When the Stat Sheet Comes Back Empty: 347 Spectators, 214 Matches, and One Rule That Cannot Be Broken

The Round 12 WK League match between Incheon Red Angels and Gyeongju KHNP drew 347 spectators. I know that number because I spent the entire first half sitting at gate three handing out backup tickets, and in those forty-five minutes I counted every person who walked through the door. On the B stand, the media area had exactly one camera, mounted on a tripod at the halfway line, roughly four metres above the pitch, its standard frame a rectangle swallowing both penalty areas. For the whole first half, it never once panned to the left flank.

In the 23rd minute, Lee Min-a scored the opening goal. On the main recording, the goal appears as a broken sequence: a long ball from midfield, a gap, then the ball in the net. No high press was captured. No diagonal run from a wide midfielder. Nobody saw Gyeongju's left back pulled eleven metres out of position before the ball reached Lee Min-a's feet. Four hundred thousand people can rewatch that clip. Three hundred and forty-seven were at the ground. Nobody in either group understood what had actually happened.

That afternoon I rigged a second small camera at a low angle, level with the touchline, about twenty metres from the left corner flag. It had no operator, just a cheap tripod and an SD card. But when I spliced the two feeds together that evening, what I saw was no longer a goal. It was a system.

The second camera is not a low starting point – it is the angle the stands have never seen.

From that low angle, I counted four runs that never received the ball in the twelve seconds before Lee Min-a shot. The right midfielder ran inside to drag the opposing full-back with her. The centre-forward dropped ten metres to open the space behind. The opposing centre-back stepped up half a metre too high. Those three movements appear on no stat sheet, because there is no event to count. No pass was made. Nobody touched the ball. And precisely for that reason, they are invisible.

A missed penalty in the 88th minute has little to do with finishing technique and everything to do with whether anyone counted how many times that player repeated that motion under pressure, across how many training sessions, in front of how many spectators. Nobody counted. So the default story becomes: she lost her nerve.

Context: a league filmed through a single lens

The WK League is South Korea's top professional women's football division, operating with eight teams for most recent seasons. Each matchday, the organiser assigns one filming crew per game, usually two people: one camera operator, one handling audio and transmission. No second camera. No overhead camera. No positional tracking data.

That is not negligence. It is a budget problem solved before kick-off. One fixed camera, operated by one person, on an afternoon with three hundred and forty-seven spectators, is a reasonable investment by every spreadsheet the organiser has. The trouble is that a correct spreadsheet produced a faulty product. It returned a truncated match, and returned it silently, with no warning attached.

A missing frame does not announce that it is missing. It simply does not exist, and viewers default to assuming that what they saw is all that happened. This is the most dangerous property of sports imagery: it has an outer boundary, and that boundary never appears on screen.

I have watched the same mechanism repeat across sports. In a women's basketball game, a main camera placed at seated eye level turns an entire zone defence into a jumble of bodies, and the viewer can never see its rotation principles. In a women's volleyball match, a camera at a diagonal angle means nobody sees where the outside hitter starts her approach, so every spike looks identical. Each time, part of the game vanishes from history, and the part that vanishes is usually the part that explains why one team beat the other.

From then on I developed a habit: before analysing anything, I check the image source. How many lenses? Where are they placed? What falls outside the frame? How many people are counting? What definition are they using? If the answer is one lens, one counter, and no recorded definition, then every accompanying stat sheet must be read as an incomplete document, not as a fact.

This is not professional cynicism. It is the minimum working condition.

The 2026 lesson: possession is the smoothest liar

On 10 July 2026, in Saint Petersburg, France met Belgium in a World Cup semi-final. The score was 1-0 to France. The goal came from a 51st-minute corner, headed in by Samuel Umtiti. Read only the post-match stat sheet and you could construct a completely different story: Belgium had more possession, Belgium completed more passes, Belgium controlled the tempo for most of the match.

But that stat sheet does not explain the match. It describes only its surface.

When I rewatched the footage, what struck me were very short windows, around eight seconds, when Belgium shifted from defence to attack. In one such sequence I counted twelve consecutive passes. But in those same sequences, I counted only three occasions when they actually moved the ball into a finishable zone. Three moments of a match produced the result, while sixty percent of possession produced only a feeling.

