The Goalkeeper's Plant Foot: A Signal Missing From the Data Export

In 2026, when global football stopped, I took a data-coding job for a national U19 women's team. They played twelve matches all year. Across those twelve matches, the first-choice goalkeeper saved 43 percent of the penalties she faced — nearly double the average I had recorded among men's teams in the same age group. In my export file, the corresponding row held a single cell: saved. No cell explained why.

I sat with her for forty minutes in a borrowed changing room. She talked about the plant foot — the way a shooter sets the standing leg, rotates the hip, drops the shoulder before contact. It is a movement smaller than a breath, and she read it from eleven metres. I listened to a goalkeeper describe how she reads the belly step, a thing that does not exist in a data export. I did not open my laptop once.

Context

In August 2026, Chelsea triggered Kepa Arrizabalaga's release clause at Athletic Bilbao. The fee was recorded at 71.6 million pounds, the highest ever paid for a goalkeeper at the time. One month earlier, Liverpool signed Alisson Becker for 66.8 million pounds. Two deals inside six weeks, both explained with the same sentence: modern football needs a goalkeeper who can pass.

I do not object to goalkeepers being able to play with their feet. I object to the weighting. In most datasets I have processed, a goalkeeper's pass volume and pass accuracy occupy far more space than the goals they prevent — while the second category is the one that decides points. A goalkeeper who plays forty passes a match looks extremely modern on a chart. A goalkeeper standing half a metre outside the ball's trajectory does not appear on any chart at all.

For leagues with narrow financial margins, the question weighs more heavily. A mid-table club cannot afford an elite ball-playing goalkeeper, but it can always train the reading of movement — a skill with no transfer fee attached.

My working method since 2026 has barely changed: hand-code every touch, log the coordinates, then compare against aggregate metrics. It is slow. One match takes six to eight hours, sometimes longer. But spatial data only has value when you know why the ball left the foot — and that why never fills itself into a table.

Core

On 15 June 2026, in Sochi, Spain drew 3–3 with Portugal. I was then a part-time statistics assistant for a football website in Singapore, tasked with coding every situation in that match. When the sheet was finished, one line stopped me: Cristiano Ronaldo's maximum speed in the game was 9.8 km/h, below Portugal's team average maximum of 11.2 km/h.

A forward scores three goals while running slower than his teammates. The familiar explanation is that he reads the game. But reading the game is an unverifiable answer, and I did not want to publish it. I went back to the coordinate map. All five of Ronaldo's shots on target came from close-range situations, within a radius where a defender needed only half a step to intervene. And the Sochi pitch was unusually narrow, compressing the wide corridors.

My analysis of that pitch width drew more than 200,000 views and was shared by a Spanish journalist. What I learned did not come from the view count. When Arnold Schwarzenegger says “I'll be back,” he is not talking about speed. Neither is Ronaldo at 9.8 km/h. A player's value — and a goalkeeper's value — lies in position, not velocity.

Bring that logic back to the goalmouth. The U19 girl's 43 percent penalty save rate does not come from reflexes. Reaction times among elite goalkeepers differ very little, often by a few hundredths of a second — a gap the naked eye cannot separate and a camera at twenty-five frames per second cannot fully record. What creates the distance between 43 percent and the average is the ability to read the movement chain before the shot: the run-up angle, the tilt of the torso, and the plant foot.

When I cross-checked her data across twelve matches, the structure was clear. Against shooters with a wide plant foot, she committed roughly a tenth of a second later than usual and kept her torso higher. Against shooters who opened their hips early, she chose her direction before the ball left the foot. Not intuition. A repeating pattern — nine times in the eleven penalties I coded.

There are numbers that never appear on a statistical sheet; they sit between two touches. The belly step is one of them.

This is why I trust post-shot expected goals — the gap between the probability a shot scores and whether it actually did — far more than raw save percentage. Raw save percentage cannot separate a difficult stop from a shot hit straight at the keeper's standing position. Post-shot metrics can. But both stay silent on the question I care about most: where was the goalkeeper standing three seconds earlier?

In 2026, as a second-year student, I launched the blog Data Corridor with a piece on Mesut Özil's seventeen key passes in the Premier League, plus a chart showing Arsenal's xG ranking falling in matches he did not start. The first reply I received was a rhetorical question about what a girl could possibly know about football. I did not delete the post. I added three more charts, each annotated with per-match data sources. A season is built from the repetition of seventeen forgotten passes, more than from the sum of thirty-eight fixtures. That lesson applies to penalties too.

Contrarian

Here I have to stop myself. Twelve matches is a small sample. A goalkeeper saving 43 percent across eleven penalties may simply sit in the tail of a distribution, and if she saves 20 percent next season, my data was not wrong — my conclusion was. I once wrote a highly confident piece on the correlation between a goalkeeper's pass volume and a team's points, then discovered that once high-possession teams were removed, the correlation almost vanished. Correlation is not causation. I had to print that line and pin it to the wall.

The transfer market holds no such habit. It pays for what can be seen. A goalkeeper good with his feet produces attractive sequences, appears in highlight reels, and is easier to sell to an audience than a goalkeeper who stands in the right place so the ball arrives within reach. The 71.6 million pound fee for Kepa was a rational decision inside that frame of logic. The problem is that the frame was built from what cameras capture, not from what wins points.

The same mechanism operates elsewhere. Inside the VAR room, the standard of a clear and obvious error reads as very tight on paper, but the space for subjective judgment within it is far wider than viewers imagine — especially on situations where a goalkeeper leaves the line a few tenths of a second early. And at the development level, satellite club systems let large clubs rotate young talent through multiple smaller leagues to sidestep domestic training requirements. Those players grow up, get valued, get resold — and nobody ever codes their plant foot.

In a corridor, if you only look toward the light, you will miss whatever is standing in the dark.

Takeaway

Based on my experience tracking matches every week, the most valuable part of a penalty usually happens before the ball is struck. Next round, I will switch off the statistics panel and watch the shooter's standing leg before I watch the goalkeeper. Three signals I will follow: how far the goalkeeper stands off the goal-line axis at the moment the ball leaves the foot, their post-shot expected goals figure over the last ten matches, and how often they commit to a direction before the ball travels. If that 43 percent is real, it will hold through a second season. If it does not, I will write it again.