. Compare the odds on your pick, too: 2.05 instead of 1.98 is a 3.5% bigger payout on every winning bet.
A bookmaker does not need equal money on every result, because it makes its money by pricing accurately. In our database of 104,000 matches, favourites priced at 70–80% won 75.5% of the time, against an average implied chance of 74.2%.
Read more: How bookmakers make money, How to calculate fair odds.
Value bets and surebets: an edge on paper and in practice
- A value bet is a price higher than the real chance justifies. Its expected value: EV = chance × odds − 1. Any price above the fair odds is value: on a fair 2.00 (a 50% chance), odds of 2.10 are worth 0.5 × 2.10 − 1 = +5%. The catch is the margin. A bookmaker with a 5% margin offers about 1.90 on a fair 2.00, so for value you need a real chance above 52.5% where the market sees 50%: you have to beat the market's own estimate by 5%.
- A surebet (arbitrage) is a set of odds at different bookmakers that together cover every result at a profit. The condition: the implied chances (1 / odds) of all the results add up to less than 1. Typical surebets give 1–3%: over 2.5 goals at 2.10 at one bookmaker and under 2.5 at 2.05 at another add up to 0.964, a profit of 3.7% whatever the score.
- A surebet worth more than 10% on a popular match is more often a pricing error than an opportunity.
- Where an ordinary bettor finds an edge: apart from surebets, mostly in promotions. Odds boosts, free bets and cashback are where a price most often climbs above fair, and a surebet calculator then helps you turn a promotion into cash at little cost.
- Value bets and surebets both end in stake limits sooner or later.
Read more: Value betting vs arbitrage betting, What is arbitrage betting, Surebet calculator.
The market knows more than you think
- Closing odds with the margin removed are accurate to within fractions of a percentage point. In 104,000 matches from 37 leagues (2015–2026), favourites priced at 60–70% won 64.7% of their matches (implied 64.4%), and draws came up in 25.8% (implied 25.6%).
- Our goals model predicted better than the league average and as well as an Elo rating, but none of 89 betting rules passed the test on new seasons at Pinnacle's prices. Where the model gave one side at least 10 percentage points more than the market, the bets lost 8–15% of their stakes. A big disagreement with the market is usually a mistake in the model, not an opportunity.
- Pinnacle is widely seen as the best bookmaker in the world. It is a typical "sharp" bookmaker: a 1X2 margin of often 2–3% in the top leagues, high stake limits, and no limits on winning bettors, because their bets quickly improve its prices. That is why the industry uses its closing odds as the benchmark for pricing a match. We use its prices purely as a measure.
- The closing line, the price just before kick-off, is the best available estimate of the chances. The quality of tips is measured by CLV (closing line value) = your odds / closing odds − 1, for example 2.10 / 1.95 − 1 = +7.7%.
- The odds move right up to kick-off, but not evenly. In 7,796 matches at Pinnacle the price barely moved from three days to one day before the match, and the market repriced most on match day (5:00–14:00 UK time) and when the line-ups came out. A day before kick-off almost half of the odds would still move by more than 5%. These moves predict results; only the last 30 minutes are noise. That is why we publish our tips at 12:00 UK time on match day, and for matches kicking off before 15:00, 3 hours before the start.
- An Elo rating turns team strength into chances: roughly 5 percentage points for every 50 points of difference. Attack and defence ratings show how a team wins. They are a reference point for your own judgement, not a ready-made tip, because Elo knows nothing about line-ups, injuries or motivation.
Read more: Why a betting model is not enough, When do odds stop moving?, Club Elo ratings, Attack and defence Elo ratings, Elo ratings to match odds.