Possession percentage is the most deceptive metric in football, because it aggregates every sideways pass into a number that looks like authority.

A sideways pass at midfield at 0-0 is worth roughly nothing. The same pass in the 85th minute, when your team is a goal down, is worth less than nothing, because it burns time. But the stat sheet adds both together. It does not distinguish score context, field position, or intent. It just counts.

I wrote that analysis on my university blog, titled Women Need to Understand Pressing Too. It drew 126 reads in its first week. A lecturer emailed me to say he had used it as material for his tactics class. The number 126 changed nothing in the sports industry. But it taught me something I have carried through my whole career: female fans are not short on analytical ability. They are short on content written the right way.

If a piece is read by only twelve people but makes one of them change how they see the game, that piece has done its job. That is why I stopped measuring quality by read count.

The language problem: why women's sports writing goes wrong at sentence one

Before moving on to data, I need to address language, because this is where most women's sports coverage is bent out of shape from the very first line.

When a female athlete plays well, the default description is usually: she fought with everything she had, she showed fighting spirit, she never gave up. When a male athlete plays well, the default description is usually: he read the pass, he held the right distance, he chose the right moment.

One side receives adjectives of emotion. The other receives verbs of action.

The difference sounds small. It is not small at all. When an entire women's football match is written in adjectives, the reader accumulates no tactical knowledge by the end. They accumulate only feelings. And feelings cannot transfer from one match to the next.

Instead of writing that the defender fought selflessly, write that the defender dropped five metres and cut off the diagonal pass into the box. Instead of writing that the midfielder played with great effort, write that the midfielder executed twelve attacking switches in the second half, and ten of them went to the right flank.

Written that way, the reader starts seeing the match structurally. And once the reader sees structurally, they will spot for themselves what went wrong, instead of waiting for someone to tell them who the hero was.

This is the entire reason for writing with annotated frames, describing player positions, gaps and directions of movement. It turns a report into a teaching tool.

214 matches, 214 problems

In 2026, leagues around the world stopped. I had just finished my master's and had a stretch of time I had never had before. I decided to do something nobody asked for: rebuild the complete dataset of the South Korean women's national team from 2026 to 2026, from video.

Two hundred and fourteen matches in total. I rewatched each one, counted each goal, classified each goal by how it was created: open play, direct set piece, indirect set piece, counter-attack, individual error, second-phase set piece. The result left me sitting quietly in front of the screen for a long while.

The South Korean women's national team scored 23.7 percent of its goals from set pieces. Japan, over the same period, scored 41.2 percent. That gap of nearly twenty percentage points had nothing to do with fitness, nothing to do with individual technique, and nothing to do with speed. It sat in one very specific place: the number of training hours spent on dead-ball situations, and the quality of the person delivering the ball.

214 matches, 214 problems: the pandemic did not stop football, it only changed how we read the game.

I wrote the report, sent it to the women's national team head coach, and waited. Three weeks later I received a reply email inviting me to collaborate on opponent analysis during the October camp. A small, pragmatic act, carrying no grand appeal. It opened an entirely new direction.

The lesson was not that data has power. That is a platitude. The real lesson is that data only has power when it is packaged as a question someone else needs answered. Twenty-three point seven percent means nothing on its own. It means something only next to forty-one point two percent, and only when the person reading it is preparing for a qualifier.

How to build a small dataset without lying

I get a lot of questions from young people who want to do sports analysis, and the most repeated one is: how do I start when I have no access to professional data.

My answer is always the same: stop trying to do everything. Pick a single event type, define it tightly, then count it across twenty matches.

For example, choose: the number of times a team wins the ball back within seven seconds of losing it in the opponent's half. The definition needs three elements: a maximum time, a geographical zone on the pitch, and a way to determine the moment the ball was lost. Miss any one of the three and your number cannot be compared to anyone else's, including your own next season.

After twenty matches, you have earned the right to say one sentence about a trend. And when you say it, you must include the definition. Without a definition, your number is just a number.

There are three mistakes I see most often among newcomers.

The first is concluding from three matches. Three matches is an accident, not a trend. In football, natural variation between matches is large enough that three matches rarely say anything.

