When the Data Falls Silent: Notes from a V.League Analysis Room

A spreadsheet with no rows

21:47, the last Saturday of the month. In a sports complex on the outskirts of a northern city, the analysis room is lit by a single lamp. The laptop lies open on the desk, the screen bright.

The spreadsheet has forty-seven columns and not a single row of data.

The column headers line up neatly: player name, minutes played, passes, pass completion rate, successful dribbles, ball recoveries in the final third, xG, xA, PPDA, duels won, distance covered. Forty-seven columns, each with a name, a unit, and a small definition note in the second row. And beneath all those names, blank space.

The cursor blinks in cell A3. It blinks steadily, patiently, like breathing. On the big wall screen, the match recording plays on mute. The eighty-third minute. A player in a yellow shirt receives the ball at the edge of the box, turns, and puts it into the far corner. The crowd on the recording erupts, but the sound was muted long ago.

The analyst calls the data provider for the fourth time. It rings. No answer. He puts the phone down, looks at me, and says something I will remember for a long time: "What if we just write what we saw?"

I do not answer straight away. In a room with forty-seven empty columns, that sentence sounds more like a confession than a suggestion.

I have been in this trade for forty-nine years. I have sat at eight Olympic Games, eight World Cups, written about the Giro d'Italia and the Tour de France, and once believed there was no match that pen and paper could not keep up with. Then, one night in Osaka, when I began writing a series of poems about the J-League for an online sports site, I understood something far simpler: most of what makes a football match does not live in any column.

Tonight, those forty-seven columns are silent. And that silence, to me, is a more newsworthy event than any win.

Ten years of data walking into the dressing room

Vietnamese football has come a long way in bringing numbers into the profession. Over roughly the past decade, V.League clubs began signing contracts with international data providers, hiring scouting services, buying packages of clip-by-clip footage, equipping players with heart-rate monitors and GPS vests in training. Youth academies such as PVF and the joint-venture academy in Ham Rong brought digital technical profiles into their recruitment processes early. The national teams, under several foreign coaches, grew used to receiving opposition reports dozens of pages thick before every match.

That is real progress. I do not dismiss it. A football culture that cannot measure itself finds it very hard to fix itself.

But alongside that progress, something else grew: the belief that if the data is complete enough, the answer will simply appear. That if the spreadsheet is thick enough, with enough columns and enough rows, every argument about tactics, about people, about form, can be settled with one bold number.

I have sat in many such meetings. I have seen reports presented with coloured charts, arrows showing passing directions, red-hot heat zones in both wide corridors. I have seen coaches nod, then walk out onto the pitch and do the exact opposite of what the report advised.

At first I thought that was conservatism. Later I understood it was instinct.

The winter of 2026 stretched into early 2026, and Vietnam won the Southeast Asian championship over two legs against Thailand, 5-3 on aggregate, with the first leg on 2 January and the second on 5 January. In the second leg in Bangkok, striker Nguyen Xuan Son — who had scored in both legs — suffered a serious injury and was carried off on a stretcher. That was a moment no spreadsheet predicted, and no spreadsheet records in full.

I tell that story here not out of nostalgia. I tell it because it is a clean example of what I want to say: data can describe a great deal, but it cannot describe the moment a team has to keep living after losing the thing that mattered most.

And tonight, in this room, the empty spreadsheet is telling me another story — the story of the measuring trade itself.

Forty-seven columns and what they cannot hold

Before I get to how the data pipeline went silent, I want to talk about what even a perfect pipeline cannot hold.

Take xG — expected goals. It is the most widely used metric when discussing attacking efficiency. It assigns each shot a probability of becoming a goal, based on location, angle, shot type, and the number of defenders between ball and goal. The method is sound, and it is genuinely useful.

But xG does not know that the player has run eleven kilometres in the first half. It does not know he missed a penalty in the twelfth minute and is trying to atone. It does not know he is playing his third match in seven days, on a surface the September rains have made slick as stone. It does not know the coach shouted at him at half-time. It does not know his small child is in a hospital in a city four hundred kilometres from the stadium.

All of that is part of the probability of scoring. We simply cannot measure it.

