If you sack your manager when your team is in the bottom three, on average the next manager delivers +0.1 points per game more than the one you sacked. If you hire a genuinely top-class coach, you can get +1.0 PPG. The median bounce is the noise between those two extremes, and most of what fans attribute to the new-manager bounce is just normal mid-season drift.
What we measured
We tracked every permanent mid-season managerial change in the five major European leagues from 2019-20 to 2024-25, where the change took effect between matchday 8 and matchday 30. For each, we compared the outgoing manager’s points per game (PPG) over his last five league matches to the incoming manager’s PPG over his first five.
That gives us 147 qualifying changes across 30 league-seasons.
As a control, we took the same statistic, PPG over a five-match window, for the top-five finishers in each league-season, comparing their PPG in matchdays 15-19 to matchdays 20-24. These teams almost never change manager, so any drift in the control group is baseline form variance, not a manager effect.
The data
| Group | Avg PPG before | Avg PPG after | Avg bounce | % with bounce |
|---|---|---|---|---|
| Changed-manager teams (147 changes) | 0.65 | 0.77 | +0.12 | 55% |
| Top-five control (no change) | - | - | +0.05 | ~52% |
| Net bounce attributable to a change | +0.07 |
Across all five leagues the per-league bounce sits in the same band: +0.08 PPG in La Liga and Ligue 1, +0.10 in the Bundesliga, +0.16 in Serie A, +0.20 in the Premier League. The Premier League’s higher average is partly because most of its mid-season sackings happen in genuine crisis situations, where the outgoing manager is far below replacement level and the league’s high overall points baseline makes any recovery visible.
The control group bounces too, by about +0.05 PPG. That is just normal form variance over a five-match window. It is the floor below which a bounce is meaningless.
The net difference attributable to actually changing the manager is roughly +0.07 PPG per game. Across a 15-game run, that is one extra point. It is not nothing, but it is close to nothing.
What the bounce actually looks like
The +0.07 PPG headline hides the real story, which is that bounces are bimodal. Most bounces cluster around zero, give or take 0.4 PPG. A small set of bounces, the famous ones, are clear outliers above +0.7 PPG and are sustained over many matches.
Bounces that were real
Antonio Conte at Tottenham in November 2021 took Spurs from Nuno’s 1.0 PPG to 2.0 PPG in his first five matches, sustained through the season, and qualified the club for the Champions League. Unai Emery at Aston Villa in November 2022 produced a +1.4 PPG bounce from Steven Gerrard’s 0.8 PPG and lifted Villa from 17th to 7th. Xabi Alonso at Bayer Leverkusen in October 2022 took a 17th-placed side to sixth and into the Europa League, then went on to win the Bundesliga unbeaten the next season. Hansi Flick at Bayern Munich in November 2019 took over a fourth-placed Bayern and won the treble. Eddie Howe at Newcastle in November 2021 lifted the team from 19th to 11th and built the side that qualified for the Champions League two seasons later. Stefano Pioli at AC Milan and Carlo Ancelotti at Everton produced smaller but still clear bounces in the same window.
Bounces that were regression
Most mid-season sackings happen at the bottom of the table, where the outgoing manager has been performing well below replacement level. Sacking him gives the new manager a low base from which to bounce. The new manager’s first five matches are usually a slight improvement, often just enough to confirm the change was overdue.
Examples across the dataset: Watford sacked three managers in 2019-20 and were still relegated. Genoa changed managers twice in 2019-20 and went down. Sampdoria in 2022-23 changed manager at matchday 10, briefly improved, then finished bottom with 19 points. Schalke 04 in 2020-21 went through five managers and finished with the lowest points total in Bundesliga history.
The contrast is not subtle. The famous bounces came from clubs that had the squad and structure to support a manager change. The relegation-bound clubs sacked their way through the season without changing the trajectory.
What we tried that didn’t work
We first treated every mid-season change as a uniform bounce of around +0.3 PPG. That systematically overestimated the bounce for bottom-of-the-table clubs, where the new manager rarely had the squad to capitalise, and underestimated it for top-half clubs that hired genuinely elite managers.
We then built a per-league fixed adjustment, +0.2 PPG in the Premier League, +0.1 PPG in Ligue 1, and so on. That was closer, but missed the bimodal distribution. Treating a Conte appointment and a Schalke carousel as the same kind of event was the wrong shape.
What we arrived at is a tiered adjustment that looks at the quality of the incoming manager and the trajectory of the squad. Top-class appointments carry a large adjustment, around +0.7 PPG, sustained for 15-20 matches. Mid-tier appointments carry +0.2 PPG, mostly fading by mid-season. Bottom-table changes carry zero or negative adjustment. The adjustments tighten as the actual record of the new manager emerges.
Known limitations
Five-match windows are short. The natural variance in PPG over five matches is high, and a single freak result, a derby win or an injury-time equaliser, can swing the delta by 0.4 PPG.
The dataset does not include cup matches. A manager bounce that comes mostly from cup performance will look weaker in this data than it really was.
Caretaker stints of fewer than five matches are excluded, which removes some of the noisier changes but also removes some real rebounds.
COVID-disrupted seasons (2019-20 Ligue 1 and Bundesliga especially) have unusual fixture patterns that distort the before/after windows. Those rows are flagged in the source dataset but not removed.
The control group uses top-five finishers, which are almost never sacked. A better control would be teams that sacked a manager earlier in the season and are now on a non-permanent appointment, but that is a much smaller sample.
What this means for an analysis
A model that treats every mid-season managerial change as a +0.3 PPG boost will systematically over-predict the form of bottom-table clubs and under-predict the form of clubs that hire elite coaches.
The right adjustment is bimodal. A change at a top-half club to a coach with a strong track record warrants a meaningful boost. A change at a bottom-table club to a journeyman coach warrants nothing, and may warrant a small negative if the squad is also damaged.
For a reader following coverage across leagues, the new-manager bounce is one of the most volatile inputs the model uses. A 50-50 call on a match played two weeks after a manager change at a relegation-threatened club carries much wider tails than the same call played three months into an elite coach’s tenure.
A read is a lens for separating signal from noise. The new-manager bounce is mostly noise. The exceptions are real, but they are not the average.