Premier League Tiebreakers: How Level Points Are Split

Arsenal went to Anfield on 26 May 1989 needing to win by two clear goals. Not one. Two — because a single-goal win would have left both clubs on 76 points with an identical goal difference of plus 37, and the third criterion, goals scored, was where Arsenal held their advantage. They won 2-0 in the final minute of the season and took the championship on that third column.

That is the sharpest illustration English football has of what happens when points run out. The sequence has changed since, the country has changed competitions since, and the arithmetic still works the same way.

The order the Premier League table is sorted in

Three criteria, applied strictly in order, and nothing else.

Points first. Then goal difference — goals scored minus goals conceded across all thirty-eight matches. Then goals scored, in isolation, which rewards the club that has been in more high-scoring games rather than the one that has kept things tight.

1. 1PointsThe clubs finish the season on the same total. Nothing about the results between them is consulted at any stage.
2. 2Goal differenceGoals scored minus goals conceded across the whole season. The great majority of ties end here.
3. 3Goals scoredIf goal difference is identical too, the club that has scored more finishes higher, irrespective of what it conceded.
4. 4Same positionIf all three are equal the clubs are treated as occupying the same place in the table.
5. 5Deciding matchWhere the tie settles the title, relegation or a European place, a play-off at a neutral venue is arranged by the Board.

The sequence set out in the competition's ordering rules, described in outline.

The critical thing about that order is what is missing from it. There is no head-to-head step. There is no away-goals step. There is no consideration of who beat whom, when, or by how much, beyond the aggregate goal columns that every match feeds into equally. A club that has done the double over its rival and finished level on points and goal difference gets no credit for it whatsoever.

The rule applies to every position from first to twentieth. There is no separate arrangement for the title, no different sequence for the relegation places, and no adjusted version for European qualification. What changes at those three points is not the sequence but the consequence — a tie that cannot be broken by any of the three criteria is left standing everywhere except where something material hangs on it, and there the Board has to arrange a match.

There is a small practical wrinkle inside "deemed to occupy the same position". A printed table still has to list the clubs in some order, and the convention in England is alphabetical where every criterion is level — a presentational choice with no competitive weight attached to it at all. Nothing follows from being listed above a club you are genuinely level with. Central distributions are a different matter, because the Premier League pays merit money on a ladder tied to final position, and a true dead heat would have to be resolved somehow for payment purposes even if the table were content to leave it standing. How that ladder works is set out in the guide to Premier League prize money.

Points deductions are applied to the points column before any of this runs, which means a sanction can create a tie that would not otherwise have existed. How deductions are imposed, and how they interact with a season already under way, is covered separately in the guide to Premier League points deductions.

Why goal difference comes before goals scored

The choice of second criterion is a statement about what the competition thinks a season is worth measuring, and goal difference is the neutral answer. A goal prevented counts as much as a goal scored. A side that wins 1-0 nineteen times finishes on the same goal difference as one that wins 4-3 nineteen times, and the table treats them as equals until the third column.

Goals scored, sitting underneath it, then breaks the deadlock in favour of the more attacking side. That ordering is deliberate and it is not the only one available. A league could put goals conceded third and reward the meaner defence; England does the opposite.

The practical effect is that the third criterion almost never gets used. Two clubs finishing on identical points is uncommon at any given position; two clubs finishing on identical points and an identical goal difference is rare enough that the English top flight has produced it once in the modern era, in 1989. Everything else has resolved at goal difference.

Which raises a fair question about whether the third criterion earns its place at all. It does, for a reason that has nothing to do with frequency. Without it, the fallback would be a play-off, and a play-off match after a thirty-eight-game season is a competitive absurdity — one afternoon deciding what nine months could not, with both squads mid-holiday and a neutral ground to fill at short notice. Goals scored is a cheap way of making that outcome vanishingly unlikely.

There is a second-order effect worth noticing. Because the third criterion favours goals scored rather than goals conceded, a club chasing a tight finish has no incentive to shut up shop in a match it is already winning. The table rewards the fourth goal.

Season length changes the shape of this more than people expect. The Premier League ran with twenty-two clubs and a forty-two-match programme for its first three seasons before reducing to twenty clubs from 1995-96. Four extra fixtures per club gave the goal-difference column four more matches in which to spread out, and a column that has spread out produces fewer exact ties. Shortening the season made identical finishes marginally more likely rather than less. That was not a consideration anybody raised when the reduction was agreed, and in three decades since it still has not produced a play-off.