Betting markets: how they settle
| Market | How it settles | What to watch out for |
|---|---|---|
| 1X2 | 1, X or 2 after 90 minutes plus stoppage time | the draw is a full third result: it comes up in 26% of matches, and at the three recreational bookmakers we track it carries an 8.5–10% margin |
| Double chance (1X, X2, 12) | two results at once | lower odds, and 1X is not "the home side win" |
| Draw no bet | a draw refunds the stake | you pay for the refund up front, in lower odds: built from 1X2 prices of 2.20 / 3.40 / 3.50, DNB carries a 4.87% margin against 3.44% on 1X2 |
| Asian handicap | a quarter line is two bets of half the stake each | −0.25 is not −0.5: on a draw you lose half the stake, not all of it; one bet in five on the main line ends as a push or half a win or loss |
| European handicap | three results, no refunds, 0:1 = home side −1 | a result exactly on the line is a separate "handicap draw", and the market's margin is higher than on 1X2 |
| Over/under | over 2.5 needs at least 3 goals | a whole line (2.0) can push, and a quarter line (2.25) can return half |
| Team total goals | like over/under, counting one team's goals | an own goal counts for the team it is credited to |
| BTTS (both teams to score) | each team has to score | what counts is who scores, not how many: at 4-0, BTTS yes loses |
| Correct score | one exact score, no refunds | even the most likely score comes up in only a few to just over ten per cent of matches |
| HT/FT (half-time/full-time) | the result at half-time and after 90 minutes, nine options | a reversal, 2/1 or 1/2, lands in 4.6% of matches, so their fair odds are about 38 and 50, and the HT/FT margin is 20–37% |
| Corners | only corners taken in 90 minutes | a corner awarded but not taken before the end of a half does not count |
| Cards | yellow cards only, card points (yellow 1, red 2, two yellows 3) or booking points (10 for a yellow, 25 for a red, 35 at most per player) | one bookmaker's "Number of cards" is card points: over 4.5 lands in 42% of matches on that count, and in 36% on yellow cards only |
| Fouls | over/under on fouls in the match or by one team | some bookmakers settle on Opta data, and a foul after which the referee plays advantage does not count |
| Shots on target and offsides | a shot on target is a goal or a save, and an offside counts once the referee gives it | "shots" is a different market: on Opta's count it includes blocked shots, 22 a match on average |
- A push really is a refund. £100 at 1.95 on Asian handicap 0 gives you back £100 on a draw: no profit and no loss.
- Many markets are the same bet: 1 = AH −0.5, 1X = AH +0.5, DNB 1 = AH 0, AH −1.5 = EH −1. Compare the odds on these "twin" markets and bet where they are higher.
- In football what counts is 90 minutes plus stoppage time. Extra time and penalties count only on markets such as "to qualify" and "to lift the trophy".
Read more: 1X2 betting, Draws in football, Double chance, Draw no bet, Asian handicap, European handicap, Over/under betting, Team total goals, Both teams to score, Correct score betting, HT/FT betting, Corner betting, Card betting, Fouls betting, Shots on target and offsides.
Accumulators, bet builders and player bets
- Accumulator (acca): the odds are the product of the odds, and the chance of winning is the product of the chances. In our data an acca of 5 favourites priced at 70–80% landed 24.1% of the time, and one of 10 favourites 5.9%. The margin multiplies too: with 5 legs at a 5% margin, £78.35 of every £100 is left on average, against £95.24 on a single. The sensible way to play one: 2–3 legs and a small, fixed stake. A k from n system is several accas of k legs each, and on average it returns as much as one of its combinations: more than an acca of the same legs, but less than singles. More in our piece on system bets.
- A void leg on an acca counts at odds of 1.00, and the rest of the slip plays on.
- Bet builder: picks from one match are linked, so the odds are not the product of the prices. Over 2.5 goals and BTTS go together (in our example the product is 2.96 and the fair odds 2.34), and some picks clash. The margin grows with every pick: in the match we checked, 1X2 inside the bet builder carried an 8.1% margin, against 7.0% in the main markets.
- Player bets: the price depends on the player's average and minutes on the pitch. With 1.2 shots on target per 90 minutes and 80 minutes played, the chance of at least one shot on target is 65.6% (fair odds 1.52). A player who does not take the field usually means a void bet; if they come on for even a minute, the bet stands.
