After 28 matchdays of the 2025/26 Premier League season, Manchester City sit four points clear at the top with 67 points. On the surface, a dominant campaign from a dominant team. But our prediction model tells a rather more interesting story.
The points are right, but the margin is not
City are performing well — the model agrees. But based on the quality of chances created and conceded across 28 matches, we would expect roughly 62 points, not 67. That five-point gap represents a team who have been slightly fortunate in close games, winning three matches where the underlying data suggested a draw was the most likely outcome.
That is not unusual for a title-winning side. But it is worth flagging: City's buffer at the top is roughly two games more fragile than the raw standings imply.
Arsenal: the case for genuine improvement
Arsenal sit second with 63 points. Our model puts their predicted total at 65 — they are actually underperforming their underlying data. Their two-goal loss to City in December was a statistical outlier; City's prediction confidence in that match was only 54%. Remove three or four high-variance results and Arsenal, on merit, should be level-pegging at the top.
Arsenal are a better team than their position in the table suggests. The data says they deserve to be leading this league right now.
Who else is living beyond their means?
Chelsea sit fourth with 52 points — four above their predicted 48. Tottenham are broadly in line at 48 actual versus 47 predicted. Man United at 41 are, if anything, slightly over-performing their genuinely poor underlying statistics.
What to watch for in the run-in
Our current model gives Arsenal a 38% chance of winning the title, up from 22% at the halfway point. City remain favourites at 54%. The table is not lying — but it is not telling the whole truth either.