What BTTS Is and What It Is Not
Both teams to score is the cleanest two-outcome football market on the board. Yes, both sides score. No, at least one side does not. No draws, no half-result settlement, no three-way complexity. On a well-run book the two prices sit close enough together that you can convert one into the other and back.
The reason it is dismissed as a coin flip is that the superficial comparison is a coin flip. Priced at 1.80, yes needs 55.6 per cent, which feels uncomfortably close to a flip. But a flip only pays at 2.00, and a market that pays 2.00 on a genuinely even chance is one of the few generous things in betting. The 1.80 price exists because the book knows which side is more likely, and that knowledge is built from a very small number of inputs.
More importantly, BTTS is not the over 1.5 market, and that difference is the source of most losing BTTS bets. Over 1.5 needs three goals; 2-0 satisfies it and fails BTTS. Yes-plus-Over-1.5 needs three goals with one on each side, so 1-1, 1-2 and 2-1. Treat these as three different markets with three different implied probabilities and you stop conflating them, which is where the edge lives.
Settlement Rules Worth Reading Twice
Most of the complications in BTTS are boring and all of them are expensive. Regulation time plus stoppage time is the standard. A goal in the first minute of second-half stoppage time counts. Own goals count towards the scoring team, so a scrappy own goal that decides a BTTS bet is a BTTS loss for the side that was defending one. Penalty shootouts never count.
Abandoned matches are void and your stake returns. Abandoned after both teams have scored is usually settled as a win, but the practice varies. Delayed matches that kick off after a long delay may be voided at the book's discretion.
Now the one that catches people: 90-minute BTTS versus 95-minute BTTS. The 95-minute market includes first-half stoppage time but excludes second-half stoppage time, which sounds absurd and is a real product. If a book lists both, take the 90-minute one unless the 95 price is clearly better.
Live BTTS is a separate product with its own trap. Once the first goal goes in, the live BTTS no price collapses to about 1.05, and the yes price becomes a live over 0.5 for the trailing side wearing a BTTS label. Price it as an over 0.5, not as an original BTTS decision.
What Each Market Implies About the Same Game
All of these can be offered on the same fixture. They are correlated, they are not equal, and conflating them is the most common BTTS error there is.
| Market | What settles it | Typical odds vs BTTS Yes at 1.80 |
|---|---|---|
| BTTS Yes | At least one goal for each side in regulation | 1.80 |
| Over 1.5 goals | Three or more goals total, any distribution | 1.20 to 1.40, much more likely |
| BTTS Yes + Over 1.5 | Three or more goals with one for each side | 3.20 to 3.80 |
| BTTS No | Either side finishes on zero | Around 2.00 when Yes is 1.80 |
| Over 0.5 first half, both | First half contains at least one goal | 3.00 to 4.00, a much longer shot |
Why BTTS Is a Direct Function of Expected Goals
Here is the model, and it is short enough to hold in your head.
Suppose team A is attacking the defence of team B. Team A's expected goals is 1.62. The probability that team A scores at least once is 1 minus e raised to the power of minus 1.62, which is 0.80.
Now the same for team B. Suppose their expected goals is 1.15. The probability of B scoring is 0.68.
Assuming the two scoring events are roughly independent, the probability of both teams scoring is 0.80 multiplied by 0.68 = 0.546. Just over 54.6 per cent. At a 1.80 price, break-even is 55.6 per cent, so a fair price for that fixture is about 1.83. The book offers 1.80. That is a 1.7 per cent margin on a market that takes twenty seconds to evaluate.
You can do that in your head for any fixture once you know the two xG numbers, and that is why BTTS is tradeable: it is one of the few football markets where the price is a nearly pure function of two published inputs, with almost no narrative adjustment on top. The book does not move BTTS because a manager looked tired in a press conference. It moves BTTS because a midfielder pulled up in training. Our expected goals guide covers how to source and adjust those numbers.
The Correlation With the Over Line Is the Whole Trick
BTTS and the over line share both inputs. That shared dependence is what makes pairing them interesting, and it cuts both ways.
When a total line is high, say over 2.75, it is usually because one team has a strong attack. But a high total is not the same as both teams being dangerous. Take a 2.75 total with an xG split of 2.20 and 0.55: a dominant favourite against a parked bus. BTTS yes there is 0.89 multiplied by 0.42 = 0.37. The book knows this, and a 3.00 total in that fixture comes with BTTS yes at 2.70. Taking it at 2.00 is the most common losing BTTS bet there is.