What every football metric silently assumes — and the assumption is always wrong — is that every player steps into the same passage of play in the same state.

The same shot from the edge of the box, at the same angle, under the same pressure, can carry three times the scoring probability depending on who is taking it, on which day, and in which circumstances. Data knows this at the collective level. It does not know it at the human level.

The best analysts I have met all admit this. They tell me they treat the number as the beginning of a conversation, not the end of one. The problem is that in most meeting rooms, people only hear the ending.

In my notes on V.League matches, there is one detail I always record that I have never seen in any dataset: which player fetches the ball in the ninetieth minute. It is a small act, costing no fitness, counted as no metric. But it tells me what that team still believes in.

A team that still believes still has someone fetching the ball. A team that has let go does not.

Sampling frequency: why data often gets the rhythm wrong

There is a concept in signal processing that I think Vietnamese football should discuss more: sampling frequency. If you sample a signal too sparsely, you will reconstruct its shape incorrectly. Not merely missing detail, but inventing shapes that do not exist.

Football is sampled at a low frequency. A match lasts about ninety minutes and is divided into roughly twelve hundred to fourteen hundred recordable events. Each event is a point. Between two points lies a gap in which most of the match actually happens.

We record the pass, but not the three seconds before it, when the midfielder turned his head twice to check his teammates' positions. We record the tackle, but not the half-second of slower running that made the tackle necessary. We record the team's average position, but not the fact that the back line shifted two metres to the left across the final twenty minutes of the second half.

For a team that plays slowly and controls the ball, event data reflects things fairly honestly. For a counter-attacking team, whose tempo is compressed into very short moments, event data captures only the visible tip. For a low-block defensive team, most of what they do is off-ball movement and holding shape — something the stat sheet captures only through a meaningless number: clearances.

A good defensive team will have low defensive numbers, and a bad defensive team can also have low defensive numbers. The number cannot tell the two apart, because it does not know who forced the opponent to pass toward them.

This is why opposition reports read so convincingly and prove so useless on the pitch. They describe accurately what happened. They do not describe accurately why it happened.

I have covered eight World Cups. Every time I see the same thing: before the tournament, a flood of prediction models is published. After the tournament, not one model is publicly reviewed. Nobody asks why a model that gave a team a twelve percent chance of winning was right, or why it eliminated a team from the knockout rounds. Models are allowed to be wrong without paying a price. That is a privilege no coach has.

Vietnamese diacritics and the quiet death of a data pipeline

The next morning, we found the cause. The provider had sent the data. It was in the inbox, on time, in the right file format, for the right number of matches. But when the club's system read that file and matched it against the player list in the internal database, not one row matched.

The reason lay in the diacritics.

In the provider's file, player names were written without diacritical marks, in the international standard: Nguyen Van A. In the club's database, names were written with full diacritics: Nguyễn Văn A. The matching system treated those two strings as two different people. Forty-seven columns of data, correct, complete, detailed down to the last pass, all suspended in the void between two ways of writing a single name.

I recount this technical detail not because it is amusing. I recount it because it is too perfect a metaphor to ignore.

An entire match, measured precisely down to the last pass, then vanished — because of a circumflex. An entire collective, fully recorded, turned into blank space — because the system was not designed to recognise that "Nguyễn" and "Nguyen" are the same person.

There is a technical lesson here, and it deserves to be said plainly: Vietnamese clubs are buying data from systems built for European football, where player names carry no diacritics. Every time data crosses a border, it passes through a conversion layer that very few people inspect. When that layer fails, it does not raise an error. It simply returns blank space.

But there is a larger lesson, and that is what I brought home from that night.

A system does not fail when it returns a wrong number. It fails when it returns blank space and nobody in the room recognises that the blank space is an answer.

Because throughout that evening, our first reflex was to fill the spreadsheet. We discussed calling another provider, hand-entering data from the recording, using the first-leg report as a stopgap. Not one of us stopped and asked: if the system cannot read our players' names, how many other things can it not read?

That was the right question. And it only appeared when we were willing to sit still in front of the blank space instead of rushing to cover it up.

