When the Spreadsheet Is Empty: The Art of Humility in Football Data Analysis

In July 2026, I sat in front of a computer screen with an empty spreadsheet. The club I had been working for as a part-time data consultant had just been dissolved. The national league I was tracking had been suspended indefinitely. In my entire database, the column labelled 'matches played' displayed a single value: zero. No xG. No PPDA. No passes, no touches, no metric of any kind to analyse. I sat there for a long time, staring at the blank cells stretching across the monitor, the hum of the laptop fan filling the room.

In football data analysis, there is a truth few are willing to admit: sometimes, the most honest thing data can tell you is that it does not know. That was the hardest lesson of my thirteen years in this industry.

When the league returned late that year, I volunteered to run performance analysis for a national women's U19 side. Across the entire year, they played twelve matches. Twelve. Less than a third of a Premier League club's fixtures in the same period. Yet inside that data scarcity, I found one of the most important discoveries of my life: their goalkeeper saved 43 percent of penalties, not through reflexes, but through reading the run-up of the taker. A signal that exists in no data export.

I once heard a goalkeeper describe how she read the belly of a run-up, something that never appears in an export file.

When data became the new religion

Over the past decade, football has undergone an irreversible data revolution. European clubs spend millions of euros a year on analytics systems, from Stats Perform's Opta to StatsBomb, with hundreds of metrics encoded for every second of every match. Brendan Rodgers once said he spends more time with data analysts than with his coaching staff. Liverpool under FSG turned data into a core competitive advantage to recruit Mohamed Salah, Sadio Mané, and Roberto Firmino, none of whom were the brightest names on the market at the time. Brighton & Hove Albion, on a budget modest against the giants, built a data-driven scouting model so effective that they became a template for a generation of clubs trying to compete with brains rather than cash.

At the same time, a paradox emerged. The more data there is, the rarer the ability to read it correctly becomes. The rise of xG (Expected Goals) has created a generation of fans who believe a single number can tell the whole story of a match. But xG only measures chance quality based on location and situation. It does not measure psychological pressure, it does not measure a centre-back playing through injury, and it certainly does not measure a midfielder covering for a team-mate who has lost form.

During this regular season, I have followed countless debates on Southeast Asian football forums. What I notice is that people cite metrics more than ever, yet question the context that produced them less than ever. A low PPDA (Passes allowed Per Defensive Action) means high pressing. But if your opponent deliberately plays long, your PPDA will look artificially low without reflecting true pressing intensity.

This is the first trap every data analyst must face: confusing data with truth. Data is raw material. Truth is the product of a refining process, and that process is never perfect.

The trap of misread metrics

The story of Cristiano Ronaldo at the 2026 World Cup in Russia is a classic example. In the 3-3 draw between Spain and Portugal, Ronaldo scored a hat-trick for the ages. But when I analysed the positional data of that match as a part-time statistics assistant for a Singapore football site, I found something strange: Ronaldo's top speed in the match was only 9.8 km/h, notably below Portugal's team average of 11.2 km/h.

Read that number alone, and you conclude Ronaldo underperformed. But place the data on a spatial map and an entirely different story appears. All five of his shots on target came from situations close to goal inside the sixteen-metre box, where sprint speed is no longer the decisive factor. What mattered was his ability to choose position, read the defender's movement, and release the shot at the exact moment.

When Arnold Schwarzenegger said he would be back, he was not talking about speed. Neither was Ronaldo at 9.8 km/h.

This is the biggest blind spot in modern football analysis: we measure what is easy to measure, then draw conclusions about what is hard to measure. We measure running speed because GPS can record it. We measure pass counts because someone can tally them. But we cannot measure positional intelligence, because it lives in the gap between two touches, and no sensor is attached to that gap.

The case of Mesut Özil is similar. In 2026, as a second-year student who had just launched the blog Data Corridor, I wrote an analysis of his seventeen key passes in the Premier League. The piece showed that Arsenal's xG ranking fell sharply in matches Özil did not start. The reaction on a major football forum was fierce, and one of the most upvoted comments read: what does a girl know about football. I did not delete the piece. I added three more charts, with match-by-match data footnotes.

A season is not the sum of thirty-eight matches. It is the repetition of seventeen forgotten passes.

The hidden signals the spreadsheet misses

What troubles me most in football data work is how transfer decisions get made on flashy metrics rather than quiet contributions. Every summer, strikers with twenty goals are valued sky-high, while centre-backs who concede only twice a season are rated far below their true worth. In a corridor, if you only look towards the light, you will miss what stands in the shadows.

Modern football data can track every step of every player across ninety minutes. But tracking capability is not the same as understanding. A defender may be recorded with a low tackle rate, simply because he reads the game so well that he never needs to tackle. A midfielder may have few assists, simply because his team-mates cannot convert his passes.

I once spent three weeks analysing data from a lower-tier Southeast Asian league where data collection was rudimentary. No player-tracking cameras. No chip in the ball. Just a single camera on the main stand and a manual note-taker. Under those conditions, I learned to read a match without any advanced metrics. I learned that a defensive midfielder's average running distance says nothing about how many times he saved his team by closing the right gap.

