How much does a new signing improve a football team?

Less than the fee implies, in most cases. A team's output is eleven players, a system and a manager, so even an elite addition shifts one slice of it. In modelling terms a genuinely top signing in a key position might move a team's expected goals by a tenth or two per game, worth a handful of points across a season. Real, sometimes decisive, and nowhere near the transformation the headlines describe.

Do transfer fees predict how good a signing will be?

Only loosely. Fees price scarcity, age, contract length, nationality rules and the selling club's leverage, as well as ability. Two players of similar underlying output can cost wildly different amounts. A model cares about the output: minutes, chances created and prevented, goals added relative to the player being replaced. The fee is a fact about the market, not about next season.

How quickly should a rating react to a transfer?

Deliberately slowly. On announcement day a signing is a forecast, not evidence: adaptation, fit and injuries are all unknown. An honest model nudges the rating by a conservative estimate of the upgrade and then lets minutes on the pitch do the rest of the moving. The alternative, repricing a team on hype, is how supercomputer-says headlines end up wrong by Christmas.

Why do markets move on signings more than models do?

Markets price sentiment as well as substance, and a marquee signing moves both money and mood at once. A model that waits for evidence will usually move less and later. Sometimes the market is early and right; often it is early and excited. The gap between the two on transfer day is a measure of how much of the fee was story.

the upgrade

How much is a new signing worth to the model?

A £100m player does not buy £100m of goals. What a transfer honestly changes in a team's numbers: one rate, nudged a little, with a deliberate lag while the evidence arrives. Hype is not an input.

21 July 2026 · 6 min read · James Frewin

Cover illustration for How much is a new signing worth to the model?

Check it withChatGPT Perplexity Claude

Transfer season prices everything in astonishment: record fee, statement signing, a club transformed by one medical and a photo with a scarf. A model is the one voice at the party asking the boring question: how many goals does this actually add, and compared to whom? The answer is usually a small number, and the smallness is the insight.

A team is not a shopping basket

A team’s rating, in this site’s terms, is a pair of rates: goals expected for and against. Those rates are produced by eleven players, a bench, a manager and a way of playing, so any one signing owns a slice of the output, not the output. Even a striker who takes a third of his team’s goals inherits most of that third from the chances the other ten create. Swap him for someone better and you upgrade the finishing of those chances, not the existence of them.

The right comparison is never the new player against nobody; it is the new player against whoever was getting those minutes. A world-class arrival replacing a merely good incumbent is a small upgrade at a vast fee. An unfashionable arrival replacing a hole in the squad can be worth more than the marquee name. This is the argument of the £125m striker piece: the market prices the player, the model prices the difference.

Elite striker replacing an average one+0.15 goals a game

Top defender into a leaky back line+0.10 prevented

Good squad player, better depth+0.04 across a season

Marquee name, no clear starting spot≈ 0 until minutes arrive

Illustrative orders of magnitude, not measurements: what a summer change might add to a team’s per-game expected goals relative to the player replaced. The unglamorous lesson is that the replacement matters as much as the arrival.

Small numbers, real consequences

A tenth or two of a goal per game sounds like nothing and is not. Across 38 games it compounds into a handful of points, and league seasons are settled by handfuls: handfuls decide titles, Champions League places and survival. So the honest position holds two thoughts at once. No signing transforms a team, and a good signing is still one of the few levers a club controls that reliably moves the number at all. What transfer spending cannot do is compress into one August. The headline “title won in the window” assumes the fee converts to goals at face value, and it never does.

The fee measures how badly two clubs wanted the same afternoon. The upgrade is measured in tenths of a goal.

Why the model moves late on purpose

On announcement day a transfer is a forecast wearing a shirt. Fit, adaptation, language, injuries, a manager’s trust: all unknown, all capable of halving the value of a great player for a year. So the engine treats a signing the way it treats any claim: a conservative nudge now, evidence to move the rest. Team rates shift a little on arrival, then follow the minutes. This is the same discipline that governs every pre-match drift: the market can afford to trade on mood, and a model that copies the mood inherits the hangover. The gap between the market’s transfer-day move and the model’s is a live measurement of how much of the story was story.

What to ask about any signing

Skip “is he good?” and ask the model’s three questions instead. Who loses minutes, because that sets the size of the real upgrade? Which rate does he move, the goals-for or the goals-against, and by roughly how much? And how long until the evidence arrives, because until it does, every valuation is a prior. Answer those honestly and you will be ahead of most of the coverage, which is answering a different question entirely: how exciting was the photo with the scarf?

How much is a new signing worth to the model?

A team is not a shopping basket

Small numbers, real consequences

Why the model moves late on purpose

What to ask about any signing

How likely is a clean sheet, really?

Who's most likely to score, and by how much

Simulating your cup run: how hard is the draw?