xG, Expected Points and Field Tilt: The Stats That Actually Predict Premier League Success

Expected Points Already Contradicts the Early 2026/27 Table as Arsenal Lag Their xG Model

Underlying statistics aim to explain what the Premier League table cannot: why results happened and what comes next. After six rounds of 2026/27, the standings did not fully match the expected-points model, with Arsenal and Manchester City among the sides whose points totals outran the quality of chances they created.

Competition: Premier League 2026/27, analysed after six rounds
Key metric: Expected-goal difference = xG created minus xG conceded
Headline gap: Arsenal had roughly 12.6 expected points against 15 actual
Bigger gap: Manchester City had about 9.9 expected points from a maximum 18
Defensive benchmark: Arsenal conceded approximately 0.89 xG per match by mid-October
At stake: separating sustainable form from finishing streaks and goalkeeping hot spells

The numbers

Arsenal: approximately 12.6 expected points versus 15 actual after six matches of 2026/27
Manchester City: around 9.9 expected points despite collecting the maximum 18
Arsenal defensively: roughly 0.89 xG allowed per match by mid-October, the best in the division
Strongest defences by chance quality: Arsenal, Manchester United, Liverpool and Nottingham Forest
Crystal Palace: just over 2.0 xG allowed per match
Brighton: also close to 2.0 xG allowed per match despite a strong early league position
Field tilt: Arsenal, Manchester City, Tottenham and Manchester United all above 63% early in 2026/27

Why expected-goal difference is the metric to trust

Expected goals estimates the likelihood of a shot becoming a goal from its location, angle and the type of assist. A penalty is valued far higher than a speculative effort from distance because it is converted far more often. Summing every chance produces a team's attacking xG total.

Expected-goal difference subtracts xG conceded from xG created, measuring performance at both ends. A consistently positive figure points to long-term strength even while results lag. A negative figure is an early warning: narrow wins built on conceding better chances than you make rarely survive a full season.

What expected points revealed about Arsenal and Manchester City

Expected points converts chance quality in each match into a probable outcome, estimating how often each side would win or draw if the same chances were replayed rather than simply awarding three points to the higher scorer.

Manchester City's perfect start was the clearest example of results outrunning process. Their maximum 18 points against roughly 9.9 expected points did not make those wins invalid — clinical finishing, goalkeeping and individual quality are fundamental — but the gap suggested the run was stronger than the balance of chances would normally produce.

Arsenal's smaller gap, about 12.6 expected points against 15 actual, points to a milder overperformance. The statistic has limits: matches are not replayed under identical conditions, and teams change tactics after scoring. A side protecting a lead can concede low-quality shots while remaining comfortable, so expected points must be read alongside tactical context.

Non-penalty xG, shot quality and the true picture of an attack

Penalties are valuable but rare, and they can inflate an attacking record. Non-penalty expected goals strips them out, giving a clearer read on how effectively a team creates from open play and non-penalty set pieces — and revealing whether a scoring run leans on favourable refereeing decisions that may not continue.

The same logic applies to individuals. A forward with a high goals total can appear more dangerous than a teammate until penalty and non-penalty output are separated. Shot volume still matters, but quality is more informative: ten low-probability attempts from distance can pose less genuine threat than two clear chances inside the penalty area.

Expected goals against: Arsenal's defence and Brighton's warning

Goals conceded can mislead over a short period. A goalkeeper may produce several outstanding performances, or opponents may repeatedly miss clear openings. Expected goals against estimates the quality of chances a team allows and can flag defensive problems before they reach the scoreline.

Arsenal's early 2026/27 numbers showed the metric's value. By mid-October the club was allowing approximately 0.89 xG per match, the best figure in the division, with Manchester United, Liverpool and Nottingham Forest also among the strongest at restricting chance quality. At the other end, Crystal Palace were allowing just over 2.0 xG per match, while Brighton sat close to 2.0 despite a strong early league position — a sign their results came with defensive exposure that could eventually prove costly. The most sustainable defences do not merely collect clean sheets; they stop opponents creating clear chances at all.

Box entries, field tilt and why possession flatters

Not every dangerous attack ends in a shot. A poor final pass, a heavy touch or a well-timed challenge can keep a promising move out of the xG data entirely. Penalty-area entries and touches capture those situations, and the location matters: cut-backs from the byline and passes into central areas produce better shooting opportunities than hopeful crosses into a crowded box. A side reaching dangerous areas but underperforming its xG may be close to a breakthrough; a side failing to enter the box at all has a structural problem, not bad luck.

Possession measures how long a team has the ball; field tilt examines where, comparing each side's share of touches or completed passes in the attacking third. Circulating between defenders can dominate possession without threatening anyone, whereas high field tilt shows more of the match is played near the opponent's goal. Early in 2026/27, Arsenal, Manchester City, Tottenham and Manchester United all recorded field-tilt figures above 63%, with relatively high possession numbers to match — but the territorial presence was the more meaningful evidence of pressure. It is no guarantee of goals, since an organised opponent can defend its area and counter into space, yet repeatedly forcing opponents backwards is generally more predictive than possession alone.

Expected threat: valuing the pass before the assist

Expected threat, or xT, assesses how much a pass or carry increases the likelihood of a goal later in the possession. A sideways pass in defence has little value because it does not improve the attacking position; a progressive carry into the final third or a pass between the lines is rated higher because it moves the team into a more threatening area. The metric recognises players involved before the final pass, so a midfielder who repeatedly breaks defensive lines can be central to an attack without recording many goals or assists.

Talking points

Manchester City's 18 points against 9.9 expected points is either elite finishing or a warning — the answer decides whether their title challenge is built on sand.
Arsenal allowing just 0.89 xG per match is a stronger title indicator than any win they have recorded this season.
Brighton sitting close to 2.0 xG conceded a match while high in the table is the clearest sign of a fall coming.
Field tilt above 63% for Arsenal, Manchester City, Tottenham and Manchester United suggests possession stats are the wrong table to read.

What happens next

The next six rounds will test whether the expected-points gaps close. Manchester City's perfect start faces the sterner examination because large gaps between actual and expected points generally narrow, while Arsenal's defensive baseline of 0.89 xG per match gives them the platform to keep their points rate without relying on overperformance. Brighton and Crystal Palace, both conceding near 2.0 xG per match, are the sides whose results are most likely to regress first.

Frequently asked questions

What is expected-goal difference in the Premier League?

Expected-goal difference subtracts expected goals conceded from expected goals created. A consistently positive figure suggests a team creates better chances than it allows, even when results lag.

Why did Manchester City have fewer expected points than actual points?

After six matches of 2026/27, City had around 9.9 expected points but collected the maximum 18, indicating their perfect start was stronger than the balance of chances would normally produce.

Which teams allowed the fewest expected goals per match in 2026/27?

Arsenal allowed approximately 0.89 xG per match by mid-October, the best in the division, with Manchester United, Liverpool and Nottingham Forest also among the strongest.