The second is choosing the definition after you already have the data. If you count first and decide what counts afterwards, you are not analysing, you are decorating. The definition must be written before you start watching.

The third is omitting the denominator. A team scoring four goals from corners sounds impressive. But if that team took a hundred and twenty corners that season, they are at an average rate. Every number about goals must travel with the number of chances that produced them.

These three mistakes are not technical errors. They are professional ethics errors wearing technical clothing.

Tokyo 2026: seventeen and three

In 2026, I was a new hire at a sports broadcaster, assigned to women's Olympic football for the Tokyo Games. The match I was given was the group-stage fixture between Great Britain and Japan. I rewatched the footage that evening and did what I always do: counted myself.

In the second half, I counted seventeen fast counter-attacks by Great Britain. The official stat sheet published after the match recorded three. A gap of fourteen.

That discrepancy did not come from one side counting wrong. It came from two sides counting two different things under the same label. The official sheet used a narrow definition of dangerous chance: only situations ending in a shot or a pass into the penalty area were counted. My other fourteen were situations that moved the ball from their own half into the opponent's half within ten seconds, through at most three passes, but without ending in a shot. They were not counted, even though they were precisely what forced Japan's back line to keep retreating and never hold its shape.

A number is not an event, it is a definition that has been counted.

Definitions are written by people. The people writing them work under time pressure. In a women's football match, the coder usually sits alone, with one keyboard and one screen, and must publish the sheet within thirty minutes of the final whistle. Under those conditions, the human brain can only register events with clear boundaries: a shot, a corner, a yellow card. Events without clear boundaries, such as a phase that stretches the shape, a defender pulled out of position, a run that never receives the ball, disappear from the sheet, and therefore disappear from the story.

I wrote a rebuttal on how the media defines dangerous chances arbitrarily. It was shared by a K League coach as reference material for his own students. I mention this not to boast. I mention it because it established a working principle: from that day on, I no longer trust any pre-made stat sheet I have not verified myself.

An empty sheet is not a clean record

This is the part I want to dwell on most, because it is the least discussed.

In sports analysis there is a trap on the opposite side from the one I just described. Not the trap of the wrong number. The trap of the number that does not exist.

A dataset returning twenty blank rows does not mean there is no problem. It means there is no data. Those two things are entirely different, and confusing them is the most serious error an analyst can make.

Imagine a security camera that failed on the night of a burglary. Nobody says that because the camera recorded nothing, no burglary occurred. Nobody says that because the patient has not yet taken the test, the result is negative. But in sport, when the sheet is empty, people keep writing anyway.

And they write with great confidence, because the tools to write are already there: a professional template, a vocabulary set, a five-part outline. Professional formatting has one dangerous property: it confers authority before content exists. A well-presented document with tables and bold headings will be read as a document with a basis, even when there is not a single fact inside it.

I call this a void dressed as a conclusion.

It shows up everywhere. A team has not played a single match under its new coach, and people have already finished writing about the new philosophy. A patch has not been played in any tournament, and people have already built the new meta. A player has no minutes, and people already have a form assessment.

In esports the mechanism is even clearer, because change cycles are far faster. After every major update, the entire existing dataset becomes void within days. That is the window when teams have not played any official match on the new version, when win-rate samples are too small, and when every conclusion is standing on air. It is precisely in that window that the volume of analytical content spikes, because viewer demand does not wait for data.

A decent professional must be able to say this sentence: I do not currently have enough data to conclude. It is a hard sentence to say, because it sounds like an admission of weakness. But it is the most accurate sentence in a great many situations.

Esports is not a young generation's discipline, it is a discipline for those willing to read the meta before stepping onto the stage. And reading the meta, in the first two weeks of a patch, means reading very little and taking a lot of notes.

A good broadcaster is not someone who talks a lot, but someone who knows how to let the data speak at the right moment.

When the World Cup paused and the whole world held its breath, I learned that silence is also a news item.

Commercial value and competitive value

There is a theme I return to often in my writing, and it connects directly to the data story.

In women's football, money and competitive quality do not travel in a straight line. A club can sign the biggest sponsorship deal in the league, have the largest social following, sell the most shirts, and still lose to a small team through a single corner routine rehearsed two hundred times across a season.