The mirror image is a 2.25 total with a split of 1.35 and 0.90. BTTS yes is 0.74 multiplied by 0.59 = 0.44, and the book will happily push that to 1.55 because the low total makes it look like a scrappy game someone must nick. Often the total is low because both attacks are poor, which makes BTTS less likely, not more. The rule that follows: judge BTTS off the two xG numbers, never off the total line. If your reasoning began with "the over is 2.5 so BTTS is likely", stop and recalculate.
Nine Numbers That Drive the BTTS Price
Everything the book uses to price BTTS is on this list, and most of it is public.
- xG for, home. The biggest single input. Below 0.9 and BTTS yes gets expensive. Above 2.0 it is near certain unless the opponent is a fortress.
- xG against, home. The mirror. A defence conceding 1.8 xG per match will not keep a clean sheet, which inflates the other side of the calculation.
- Post-shot xG. Expected goals after shot quality. The gap between xG and post-shot xG is the defence's actual goalkeeping and blocking, and it is the best correction available.
- Shots from the box per match. The fallback when xG is unavailable. Twelve or more usually means 1.3 xG or better.
- Big chances created and missed. A side missing 0.35 xG of big chances a game is priced as if it will convert soon. Regress it.
- Clean sheet rate in the last ten. Short-term, noisy, and the input most bettors use badly. Regress it before you use it.
- Fixture congestion. A side in three games in eight days scores about 12 per cent less. The effect is real, small and free, because the fixture list is public.
- Home and away splits. Many sides score at home and not away. A season-wide xG for the away leg throws away the most reliable split in the data.
- Promoted-team baseline. A promoted side is priced closer to a top-flight average than it should be for the first eight matches. A structural error, not a one-off.
Why BTTS Is More Predictable Than the 1X2
The match result is a three-way outcome depending on the interaction between two attacks and two defences across ninety minutes. BTTS depends on a strictly smaller question: did each side score once or more, regardless of how many the other got. Fewer degrees of freedom, and fewer correlations to model.
Consider the variance. A 1X2 price of 2.20 implies roughly 45 per cent. Getting that right from two xG numbers needs a full bivariate Poisson or equivalent, because the two scorelines are dependent. Getting BTTS right needs two independent Poisson tails. The second task is materially easier, and it is why books keep the margin on the 1X2 tighter than their own modelling error.
A practical test. If your BTTS forecast lands within three points of the market, you have found nothing. If it lands six or more clear, either you have found something or your inputs are wrong. Check against a second route before staking: multiply each team's home scoring rate by the opponent's home conceding rate. If the two routes disagree by more than five points, do not bet. The wider framework is in our football betting strategies page.
League Rates and Where the Margin Is Widest
A national BTTS hit rate is a trap, because the variation between competitions is wider than most bettors assume. Leagues with two high-tempo attacking sides and defences that leak produce rates well above 60 per cent. Leagues where the leading side plays deep and narrow produce rates below 50. Knockout cups run materially below their league averages, because the mismatch is intentional.
So the procedure is not to decide "BTTS yes" in the abstract. It is to find the fixture's two xG numbers, compute the Poisson probability, and compare to the price. League averages are a sanity check on your inputs, not the input itself. Use them to spot when a book has priced a Serie A fixture at Serie A numbers and a Championship fixture at Premier League numbers, which happens more often than you would expect when a book copies a competitor rather than modelling. Our margin guide covers pricing the gap, which is the entire profit here.
Not all BTTS markets carry the same book. Pre-match main markets on major leagues from the sharper books run 103 to 108 per cent, and that is where the professional money sits. Smaller leagues and lower divisions run 108 to 118, where the errors are larger but so is your own uncertainty. Live BTTS runs 110 to 125 and grows the longer the match runs: after the first goal the no side collapses to 1.02 and the yes side becomes a single-team over 0.5, which is no longer a BTTS bet at all. Outright season markets run 130 per cent or worse. Bet them never. The value in BTTS is in mid-table European leagues where the xG data is public and the pricing is loose.
Two Real BTTS Price Mistakes and the Fix
Both are the same error in different clothes: pricing a fixture off the total line instead of off the two attack strengths.
| Mispricing | What the bettor assumes | What the model says | The fix |
|---|---|---|---|
| Over 2.5, BTTS yes at 1.55 | Three goals means both scored | A 2-0 satisfies over and kills BTTS | Compute 1-e^(-xGA) times 1-e^(-xGB) |
| Over 3.0, BTTS yes at 1.35 | Heavy total means both score | 3-0 satisfies over and kills BTTS | Weight the scoreline distribution |
| Under 2.5, BTTS yes at 1.90 | Scrappy game, someone will nick it | Low total can mean two sterile attacks | Low totals are not bullish for BTTS |
| Strong home side, BTTS yes at 1.30 | Favourite will score, so it hits | Favourite scoring is exactly half the market | Buy the away leg and price it |
| BTTS no at 1.90 vs a clean-sheet machine | 0-0 is a common scoreline | They may not register a shot on target | Check post-shot xG, not the record |
The Teams That Break the Pattern
Every rate has exceptions and the exceptions are where the money is. Three categories account for most of the profitable BTTS bets I have taken.
Defensive clean-sheet machines. The most dangerous teams to back BTTS yes against. A side conceding under 0.6 xG per match has not merely kept clean sheets, it has made it hard for opponents to generate anything. Against these, back the no, and the price is often better than 2.00 because the market prices their clean sheets off the final scoreline rather than off their chance suppression.
Parked-bus heavyweights. A dominant side playing a team with nothing to lose produces a high total, a low BTTS probability and a public that reads the total line as bullish. This is the most common way a bad BTTS bet loses.
New managers and promoted sides. A promoted side's attack is priced on last season's second division numbers for about six weeks, and books over-compensate for the step up. A new manager's first three fixtures are unpriced because training data does not exist yet. Both produce short-lived edges: take them once, do not build a model on them. One cuts the other way. Promoted sides whose forwards were sold in the summer see their xG collapse within a month and the books take three to notice. That one is repeatable across every promotion in Europe every year.
Pairing BTTS With the Over Line
Combining the two gives a bet that needs three goals with one for each side, so the winning scorelines are 1-1, 2-1 and 1-2, plus 2-2 on the quarter line. The combined price typically lands between 3.20 and 3.80, implying 26 to 31 per cent, and a book of roughly 112 per cent once you account for the two margins. That is a wide book, which makes it poor as a standalone bet and reasonable inside a small accumulator.
The sharp version is not the combination but the relationship. Once you know BTTS yes is 0.60 and over 1.5 is 0.78, the BTTS-yes-plus-Over-1.5 probability is not the product of 0.60 and 0.78, because those two events are dependent. It is the sum of the joint probabilities of 1-1, 2-1 and 1-2, which a Poisson model gives you as a specific number. Work the model out first, then look at the offered price. A combo priced above 3.90 with a model probability above 30 per cent is a real edge; a combo at 3.10 is a gift to the book.
For the mechanics of the total line itself, including quarter lines and pushes, read the over and under guide. If you want to go further into scoreline modelling rather than three-goal bands, the correct score guide is the natural next step, since BTTS-plus-Over-1.5 is literally a four-outcome correct score bet.
How to Actually Bet It, Step by Step
About twenty minutes per fixture.
One. Pull the xG figures for both teams, split by home and away where the data allows. Do not use a season-wide average for a side that only scores at home.
Two. Apply any regression correction you believe in: big chance miss rate, manager change, congestion, promoted-team baseline. Each is worth a few points on BTTS probability, which at 1.80 is a couple of points of price.
Three. Compute (1 - e^-xG_A) multiplied by (1 - e^-xG_B). Round to one decimal place in per cent, because the model is not more accurate than that.
Four. Compare with the best price across three books. Take the best price, not the most familiar book, which is the principle behind our value bet detection guide.
Five. Require a gap of at least four percentage points between your probability and the break-even. Smaller gaps sit inside the model error.
Six. Stake in fractions. The mathematics are in value betting mathematics.
Seven. Log everything, including the bets you passed on. A model that logs only its winners will convince you it works. The discipline is in our betting journal method, and it is the only way to learn whether the gap was skill or variance.
The Summary
BTTS is a two-outcome market priced from two numbers you can look up. Compute the probability, take the best price, demand a four-point gap, and stake in fractions. Expect roughly 55 per cent to hit over a season and expect the margin to take 4 to 9 points of it, which means the edge lives in a handful of fixtures a week rather than a hundred. A good BTTS bettor is right more often than the price suggests on a small number of bets, and wrong more often than the price suggests on the rest.