PPDA has been decoded, and what that says about the V.League

Among those forty-seven columns is one called PPDA — the number of passes an opponent is allowed before your team commits a defensive action. Over the past fifteen years this metric has become the standard measure of pressing intensity.

I have tracked PPDA across many leagues for nearly a decade. And I believe it is the metric that has been decoded faster than any other in modern football.

When high pressing was still an advantage, a low PPDA was the mark of a strong team. Mid-table clubs looked at it and copied. They raised their intensity, increased their sprints, pushed their PPDA down. But they did not have the same quality of player in the forward line, so their pressing produced pressure without producing the ball. The result was a kind of football that turned matches into footraces, where fitness replaced ideas.

In the V.League this shows up in a particular way. The congested calendar, the long travel, and the harsh pitch and weather conditions at many points in the year make sustaining high pressing intensity for ninety minutes a fitness gamble. Teams that press hard for the first thirty minutes usually pay for it in the final twenty. And that price does not appear in the PPDA column, because PPDA is a whole-match average.

An average metric can never tell the story of a match split into two halves played at completely different tempos.

I have watched matches where the home side pressed furiously in the first half, took the lead, then conceded an equaliser in the seventy-fifth minute and lost in the eighty-ninth. The post-match stat sheet will show they ran more, contested more, shot more. The stat sheet will not show that their legs went at the sixtieth minute.

This is where the human eye still has value. When I sit in the stands, I do not measure PPDA. I watch how the back line shifts when the ball is on the opposite flank. If the distance between the two centre-backs starts to widen over about ten minutes, I know that team is about to break. No column among the forty-seven records the distance between the two centre-backs minute by minute.

Perhaps it will exist in the future. But when it does, it will be decoded too, and we will have to go looking for something else.

The blank space is the answer

I want to state plainly what I consider the central paradox of football analysis.

The trade believes its value lies in its ability to answer. Its true value lies in its ability to ask the right question.

A complete spreadsheet can answer the question "what happened". It cannot answer the question "what is happening". And in football, the second question is the only one worth asking, because the match is not over.

The empty spreadsheet that night was not a failure of data. It was a failure of the question we had asked. We asked the system how the match had gone, when we should have asked it whether the system understood our team.

There is one more thing I want to say to the young people working in this trade in Vietnam. You are being handed tools my generation never had. That is good. But keep a habit my generation was forced to keep: before publishing any number, cross-check it against at least three independent sources. Not because you will find the truth — but because in the process of cross-checking, you will discover that some numbers do not agree. And it is precisely in those disagreements that the story lives.

I was once the person who had to take the blame for a young colleague in an editorial meeting, on a World Cup night when he shouted into the microphone that we were about to make history. He was not wrong to shout. He was only wrong to believe he knew the ending in advance. And I said nothing to him at that moment, because I knew some lessons can only be learned by letting the tide go out and then looking at the sand yourself.

That night on Russian soil, I understood that the tide goes out only to hand back a sorrow.

Our job is to stand on that sand and record what is left behind.

What I brought home

The spreadsheet was eventually filled. We rewrote the player names by hand, added a diacritic check layer, and by the following afternoon every number was in its proper place. The twenty-two-page report was printed, bound, and handed to the coaching staff.

But what I remember most from that night is not in those twenty-two pages.

It is the roughly thirty seconds when the room went quiet, the cursor blinked in cell A3, and on the big screen a player in a yellow shirt turned and put the ball into the far corner in silence. Those thirty seconds were recorded nowhere. No column holds them. No report mentions them.

I think most of the work of a football writer is going back to find thirty seconds like that.

The ball rolls through my years, and I copy it down in verse.

And if anyone asks me whether to trust data or to trust my own eyes, I will answer that the question is wrongly put. Both are ways of sampling something far larger. What deserves trust lies neither in the spreadsheet nor in the stands. It lies in the distance between the two — and our job is to measure that distance, every day, with a ruler that has no unit.

Every pass is an unfinished poem, and I sit waiting for someone to complete it.

Stoppage time is a debt I owe to time, and I pay it every night.

As for that spreadsheet — forty-seven columns, no rows — I have kept a copy. Not to remind myself of a technical fault. But to remind myself that on some nights, the only honest thing a system can say is nothing at all.