Read more: Accumulator bets, Bet builder, Player props.
Virtual sports and free prediction games
- Virtual sports work like a machine: a random number generator draws the result before you see the animation. Form and head-to-head records mean nothing. At an 85% return, 20 matches an hour at £10 a match cost about £30 an hour on average.
- Free prediction tournaments in which only the top ten win a prize reward rare, high-scoring picks rather than "bankers". Before you play, check whether the prize is cash or a bonus that has to be wagered first.
- Skrill Final Score: a free game of predicting exact scores, with prizes that depend on the number of hits. In every match it pays to pick the most likely score, and that is most often 1-1, even when the favourite has more than a 50% chance of winning. What a season of such picks is really worth we worked out on 2,660 matches, and how Skrill picks the matches for its rounds on the first six rounds of the 2026/27 edition.
Read more: Virtual sports betting, Skrill Final Score.
Bonuses: the headline is not the value
- A free bet does not return the stake. A £100 free bet at odds of 3.00 wins £200, not £300. A no deposit bonus, a risk-free bet and cashback usually end up as a free bet. In our data a £100 free bet returns between £24 and £70 on average, depending on the odds you use it at.
- Wagering (rollover): check what it is counted on. A £200 bonus with a 3x requirement means £600 of bets, but the same requirement counted on the bonus and the deposit together means £1,200. The minimum odds often apply to every selection on a slip, and a slip closed with a cash-out often does not count.
- A welcome offer "up to £X" usually comes in several stages, with several deposits and deadlines. Do not deposit more just to unlock the full amount.
- The real value of a bonus = the headline amount − the cost of wagering − the cost of hedging − mistakes.
- Odds boosts give a real edge, but they come with a low maximum stake, so it is usually worth a few pounds.
- Before you take a bonus, answer five questions: what do I get, what do I have to do, by when, what are the restrictions, and when does the money really become mine.
Read more: Free bets, Bonus hunting with a surebet calculator.
How bookmakers work: terms, pricing errors, withdrawals and limits
- Your account: only with a licensed bookmaker. In Great Britain the company behind a betting website or app must hold a Gambling Commission licence, which you can check on the Commission's public register. A bookmaker's track record and reputation tell you more than the size of its bonus. Accounts at 2–3 bookmakers let you bet wherever the odds on your pick are highest: the payout on the same bet can differ by 10%.
- The terms decide every dispute. A bookmaker can refuse a bet, cut a stake, correct the odds after an obvious error and cap the maximum payout. Before your first deposit, check the maximum payout per bet slip and how to make a complaint. How payout caps work is explained in our guide to maximum bet limits.
- A palpable error, or pricing mistake: odds that look "too good" are more often an error than an opportunity. A price 10–15% above the market on a popular match is a red flag, and it will end in a refunded stake or corrected odds.
- Settlement rules differ between bookmakers: around the world a tennis retirement is settled under three different rules, and a postponed match usually has 24–48 hours to be played before the bet is refunded. That matters most with surebets.
- A postponed match almost always means a refund: of 2,280 postponed matches in 37 leagues, 0.1% were played within 48 hours and more than half only after a month. An abandoned match happens roughly once in a thousand matches, and markets already decided before play stopped are settled.
- Verification: in Great Britain an online bookmaker has to verify your age and identity (at least your name, address and date of birth) before you can bet, and under the Gambling Commission's rules it cannot make sending ID a condition of cashing out if it could have asked for it earlier. Most "unpaid" winnings come down to unfinished verification or a deposit made with someone else's card.
- Stake limits: a bookmaker cuts stakes for regularly beating the closing line, not for winning as such. The only sensible answer is to bet in your own name at several licensed bookmakers, never through someone else's account.
Read more: Palpable errors, Bet settlement rules, Postponed or abandoned match, Bookmaker limits.