I am not saying money does not matter. I am saying money is allocated by the logic of brand, while points are allocated by the logic of set pieces.

The most expensive transfer deal is not written in the contract, it is written in the gap the player leaves behind.

The same mechanism runs on both sides. When a big club spends to buy a famous attacking player, it is paying for the ability to generate shareable moments. When a small club signs a defensive midfielder nobody reports on, it is paying for the ability to erase moments. The value of the second player never appears on a stat sheet, because the sheet measures what happened, not what failed to happen.

That is why small clubs understand better than anyone that the transfer market does not price players by actual contribution. It prices them by how memorable they are.

In women's football the problem multiplies, because publicly available data is far thinner, meaning a female player's memorability depends on whether someone happened to record her moment. A female defender who plays thirty error-free matches will have no viral video. A female forward who scores one beautiful goal will.

If you want to understand why national team squads in many countries regularly have inexplicable gaps in defence, start here. Not because there are no good defenders. But because good defenders do not generate video.

Method: count three times before writing one sentence

This is the process I use for every post-match analysis.

First pass: I watch at normal speed, taking no notes. The only purpose is to let the first impression form, so I know which way I will be biased. I always write that impression down in one sentence, because that impression is what I will have to check at the final step.

Second pass: I watch again at reduced speed and count. I count definable events: how many times a team built from the right flank, how many times the defensive line pushed past the halfway line, how many seconds between losing the ball and winning it back.

Third pass: I watch again at normal speed, but this time only to check whether the numbers I counted actually correspond to what I felt. If they conflict, I trust the numbers, and I record the conflict as part of the piece.

I do not trust emotion, I trust data. Emotion can lie, a stat sheet cannot.

But that sentence is only half true, and the other half matters just as much: a stat sheet can lie in a different way. It does not invent facts, but it omits them. A sheet that is honest about what it measures can still lead a reader to a wrong conclusion, if the writer does not state clearly what was left out.

That is why in every piece I try to include at least one passage answering the reverse question: what does this data not include.

Football is not remembered only by goals, but by the forgotten minutes of extra time. Those minutes are not in the sheet. But they are in the footage, if you are willing to rewatch.

What is changing

I began my career as an esports athlete and tournament organiser, then moved into media. That path gave me an odd advantage: I know how a match is produced before I know how it is written. I know that behind every published stat sheet there is someone sitting at a screen at eleven at night, tired, deciding whether that last phase counts as a chance.

Knowing that, you read every sheet differently. You stop asking whether the number is right or wrong. You ask: who counted, under what definition, under what conditions, and what did they skip to meet the deadline.

Eight years after the match with three hundred and forty-seven spectators, the number of cameras at Korean women's football matches has increased. Not as much as people hoped, but enough that a high press no longer vanishes from history. The number of dedicated data staff for women's competitions has risen too. And the number of women writing about women's sport, the most important figure of all, has risen as well.

Every time another woman sits down at an editing desk, another frame gets noticed. Not because women see things men do not. But because when more people look at the same match, the number of overlooked angles goes down.

That is my entire argument, and it needs nothing more.

At the same time, a new pressure has appeared: the pressure to speak faster, to have content before the match ends, to have numbers before the coder presses the final key. That pressure produces a new category of content: analysis without data, wearing the format of analysis with data.

That is what I try not to write, and also what I want readers to recognise so they can protect themselves.

If a piece does not tell you where its data comes from, read it as an opinion. If a piece contains not one specific frame, read it as a feeling. And if someone hands you an empty stat sheet with the expression of a person who has just discovered the truth, ask one question only: in that sheet, which row is the blank one?

The answer to that question is usually the match.

I still keep that small camera. It no longer works; the tripod broke long ago. But I keep it on the shelf, right beside my monitor, as a reminder that every time a frame disappears, a piece of the truth disappears with it, and the only person who can record that piece is the one who was there.

Three hundred and forty-seven people were there. One of them brought a cheap tripod. Thanks to that, Lee Min-a's 23rd-minute goal is no longer a blank on a recording, but a system that can be read, counted, and taught to others.

That is the whole job. There is nothing glamorous in it. Just a person who stays behind after everyone has gone home, rewinds the tape a third time, and counts.