Fulham vs Chelsea: how much should a new manager move a match probability?
Published
Monday night at Craven Cottage closes the Premier League’s opening matchday with two managers taking charge of a league game for the first time, and it forces a question previews normally answer with adjectives instead of arithmetic: how much should a new manager move a match probability? Our model’s 22 August run makes this 38.2% Fulham, 27.5% draw, 34.3% Chelsea — the current version is on the match page — and the betting market has the visitors odds-on. Those two readings cannot both be close to right, so it is worth being exact about how far apart they are.
Seventeen points, not “a bit of a gap”
Football365’s preview, read on 23 August 2026, lists Chelsea as 17/20 odds-on favourites, Fulham at 27/10 and the draw at 3/1. Converted to decimals that is 1.85, 3.70 and 4.00, whose implied probabilities sum to 106.1% — the bookmaker’s margin. Strip it out proportionally and that market is saying roughly 25.5% Fulham, 23.6% draw, 51.0% Chelsea.
Line the two up:
| Our model (22 Aug run) | Market, margin removed | |
|---|---|---|
| Fulham | 38.2% | 25.5% |
| Draw | 27.5% | 23.6% |
| Chelsea | 34.3% | 51.0% |
That is 16.7 percentage points of disagreement on the away win and 12.7 on the home win. For context, the entire effect of home advantage in a typical Premier League fixture is worth less than that. This is not a rounding difference; one of the two has the fixture materially wrong.
Our ratings will not separate these teams
The model’s attack and defence parameters, updated 23 August, rank Chelsea 10th of 20 for attack and 11th for defence and Fulham 14th and 18th. On those components Chelsea are the better side — modestly. Our separate Elo summary, which weights recent results differently, goes the other way and has Fulham on 1628, 10th in the division, against Chelsea on 1568, 14th. Two of our own ratings pointing in opposite directions is itself the finding: there is no version of last season’s evidence in which these clubs are far apart.
Add Craven Cottage to a component edge that small and it flips. That is the whole of the 38.2%. It is not a claim that Fulham are the better team; it is a claim that the better team’s edge here is smaller than a home ground.
How much should a new manager move a match probability?
Here is the part the market is presumably pricing and our model structurally cannot. Álvaro Arbeloa makes his managerial debut for Fulham; Xabi Alonso takes charge of Chelsea for the first time, the two having played 46 league games alongside each other at Liverpool, per the same Football365 preview. Our ratings were computed from matches neither man was involved in.
So: how big would the manager effect have to be to justify the market’s number? Approximate the published split with an independent-Poisson model — deliberately crude back-of-envelope working, not the model’s internals — and it sits at roughly 1.40 expected goals for Fulham and 1.32 for Chelsea. Now raise Chelsea’s scoring rate and see what it buys:
- Chelsea’s rate +10%: Chelsea 34.3% → 38.5%
- +20%: → 41.8%
- +30%: → 44.9%
- +50%: → 50.9%
Only at a 50% uplift in expected goals does the model reach the market’s 51%. Allow that Arbeloa’s arrival also costs Fulham 10% of theirs, and Chelsea still need roughly +40% to get there. A single change in the dugout worth forty per cent of a team’s attacking output, before a ball has been kicked under either manager, is a very large claim to make quietly inside a price.
The honest reading is that the market is not pricing the managers alone. It is pricing a squad — five new signings expected to start for Chelsea, including Maxence Lacroix, Morgan Rogers and Marco Palestra alongside Cole Palmer and João Pedro, again per Football365. Transfers are exactly the input our model has no channel for: it learns a club changed only when the new squad plays and the results move the ratings.
The absences on both sides, and what they are worth
Fulham are without Joachim Andersen, serving a suspension, and captain Tom Cairney, injured. Chelsea are missing Wesley Fofana, banned, Jordan Henderson, still recovering from a wrist injury sustained after England’s World Cup last-16 win over Mexico in July, and Emmanuel Emegha and Aaron Anselmino with hamstring problems; Danny Welbeck and Mamadou Sarr are expected to be available. All of that is from the Football365 preview and none of it is in our numbers.
Two defenders out on one side and a centre-back plus two forwards on the other roughly cancel in direction. What they do not do is close a seventeen-point gap — as we worked through when putting a number on Jeremy Doku’s absence, one missing player is worth low single digits, not double.
What we do with a disagreement this size
Nothing. We publish 38.2/27.5/34.3, date it to the run that produced it, and let it be graded like every other call in the public track record — including the ones where the market was obviously better informed and we said so at the time.
There is a version of this post that quietly nudges the number toward the price and looks cleverer on Tuesday morning. It would also make every number on this site unfalsifiable. The rest of the round is on the Premier League hub, and Fulham’s team page carries the ratings as results start to move them — which, for a squad and a staff this new, will not take long.