Calculators: what they work out and what they cannot change
- A surebet calculator splits the stakes so that the worst outcome is as good as possible. Place first the bet that is hardest to get on.
- Dutching splits one stake over several picks so that each pays the same. It changes the shape of the risk, not the value: it pays only when your chance that one of the picks lands is higher than 1 / combined odds. Three correct scores at combined odds of 2.95 need a 33.9% chance. Covering results is not the same as making a profit.
- Hedging: the hedge stake is H = W / K, where W is the payout of your bet and K the odds of the hedge. Certainty costs money: in our accumulator example, about £7.66 of expected value. A cash-out is a hedge priced by the bookmaker, and on the same slip a £90 cash-out offer was £17.30 worse than an equal hedge, so compare the two before you accept.
- Fair odds: take the margin off one low-margin bookmaker's market and only then compare it with the price you want to bet at. This is most useful for judging odds boosts.
- Asian handicap and Elo: the Asian handicap calculator settles any line, quarter lines included, and the Elo calculator turns a rating gap into chances and fair odds.
Read more: Dutching, How to hedge a bet and check a cash-out, How to settle every Asian handicap line, all our calculators.
Bankroll, a betting record and patience
- Stake 1–2% of your bankroll, with 3% as the absolute ceiling. Set a daily loss limit of 5–10% of your bankroll: once you reach it, the day is over, with no "one more bet".
- Keep only working capital in your bookmaker account and withdraw regularly. Keep the rest separately.
- Record every bet. The most important number is yield = profit / turnover × 100%. Record free bets separately, because they inflate the turnover and distort the result.
- Judge yourself over hundreds of bets. After 100 bets a bettor with a real 5% edge is down in about 41% of cases, and one with no edge at fair odds is up in 47%. The margin works on every bet: at a typical 5%, each stake costs 4.8% on average, so 1,000 bets of £100 lose £4,762 on average.
- A result close to zero over a large sample is better than most bettors manage.
Read more: 10 rules to protect your bankroll, Betting tracker spreadsheet.
Thinking traps and "systems"
- The gambler's fallacy: a run is already in the price. Teams after 5 or more defeats in a row won 23.7% of their matches, against an implied chance of 25.3%.
- A "banker" at 1.30 loses roughly one match in four, and needs a 76.9% hit rate just to break even.
- Chasing losses: £60 staked at odds of 2.00 to win back three losing £20 bets leaves you at £0 for the day if it wins and at −£120 if it loses. At a 5% margin the loss is the likelier of the two (52.4%).
- The draw progression (a martingale on draws) looks safe, because a draw comes by the 10th bet in 95.2% of cycles. But after 20 matches without a draw, the stakes at odds of 3.60 add up to £13,395.69, and runs of 20 or more matches without a draw happened 457 times in our data. At fair odds the average season comes out at exactly zero; at the odds bookmakers really pay, it loses at every budget, 7.8% of turnover in our model, which is exactly the margin.
- "Trends" from a small sample: the minutes in which a team conceded its goals do not predict its next matches (a correlation of about zero in 37 leagues).
- Tipster scams: "fixed matches" and "sure tips" for money are always a scam. Someone who really wins would make money betting rather than selling tips at £99. A credible tipster publishes every tip before the match, somewhere they cannot edit it, and has several hundred of them.
- AI tips: at the 2026 World Cup none of the four leading language models predicted results better than the odds, and in a simulation of a Premier League season every model lost money. In practice a chatbot repeats the favourite from the odds, and matching the market is not enough to cover the margin.
Read more: Betting psychology, Draw progression system, Late goals, Tipster scams, AI betting tips.
Responsible gambling
- Betting is for adults only: in Great Britain the minimum age for betting, online and in betting shops, is 18. Bet only with licensed bookmakers.
- Treat it as paid entertainment. Someone who stakes £50 a day turns over £1,500 a month and, at a 6% margin, loses about £85 of it on average.