Goal average, and the change that got rid of it

Before 1976-77, English football did not use goal difference at all. It used goal average: goals scored divided by goals conceded, expressed as a decimal, and applied as the first tiebreak after points.

The two measures behave completely differently. Goal average is a ratio, so it rewards a low concession rate out of all proportion. A side scoring 60 and conceding 30 has an average of 2.000 and a difference of plus 30. A side scoring 90 and conceding 50 has an average of 1.800 and a difference of plus 40. Under goal average the first side finishes above the second; under goal difference the second finishes above the first. Same two seasons, opposite answers.

Method	Formula	Rewards	In use
Goal average	Goals scored divided by goals conceded	A mean defence, disproportionately	Until 1975-76
Goal difference	Goals scored minus goals conceded	Attack and defence equally	From 1976-77

Goal average decided real championships. Manchester United took the 1964-65 title ahead of Leeds United on it, both clubs on 61 points. Arsenal beat Preston North End to the 1952-53 title on it. Portsmouth finished above Wolverhampton Wanderers on it in 1949-50. Those are titles awarded on a decimal fraction that most supporters in the ground could not have calculated at full time.

The Football League switched to goal difference for 1976-77, and the reasoning was partly comprehensibility and partly the sense that a ratio was quietly rewarding negative football. It is one of the few administrative changes in English football that essentially nobody has argued for reversing. The competition the change was made in no longer exists in the same form, but the rule survived the creation of the Premier League in 1992 untouched, as did most of the old First Division's competitive furniture.

The deciding match the rules provide for

If the three criteria all come out level and the tie determines the championship, relegation, or qualification for a UEFA competition, the Premier League's rules provide for a play-off. One match, neutral venue, arranged by the Board, with extra time and penalties available if it finishes level.

1. 1Season ends levelTwo clubs finish on identical points, identical goal difference and identical goals scored.
2. 2The tie mattersThe position at stake is the title, a relegation place, or a place in a UEFA competition.
3. 3The Board arranges a matchA single fixture at a neutral ground, on a date set by the Premier League after the final round.
4. 4Extra time and penaltiesIf the ninety minutes ends level, extra time follows, and a shoot-out after that if needed.
5. 5The result standsThe winner takes the higher position for every competition purpose that flows from it.

The outline the competition rules provide for. No Premier League season has required it.

It has never happened. That is not luck so much as arithmetic: three independent criteria all landing level across thirty-eight matches is an event with a very small probability attached to it, and the competition has run for a few decades rather than a few centuries.

The nearest thing English football has to a precedent sits well down the pyramid and in a different era, when the Football League used test matches and play-off deciders more freely than the modern rules would allow. The modern Premier League provision is best understood as insurance rather than as a live possibility.

One consequence is worth thinking through. A neutral-venue decider would have to be scheduled after the final round of the season, which is already crowded with the FA Cup final, the EFL play-off finals at Wembley, and in some years a European final involving English clubs. Where it would go in the calendar is not obvious, and the rules sensibly leave that to the Board rather than specifying a date.

Head-to-head, and why English football does not use it

Nowhere in the Premier League's ordering rules do the results between two tied clubs appear. This surprises people, because a large part of Europe does the opposite.

La Liga applies head-to-head before goal difference. Two clubs level on points in Spain are separated by the aggregate of their two meetings, and only if those are level does the wider goal difference come into play. Serie A has used head-to-head as well, alongside a provision for a deciding match at the top of the table. The Bundesliga sits closer to England, taking goal difference first.

The English argument against head-to-head is that it is a small sample doing a large job. Two matches out of thirty-eight, played months apart, one at each ground, with different personnel available and possibly a red card in one of them, is a thin basis for deciding a championship. Goal difference uses every match the club has played. It is a worse measure of the specific rivalry and a better measure of the season.

The counter-argument is real and worth stating. Head-to-head answers the question supporters actually ask — who was better against the club they finished level with — and it has the merit of being decided on the pitch between the two teams concerned rather than by results against Burnley in November.

England has simply picked the other side of that argument and stuck with it for a very long time. There is no live campaign to change it, and the Football Association's wider rulebook takes the same approach in most competitions it runs.

Ties at the bottom of the table

Relegation works on the identical sequence. Points, goal difference, goals scored, and no head-to-head anywhere.

Fulham finished the 2007-08 season on 36 points. So did Reading. Fulham's goal difference was better, and that alone is the difference between a club that stayed in the Premier League and one that went into the Championship — with everything that follows from it, including the parachute payments and the wage restructuring covered in the guide to the cost of relegation.

