How Long Does the New Manager Bounce Actually Last?

The new-manager bounce describes an improvement in results after a coaching change, but it does not tell you how long that improvement will last or what caused it. Manchester United, Crystal Palace and Everton have produced very different versions of the story. Their results show why a manager’s arrival needs more investigation than simply adding extra confidence to the next fixture. When using our AI football predictions, the useful question is what has changed on the pitch and whether it matters against the next opponent.

What research actually says about the new-manager bounce

A 2016 study by Lucas Besters, Jan van Ours and Martin van Tuijl examined managerial changes in the Premier League across the 2000/01–2014/15 seasons. Its final matched sample contained 61 changes. The researchers compared clubs that changed managers with similar situations in which the same clubs, in other seasons, kept their managers. On average, replacing the manager did not improve performance relative to those comparisons, although individual changes had different outcomes. This distinction matters because a team can recover after a dismissal without the dismissal being responsible for the recovery.

The statistical problem is known as regression towards the mean: a particularly poor sequence can be followed by results closer to the team’s usual level. However, the research is not unanimous. A 2020 study by Yvon Rocaboy and Marek Pavlík, using English and French top-flight data from 2015/16 and 2016/17, found that dismissals could improve performance when teams had previously underperformed. The studies use different methods and samples; neither gives us a universal three-match or five-match period to apply to every appointment. We need evidence about the individual team.

Solskjaer’s Manchester United: 12 unbeaten games did not secure the improvement

Ole Gunnar Solskjaer’s opening Premier League run at Manchester United in 2018/19 consisted of ten wins and two draws. That delivered 32 points from 12 games, or 2.67 points per match. United then took eight points from their remaining nine league fixtures, with two wins, two draws and five defeats. Their rate fell to 0.89 points per match. The early results were substantial, but treating them as United’s new permanent level would have badly overstated the strength of the side during the run-in.

There were warning signs before the season ended. In a Premier League tactical analysis published on 19 April 2019, Adrian Clarke noted that United had allowed 20 shots against Watford and 18 against West Ham despite winning both home matches. His comparison showed shots conceded rising from 11.3 per 90 minutes across Solskjaer’s first ten league games to 14.5 across the next six. United had kept five clean sheets in the first window and none in the second. The defensive deterioration was measurable even while some results remained favourable.

That gives us a more useful explanation than saying the players stopped responding to a new manager. We could see opponents getting more attempts and United struggling to protect their goal. It did not make defeat certain in the next match, but it weakened the case for continuing to judge the team by its initial unbeaten run. For prediction purposes, a recent win should carry less reassurance when the performance behind it is moving in the opposite direction.

Glasner’s Crystal Palace: the strongest spell came later

Oliver Glasner’s first match in charge of Crystal Palace brought a 3–0 home win over Burnley on 24 February 2024. That looked like an immediate bounce, although Burnley played much of the match with ten men. Palace then went five league matches without another victory, leaving Glasner with one win from his first six. A rule that assumed the boost had already expired would have missed what followed: Palace won six of their final seven Premier League matches, drawing the other.

The quality of those later results matters. Palace won 1–0 at Liverpool, beat Manchester United 4–0 and finished with a 5–0 victory over Aston Villa. The club’s own season review recorded that five of the teams beaten during the run had played European football that season: Liverpool, West Ham, Newcastle, Manchester United and Villa. The improvement therefore extended beyond collecting points against the division’s weakest opponents, although those opponents’ condition and priorities still differed from match to match.

Player availability also changed. Palace’s report on the 4–2 defeat by Manchester City on 6 April records Michael Olise returning from injury for the closing stages. That was the final match before their seven-game unbeaten finish. His return does not isolate the cause of the turnaround, but it is a concrete change that a “new manager” explanation alone leaves out. Glasner had more time with the squad while an important attacking player became available again; both belong in an assessment of the improvement.

Two managerial changes, very different league-result patterns

Manager and club	Period	Wins	Draws	Defeats	Points per match
Solskjaer, Manchester United	First 12 league games, 2018/19	10	2	0	2.67
Solskjaer, Manchester United	Remaining nine league games, 2018/19	2	2	5	0.89
Glasner, Crystal Palace	First six league games, 2023/24	1	2	3	0.83
Glasner, Crystal Palace	Final seven league games, 2023/24	6	1	0	2.71

Points per match is calculated from three points for a win and one for a draw, rounded to two decimal places. These are successive league-result windows within each manager’s first season, with cup matches excluded. They describe what happened; they do not measure how many points the manager personally created. United’s strongest results came immediately, while Palace’s strongest sequence followed a difficult opening spell. Averaging these examples into a supposed standard bounce duration would hide the difference we need to understand.

Dyche’s Everton: beating Arsenal did not settle the next meeting

Sean Dyche began his Everton tenure by beating league leaders Arsenal 1–0 at Goodison Park on 4 February 2023. James Tarkowski headed in a corner from Dwight McNeil, two players Dyche had previously managed at Burnley. It was an identifiable way of winning the match: a set-piece goal alongside a clean sheet against strong opposition. Yet on 1 March, only 25 days later, the same Everton manager faced the same Arsenal manager at the Emirates and lost 4–0.

Those two results cannot tell us that Dyche’s influence lasted exactly 25 days. They can tell us that the first win did not remove the need to reassess the venue and the matchup. Arsenal remained capable of a much stronger performance in the return fixture. The opening result also supplied evidence about Everton’s ability to compete in a tight game, rather than evidence that they had become a consistently high-scoring side. A managerial improvement can matter to one prediction market without supporting every other outcome.

How to assess a team after a managerial change

Start by separating the results under the new manager from those under the previous one. After two competitive matches, a last-five form record still contains three matches from before the change. That older evidence remains useful for understanding the squad’s strengths and weaknesses, but it cannot all be presented as the new coach’s work. we exclude friendlies from its last-five and head-to-head records as a match-quality filter; when reviewing a managerial transition, also check who was in charge for the competitive performances you are interpreting.

Then look for a change that survives beyond the scoreline. United’s rising shots conceded exposed a defensive problem despite victories; Palace’s recovery included the return of a player who could alter its attack. Ask whether the team is allowing fewer dangerous chances, creating better opportunities or using available players more effectively. Where reliable xG data or footage is available, use it to examine those questions. A formation graphic alone cannot establish an improvement, and more running does not automatically mean better defending.

Finally, test the next opponent against the change you have identified. Palace’s win over ten-man Burnley required different interpretation from its later victory at Liverpool. Everton’s home win over Arsenal was not sufficient evidence for the away return. Check the venue, opponent strength and relevant absences, then decide whether the previous improvement has a credible route into this fixture. The evidence may strengthen after several matches or weaken while a winning run is still intact; the number of games since the appointment cannot make that decision for you.

How to use FootballPredictAI when a new manager takes over

Use the current fixture analysis to compare the primary pick and both alternatives, with each confidence score attached to its named market. A tighter defence may give you a different reading of the goals markets from the one you get through our Ai football analyzer. Read the reasons alongside those options and check current team news separately. The manager’s arrival provides context for evaluating the evidence; it does not justify adding an invented percentage to the displayed confidence.

At FootballPredictAI, we recommend prioritising predictions with confidence of 70% or higher, while lower-confidence predictions remain visible. A higher figure represents stronger model confidence in that specific outcome, and the highest-rated option may be an alternative to the primary pick. Keep that comparison tied to the fixture in front of you. Solskjaer, Glasner and Dyche show why the question is whether the football evidence still supports the prediction, rather than whether the manager is still inside an assumed honeymoon period.