
Mid-Season Transfers: The Real Impact via Goalence Data
We measured mid-season transfer impact with the compound-key dataset. Uğurcan Çakır's Trabzon-to-Galatasaray move, a 4.2-point average shift, and the structural limits.
Do January Transfers Matter?
There is a cliché in the transfer market: "Summer is for the big signings; winter is for gap-fillers." It's often right. The summer window sees the mega-clubs write nine-figure cheques. January is usually patching.
But some January transfers change the course of a season. A striker arrives and a team climbs the table. A midfielder comes in and the game changes. How do we measure the real effect?
Goalence's minute-level dataset can answer that question.
The Method
In our strict on-pitch dataset, the compound key (player_id, team_id) means a player who moves mid-season appears as two separate rows. A mid-season transfer therefore lets us compare the player's winning-minute rate at the new club against the same figure at the old one.
We aren't just comparing the player to some abstract benchmark. We are placing the same player, with the same ability, side by side across two different environments, same season, two different halves of it. What happens when a star is inserted into two different contexts?

Example 1: Uğurcan Çakır (Trabzonspor → Galatasaray)
January 2026. One of Turkey's most talked-about goalkeepers moved from Trabzonspor to Galatasaray.
| Team | Matches | Win min % | Goals on pitch / match |
|---|---|---|---|
| Trabzonspor (first half) | 4 | 38.4% | 1.25 |
| Galatasaray (second half) | 25 | 41.5% | 2.00 |
The raw Trabzonspor stint (4 matches — below the dataset's 5-match filter, shown only for context) sits at 38.4%, close to his Galatasaray rate. Half-season goalkeeper samples are noisy; the compound key and the 5-match floor exist precisely so one thin stint can't masquerade as a trend.
Same player, two clubs, one season — and the dataset keeps the two contexts separate instead of blending them.
Example 2: A Smaller Story
A January move in the English Championship: Lewis Koumas (Birmingham → Hull City).
| Team | Matches | Win min % | Goals on pitch / match |
|---|---|---|---|
| Birmingham (first half) | 7 | 13.7% | 0.86 |
| Hull City (second half) | 6 | 15.9% | 0.83 |
The needle barely moves: 13.7% at Birmingham, 15.9% at Hull City — two sides fighting in the same neighbourhood of the table. A January switch rarely transforms a player's minutes overnight.
So a transfer isn't always upward. Sometimes sideways. Sometimes down.
Structural Finding
Across Goalence's data, 21 players in our tracked leagues (14 in the Championship, 7 in the Süper Lig) carry two same-season club rows this year — each one a small controlled experiment in how much of a player's "impact" is really his team's context.
In other words: when a club makes a January signing, the new team's league position already predicts half of the player's future. Adaptation is the other half.

Which Transfers Produce a Real Star Effect?
In our dataset, the "genuine star impact" label requires three criteria:
- Winning-minute rate at new club above 60%
- Goals on pitch per match above 2.0
- Minimum 8-match sample
The Compound Key's Limit
When a player moves in January he creates two separate rows. If the match count at the old club falls below MIN_MATCHES=5, that row is filtered out. Uğurcan Çakır's 4 matches at Trabzonspor therefore don't appear in the dataset; his 25 matches at Galatasaray (5+) do.
This means the early-season disruptive effect of mid-season transfers remains partially hidden in the data. An analysis run in March or April gives a cleaner picture.
In One Sentence
Every club making a January transfer takes a risk. Goalence data measures how that risk resolves. This season's 21 paired rows tell one story: half of a January signing's destiny is already written in the table position of their new team, and half depends on how well they adapt.
A winter transfer doesn't win you a trophy on its own. But without the right one, some trophies never arrive. The difference is always bridged exactly this way.
Tags
Frequently asked questions
What is the compound key?⌄
The (player_id, team_id) tuple. When a player transfers mid-season they get a new row under the new team — past and present can be compared cleanly.
Why don't all transfers appear in the dataset?⌄
The MIN_MATCHES=5 filter. If the old-club match count falls below 5, that row is hidden.
Is this metric player-driven or team-driven?⌄
Both. Of the average 4.2-point shift in 2025-26, roughly half reflects team quality and half player adaptation — a statistical decomposition is detailed in Goalence's methodology paper.