The financial asymmetry is what makes tiebreaks at the bottom feel harsher than tiebreaks at the top. A club that misses the title by goal difference has still qualified for the Champions League, still has its broadcast money, still has its squad. A club relegated by goal difference loses a substantial share of its central distribution, triggers relegation clauses across its contracts, and faces a promotion campaign against clubs that have been building for it. The guide to Premier League relegation sets out how the drop is calculated and what it costs.

Three-way ties at the bottom are the case the rulebook handles least gracefully. Three clubs level on points for the last two safe places would be sorted by goal difference and then goals scored in the ordinary way, and only a genuine three-way identity across all three criteria would force the Board to arrange anything at all. What that arrangement would look like is not obvious, because a single fixture cannot separate three clubs, and the ordering provision does not spell it out. The probability is small enough that nobody has felt the need to.

Nothing in the rules acknowledges any of that. A tie for seventeenth is broken by the same three criteria as a tie for seventh, and the Board's play-off provision covers the relegation places explicitly. That evenhandedness is the right call, and it is also the reason a manager in April is doing arithmetic that has nothing to do with the opponent in front of him.

Ties for a European place

European qualification adds a layer that has nothing to do with tiebreaks and everything to do with counting.

The number of English clubs in each UEFA competition is not fixed by league position alone. Domestic cup winners take places; a club that wins a European competition can occupy an additional place; and the number of Champions League berths available to England has varied with UEFA's coefficient arrangements. Two clubs finishing level on points may therefore both qualify, or neither, depending on results in competitions neither of them played in. The full picture is in the guide to Premier League European qualification.

Where the tiebreak does bite is on which of the two goes into which competition, and that can be worth a great deal. Manchester United and Chelsea both finished 2019-20 on 66 points, United third on goal difference. In that instance both clubs reached the same competition and the ordering was cosmetic. It would not have been if the tie had fallen a place lower.

There is also a subtlety about cup winners that catches people out. If the FA Cup or EFL Cup is won by a club that has already qualified for Europe through the league, the place its win generates passes down the table — and it passes to a specific league position, which means a tie for that position becomes suddenly material. A club sitting seventh, level on points with the club in eighth, may find that a Wembley result in May has turned an ordering nobody cared about into the difference between European football and none.

The title races that turned on the last two columns

Manchester City and Manchester United both reached 89 points in 2011-12. City finished with a goal difference of plus 64, United with plus 56, and the title went to City on that eight-goal margin after a final-day match at the Etihad that turned in stoppage time.

• 89Points, each club
• 64Manchester City goal difference
• 56Manchester United goal difference

Final figures for the two Manchester clubs. Goal difference separated them; points did not.

The arithmetic of that season contains a detail that most retellings skip. City won 6-1 at Old Trafford in October. That single result moved City's goal difference by five and United's by five in the opposite direction, a relative swing of ten goals — more than the eight-goal margin the title was eventually settled by. A derby in autumn decided a championship in May, and it did so through a column nobody was looking at when it happened.

Arsenal's 1988-89 win at Anfield is the only occasion the English championship has been settled by the third criterion. Both clubs finished on 76 points. Both finished with twenty-two wins, ten draws and six defeats. Both had a goal difference of plus 37. Arsenal had scored 73 and conceded 36; Liverpool had scored 65 and conceded 28.

Arsenal scored73goals

Arsenal conceded36goals

Liverpool scored65goals

Liverpool conceded28goals

Both clubs finished on 76 points with a goal difference of plus 37. Goals scored separated them.

Show the numbers

Arsenal and Liverpool, 1988-89 First Division

Item	Value
Arsenal scored	73goals
Arsenal conceded	36goals
Liverpool scored	65goals
Liverpool conceded	28goals

Liverpool had much the better defence and lost the title anyway, which is the clearest possible demonstration of what the third criterion is for. It is not a neutral measure. It picks a side. The list of English champions and the seasons behind them is set out in the guide to Premier League title winners.

How the EFL orders its tables

The Championship, League One and League Two run on the same first three criteria — points, then goal difference, then goals scored — and any supporter moving between the Premier League and the EFL will find the tables ordered the way they expect.

Two structural differences change how much the tiebreak matters.