There is another issue I want to give more space to: goalkeeping distribution is being sanctified. Top clubs pay enormous sums for goalkeepers with beautiful distribution metrics and high long-pass accuracy. But when you look at these goalkeepers' data across multiple seasons, a different pattern emerges. Many goalkeepers valued highly for their distribution have basic shot-stopping metrics in decline. The goals they concede do not appear in the flashy metrics clubs use to price them. I have seen this in at least three major transfers in the past five years, where a goalkeeper was signed for high distribution metrics but had a save rate on long-range shots well below the league average.

This is why I always dedicate at least a third of every analysis to older players or under-covered leagues. Not because I want to appear different, but because that is where the real signals live. Where the empty spreadsheet holds the most stories.

The VAR paradox and the grey zone of judgement

In this context, the story of VAR (Video Assistant Referee) in major leagues is a perfect example of how data can create an illusion of objectivity. When VAR was introduced, fans believed controversial decisions would fall thanks to technology. But across many seasons of watching, I have found the opposite: VAR does not remove subjectivity, it relocates subjectivity into a closed room with a monitor.

The 'clear and obvious error' clause that VAR officials must follow is one of the most ambiguous provisions in the laws of football. How do you determine that an error is clear? Clear to whom? Obvious to what degree? In a match decided by speed, a collision can look clearly like a foul when replayed in slow motion, yet be entirely legal at real speed.

The space for subjective judgement in VAR is far larger than people think. And this is what data cannot resolve, because the very definition of a foul is a social convention, not a physical fact. You cannot digitise a judgement.

The same applies to semi-automated offside. In a Premier League match last season, a goal was disallowed because a striker's toe was a few millimetres ahead of the opposing defender. In data terms, the decision was absolutely precise. But in the spirit of football, should a distance the human eye cannot distinguish decide the fate of a match? This is a question no algorithm can answer, because it is not a question of precision, but a question of value.

The counter-intuitive view: when empty data is more honest than full data

There is a paradox few football data analysts dare admit: an empty dataset, honest about its emptiness, is worth more than a dataset filled with numbers born from assumption. When I sat in front of that empty spreadsheet in July 2026, the worst thing I could have done was invent a story to fill the void.

I have seen this happen far too many times in the industry. A club needs a report on a player for whom it lacks sufficient data. A young analyst, under pressure to deliver, uses data from a similar league, adjusts coefficients, and draws a conclusion. The result is a report that looks formally perfect yet contains biases disguised as precision. In one football database I worked with, there was an internal notes column that displayed a single line for a few players: insufficient information. That was the most honest line in the entire file. And I believe every football data analyst should have such a column in their toolkit. The ability to say I do not know is the most important professional skill, not a weakness.

Clubs dissolve, football stops. But data never stops telling stories.

Once, I analysed a national women's U19 side with twelve matches across the year. A data sample so small that any statistician would say no conclusion could be drawn. But precisely within that limitation, I was forced to look at each individual situation rather than run general models. And I found the goalkeeper with a 43 percent penalty save rate through reading the run-up of the taker. A metric that exists in no data export, because no one encodes the run-up as a variable. That team's coach told me something I have carried through my career: you see what men do not see. I do not think it was a comment about gender. I think it was a comment about reading signals that standard data systems were never designed to record.

Bias does not score goals, but a chart can.

The problem of satellite clubs and youth development

Another aspect of football data I care deeply about is how data systems can be used to conceal structural problems. The satellite-club system many European giants are building is one example. In theory, satellite clubs help develop young talent and create playing opportunities for young players. In practice, they allow giants to circumvent homegrown regulations.

A young talent in a smaller league can be signed and placed into a satellite system, where he is recorded as a product of the parent club in data terms, yet in reality is never trained there. Talents from small leagues become satellite assets - numbers on a balance sheet rather than people on a pitch. This is a form of data camouflage I call valid but unethical data. The numbers are all accurate. The rules are all followed. But the spirit of football - where a child in a small league can be properly developed - is damaged. When analysing football data, questioning where a number comes from matters no less than reading the number itself.

Looking forward: the signals of the next cycle

As this regular season enters its decisive phase, I am tracking a trend I believe will shape how we analyse football in the coming years: the shift from event data to tracking and spatial data. Europe's leading clubs have begun using machine-learning models to predict not only which players will succeed, but which players will succeed in a specific tactical system.

But I am also tracking a quieter, less noticed counter-trend. Football data specialists in Southeast Asia, where data collection is still young, are developing their own methods for reading matches without expensive technology. They use handheld cameras, manual analysis, and compensate for the lack of tools with powers of observation.

I believe the most notable signal of the next cycle will not come from the clubs with the biggest data budgets. The early warning will come from where the spreadsheet is empty, where analysts are forced to learn to see what is not recorded. There are numbers that do not appear on the spreadsheet; they live between two touches. And when the ball stops rolling, I still hear the sound of data falling.