Rolling in the Dip
People can come up with statistics to prove anything. 40% of all people know that.
Nobody was surprised that Liam Rosenior did not see out his six year contract. Why would he when working at a club that hasn’t seen any of a parade of illustrious managers last even three seasons in the previous two decades?
Most onlookers might have expected him to last slightly longer than three months though. If one wanted to defend Rosenior, one could point to Chelsea’s underlying data. There were signs of improvement at both ends of the field early on.
At worst, he left them where he found them after 13 league games. That’s about as much as you can hope for when promoting a young coach way above his experience level.
Aside from it representing one third of a season, a 13 game sample is an odd number to choose unless it’s the full duration of a manager’s tenure. Ten makes more sense, humans like a round number.
It works well when assessing the impact that Michael Carrick has had at Manchester United. His predecessor left after overseeing the first 20 league games of 2025/26. Have the Red Devils improved since the much-derided Ruben Amorim got the boot?
We will gain an insight into whether United are getting smarter as a club based on the decisions they make this summer. Positive vibes only last so long, they can only carry a team so far.
With the managerial changes that have occurred in the Premier League this term, as a little as rolling five-game sample size can show how certain clubs performed under certain coaches.
There are four managers who have taken charge of exactly this many matches, with Graham Potter the first of them to be dismissed. As unlikely as it looks that Nuno Espírito Santo will keep West Ham up, they eventually began moving in the right direction.
If they go down, it will be at the expense of the ninth richest club in the world managing to escape the drop. Slow hand clap for Tottenham, everyone.
Spurs are a good team to assess in this way, as Igor Tudor oversaw five league matches before Roberto De Zerbi has since done likewise. The latter has proven more effective, aided somewhat by a kinder schedule over their equally small bodies of Premier League work.
Nottingham Forest have also barely escaped relegation through a succession of relatively small managerial bursts. Nuno had three games, Ange Postecoglou five, Sean Dyche 18 with Vitor Pereira handling the final 12 (assuming Evangelos Marinakis doesn’t ditch him now the club are safe; you never know).
We can be confident that Forest’s owner prefers mood music to underlying numbers, attitude to trend lines. How ever Marinakis landed on Pereira, he may to have finally made the right choice. Eenie, meanie, minie, Vitor.
The rolling five game graph makes for interesting reading back at our first club, Chelsea. It highlights short-term ups-and-downs that are smoothed out in the 13-game chart. We more readily see the upsurge that came under Rosenior before the slump that saw him leave faster than Abe Simpson in the Maison Derrière.
It’s also worth a once over as Mr High Performance’s first five games saw the Blues earn 13 points. Nobody would have imagined he had just eight league matches remaining after a start that hot.
Five-game spells have been hugely relevant to Liverpool’s campaign too. They had a perfect record and a five point lead atop the Premier League table after five matches. The points tally hadn’t risen four games later after a run of defeats.
The Reds have also had a run of four consecutive draws, with 11 the most points they have earned in any stretch of five league games since the first five fixtures were in the bank. Had the season begun on September 25th, for argument’s sake, Liverpool would be eighth; only four points adrift of fifth, but 12 off fourth and closer to the bottom three points-wise (16) than the leaders (25).
What does their rolling five game chart look like?
It looks like the EastEnders logo if someone poured red paint into the Thames somewhere around City Airport. Will the red river sweep Arne Slot out towards his homeland?
As difficult as 2025/26 has been, you would hope to see improvements by this point, not a slump. Not that tired minds allied to tired bodies, on and off the pitch, will have made it easier to eke that out of the squad.
The graph sums up the rollercoaster of a campaign. In very broad terms, when expected goals have increased at one end they have also gone up at the other.
The Reds recently recorded their best five-game sample for underlying numbers, only to immediately follow it with their worst. Only five of the 32 overlapping five-match runs have seen Liverpool drop below an average of at least +0.4 non-penalty xG difference but only two have been above 0.7.
It’s mildly dominating games without doing enough to consistently win matches, over and over and over again. The good, the bad and the ugly frequently show up in any given game, with the whims of the referee or the bounce of the ball capable of deciding so much. It has been a season of ‘okay, but’.
For comparison, here is Manchester City. They funnelled into a mid-season blip before emerging out of it in style. Even with a red patch, the Cityzens have dominated many more of their matches than the Reds.
Arsenal have been in credit for every five-game streak, usually by a decent margin. It would be fascinating to see their figures for open play chances only. Even the Gunners haven’t entirely escaped Liverpool’s trend of seeing xG rise at both ends at the same time.
A rolling 10-game chart for Liverpool is the right way to finish. Smoother than the five-match version, it enables us to look at their unbeaten run of that length in the middle of the campaign.
Slot’s decision to tighten up the team after the trio of three-goal defeats led his side to have their best 10-game run defensively this season. The failure to sustain it has seen the xG Against line creep up towards the attack, with the team’s passive approach doing little to prevent opposition chances to the extent that it should.
This has created a bottleneck on the graph as Liverpool approach their final two matches. Maybe the lines will cross, snipping Slot off ahead of 2026/27. Never cross the streams, Arne.
In case you’ve not seen, the newsletter is ending this month. Here’s why:
