A manager’s points per game is a squad stat
Autumn in the Premier League means sack-race season, and the case for or against every embattled manager is made with the same number: points per game. It is quoted on broadcasts, compared against predecessors, and used to declare whether a new appointment is working. It is a reasonable summary of results. It is a poor measure of management, because over the length of a typical managerial stint it is dominated by things the manager did not choose: the money behind the squad, the order of the fixtures, and the randomness of a low-scoring sport.
Money explains most of the table
The most robust finding in football economics is how closely league position tracks wage spending. Across long periods, a club’s wage bill relative to its rivals explains the large majority of where it finishes. That does not leave managers irrelevant; it means the range of outcomes available to any given manager is set largely before he arrives. A coach at a club with the fifth-highest wage bill who finishes fifth has done roughly what the squad’s cost predicts. A coach at a club with the fifteenth-highest wage bill who finishes tenth has outperformed by a wide margin, even though his points per game is far lower.
The fixture list front-loads the noise
Premier League schedules are not balanced over short windows. A manager whose first ten matches include several of the league’s strongest sides will post a weak points-per-game figure that says more about the calendar than about his methods, and one handed a soft opening run will look like a revelation. Over a full season the fixtures even out. Over the few months that most sacking decisions are based on, they do not. Comparing two managers’ records without adjusting for opponent strength is comparing their schedules as much as their work.
Results are a noisy sample of performance
Football is low-scoring, which makes results unusually volatile relative to underlying play. A team can create the better chances in a match and lose it, and over ten or fifteen games those swings do not reliably cancel out. Expected goals and shot-quality measures exist precisely because the underlying performance stabilizes faster than the points do. A manager whose team is generating good chances and conceding few, but whose points lag behind, is more likely to be suffering from finishing variance than from a flawed plan. The points per game is the last place that improvement shows up.
The new-manager bounce is mostly regression
The strongest argument for sacking a manager is the bounce: results tend to improve after a change. They usually do. They also usually improve when a struggling club keeps its manager. Academic studies that compared clubs that changed managers after a poor run with similar clubs that did not have found that both groups tend to recover by similar amounts, because a poor run is partly bad luck and bad luck rarely persists. Clubs sack managers at the bottom of a slump, so the recovery that follows gets credited to the new appointment. Part of it would have happened anyway.
This does not mean changes never help. Some appointments clearly improve a club, and some managers are plainly out of their depth. The point is that the bounce cannot by itself tell the two cases apart, because it shows up whether the decision was good or bad. Judging a new manager by his first eight or ten matches mostly measures how unlucky his predecessor was in his last eight or ten. The real test of an appointment comes later, once the fixtures and the finishing have had time to even out.
Squad turnover muddies the comparison
Comparisons between a manager and his predecessor assume the squad stayed constant, and in the modern Premier League it rarely does. Transfer windows, injuries to key players, and the rotation demands of European competition all change what a manager is working with. A coach who inherits a squad mid-rebuild is not competing on the same terms as the one who left with the core intact. Points per game flattens all of that into a single comparison that sounds precise and is not.
What a better scorecard looks like
A fair evaluation of a manager would compare results against what the squad’s wage bill predicts, adjust for the strength of opponents faced, and give more weight to underlying measures like expected goal difference than to points over short spans. It would also account for squad changes and injuries. None of this is exotic, and clubs increasingly use versions of it internally. The public conversation, though, still runs on the single number that is easiest to quote and least able to separate the manager from his circumstances.
When the number does mean something
Points per game is not meaningless. Over long tenures, measured against a stable squad and a balanced schedule, it does start to reflect a manager’s impact. Persistent overperformance relative to wage spending across multiple seasons is one of the clearer signals in the sport, and a handful of managers have shown it. The trouble is that sacking decisions are rarely made on that kind of evidence. They are made in October and November, on a dozen games, which is the stretch where the number is least reliable.
How to read it on StatLine
StatLine’s EPL tables carry team results, goals for and against, and the standings needed to put a manager’s record in context. When a manager comes under pressure, compare the club’s position with where its resources suggest it should be, check the difficulty of the fixtures played so far, and look at whether goal difference is tracking the points or diverging from them. A gap between performance and points is usually temporary, and it usually closes without anyone being sacked.
The honest read
A manager’s points per game, over the span on which most sacking decisions are made, is mainly a squad stat with a schedule effect and a large helping of finishing luck. Wage bills set the realistic range, fixtures distort short windows, results lag behind performance, and the new-manager bounce is largely the regression any struggling team would have experienced. The number has value over years, not months. When the sack-race graphic appears, the honest version would include the wage-bill rank beside every name.