- Set deposit and loss limits in your bookmaker account. Under the Gambling Commission's technical standards, a lower limit takes effect immediately and a higher one only after a cooling-off period of at least 24 hours. GAMSTOP excludes you, free of charge and in one go, from all online gambling companies licensed in Great Britain for 6 months, 1 year or 5 years. The free BetBlocker app blocks gambling sites and apps on your devices.
- The warning signs are betting to win back losses, raising stakes without a plan and hiding your betting from the people close to you. The PGSI test (9 questions) will show whether it is already a problem.
- Help: the free National Gambling Helpline run by GamCare, 0808 8020 133, open 24 hours a day, every day.
Read more: Betting psychology, with the warning signs and where to get help.
Summary
- Odds are the price of a chance: 1 / odds is the chance according to the bookmaker, margin included.
- Payout = stake × odds, winnings in Great Britain are tax-free, and a push returns the whole stake.
- Without an edge, at a 5% margin, about £95 of every £100 comes back on average.
- Value needs a price above the fair odds and a surebet implied chances adding up to less than 1, but both are rare and short-lived, and both end in stake limits.
- The market prices the chances very accurately, and a big disagreement with it is usually your mistake.
- Every leg on an accumulator multiplies the margin and lowers the chance of winning.
- A bonus is worth what is left after wagering and hedging, not the figure on the banner.
- The terms decide every dispute: check the maximum payout and the settlement rules before you deposit.
- A fixed stake of 1–2% of your bankroll, a daily loss limit and a betting record, with your results judged only after hundreds of bets.
- Runs, progressions, "sure tips" and chasing losses do not work. Betting is paid entertainment for over-18s.
FAQ
Multiply the stake by the decimal odds. £100 at odds of 2.00 pays £200, stake included, so £100 is profit. With fractional odds, multiply the stake by the fraction and add the stake back: £100 at 5/2 wins £250 and returns £350. In Great Britain nothing is deducted, because gambling winnings are tax-free whatever the amount.
Rarely. At a margin of about 5% each bet costs about 4.8% of the stake on average, and the market prices the chances very accurately. In our simulation even a bettor with a real 5% edge was still down in about 41% of cases after 100 bets, and only after 1,000 bets was such a bettor in profit 79.1% of the time. Without an edge, the margin takes £4,762 from 1,000 bets of £100 on average. Real edges come mainly from promotions, such as boosts and free bets, and from regularly beating the closing odds.
With a licensed bookmaker (in Great Britain you can check the licence on the Gambling Commission's public register), age and identity verification, which in Great Britain comes before your first deposit or bet, a budget you can afford to lose, and rules: a stake of 1–2% of your bankroll, a daily loss limit and a record of every bet. Start with simple markets such as 1X2 or over/under, and with singles rather than long accumulators.
In the long run, a single. On an acca the chances and the margins multiply: 5 legs at a 5% margin leave you £78.35 of every £100 on average, and a single £95.24. An acca makes sense as entertainment for a small, fixed amount, ideally with 2–3 legs.
A value bet is a single bet at odds higher than the real chance justifies: it makes money on average, over hundreds of bets and with big swings. A surebet is several bets at different bookmakers that together make a profit whatever the result. At recreational bookmakers' margins both are rare and short-lived, and both lead to stake limits sooner or later.
Only once you have done the sums. A free bet does not return the stake, wagering is sometimes counted on the bonus and the deposit together, and the minimum odds often apply to every selection. A bonus is worth it when a real amount is left after the cost of wagering and hedging, and the conditions do not push you into bets you would not normally place. In our data a £100 free bet returns £24 to £70 on average, depending on the odds.
Most often because you regularly take odds above the closing line, or bet mainly on promotions, surebets or pricing errors. A stake limit is allowed under the bookmaker's terms and is not the same as withholding a withdrawal. Do not open accounts in other people's names: bet in your own name at several licensed bookmakers.