The first is the play-off system. In the Championship, the four clubs finishing third to sixth enter a knockout, so a tie for sixth place is a tie for entry to a competition rather than for a final position, and the stakes attached to a single goal are correspondingly higher. The mechanics are covered in the guide to how the Championship play-offs work, and there is a separate, more detailed treatment of Championship tiebreakers for the divisions below the Premier League.

The second is the length of the season. A forty-six-match programme produces larger goal-difference spreads than a thirty-eight-match one, which makes exact ties slightly less likely at any given position even though there are more positions in play.

Where the two rulebooks diverge is in the wording after goals scored, and this is a point on which it is worth being honest rather than tidy. Competition regulations get amended, and the provisions covering what happens when every criterion is exhausted have not always read identically across the Premier League, the EFL and the National League. Anyone who needs the answer for a specific division in a specific season should read that competition's regulations for that season rather than assume the Premier League wording applies throughout the pyramid. The structure of the divisions themselves is set out in the guide to the English football pyramid.

Calculating goal difference correctly

Goal difference is goals scored minus goals conceded across every match in the competition. Not per match. Not averaged. A straight subtraction of two season totals.

The details that trip people up are all about what feeds into those two totals.

Own goals count. An own goal appears in the goals-for column of the club that benefits and the goals-against column of the club that scored it, exactly like any other goal. It is not credited to any individual player — the Golden Boot guide covers why that matters for a scorer's personal record — but for the table it is indistinguishable from a header at the far post.

Penalties count, including those converted in stoppage time of a match already decided. A penalty shoot-out does not count, because Premier League matches do not have them; goal difference is a league-competition figure and only league matches feed it.

Goals in a match that is abandoned and ordered to be replayed do not stand, because the fixture itself does not stand. Where the Board awards a result rather than the clubs playing it out, the scoreline it applies enters the goal columns in the ordinary way.

Cup goals never count. Neither do goals in European matches, in the EFL Cup, or in any competition other than the league itself. A club can score forty times in cup competitions and see none of it reflected in its league goal difference.

The awkward case is a season that does not finish. When the 2019-20 League One and League Two campaigns were curtailed, final placings were settled on an unweighted points-per-game basis, because clubs had played different numbers of matches and the raw points column had stopped meaning the same thing for everybody. Points per game was not a tiebreak in that instance. It replaced the first criterion outright, with goal difference then applied underneath it in the usual way. That was a competition-specific decision taken in extraordinary circumstances rather than a standing provision, and it remains the clearest example of the English ordering sequence being adapted at the top while the rest of it carried on working normally.

What a goal-difference gap does to a manager in April

Two clubs go into the final month level on points, one of them eight goals better off. The trailing side needs a win more than the leading side does, because a draw leaves the deficit intact and a narrow win does almost nothing to it.

This shapes behaviour in ways that are visible from the stands. A manager 3-0 up with fifteen minutes left, in a season where goal difference is live, does not bring on a defensive midfielder. He leaves the front three on and accepts the risk, because the fourth goal is worth something and the eighty-ninth-minute consolation conceded is worth as much against him. The same manager, in a season where his club is six points clear, takes the strikers off and protects legs for the following Wednesday.

It also affects how a heavy defeat is treated. Losing 5-0 costs the same three points as losing 1-0, and it costs four more goals of difference, which is why sides in a close race sometimes appear to be playing for damage limitation in a match they have plainly lost.

The uncomfortable version of this arrives at the bottom of the table. A club two goals adrift on goal difference with a match remaining is effectively playing for a win rather than a draw, whatever the neutral arithmetic of points might suggest, and that pushes a squad already short of confidence into an open game against opponents with nothing to lose. Goal difference is a fair rule. It is not a kind one.

Where the wording lives, and how to read it

The authoritative text is the Premier League Handbook, published for each season and containing the competition rules in full. The ordering provisions sit in the section dealing with the league competition itself, alongside the rules on fixtures, squad registration and the composition of the league. The EFL publishes its own regulations on the same basis, and the FA Handbook carries the rules of the wider game that both sit underneath.

Two habits make reading them easier. Look for the ordering rule and the play-off rule as separate provisions rather than as one paragraph, because the first says how the table is sorted and the second says what happens when sorting fails. And check the season, because a handbook is reissued annually and a rule quoted from an old edition may have been amended.

For anything historical — a goal-average title from the 1950s, a goal-difference finish from the 1980s — the competition's own archive is the record, and the seasons themselves are covered across the England guides and the wider football section.

Anyone who has ever argued about this in a pub will have met the person who insists head-to-head applies. It does not, it never has in the English top flight, and the Handbook says so in about two lines.