Heat Maps and the Reader's Trap in Vietnamese Football's Data Culture

On the morning of March 14, the training ground on the outskirts of Hanoi recorded 19 degrees Celsius and 78 percent humidity. The session ended at 9:40. On the stone bench along the touchline lay an untouched stack of A4 sheets printed with GPS data: 10,842 metres covered, 21 accelerations above 5.5 m/s², a top speed of 31.4 km/h, 47 ball contacts in an 8-versus-8 drill. Nobody read it. By 10:15, when the squad moved to set-piece work, the sheets had been folded into four and slipped under a glass of tea. The water spread into a brown oval that covered, precisely, the column labelled “number of times the pressing line was broken”.

That is the image I have kept longest from my first working trip back to the V.League after years of reporting abroad. A sheet of data used as a coaster — a small detail, but one that tells almost the whole story of the gap between the volume of information Vietnamese football now owns and its ability to turn that information into decisions on the training pitch.

A decade of data, a decade of misreading

Over the past four seasons, the volume of data a V.League club collects each week has far outstripped the processing capacity of its own analysis department. GPS vests record distance, accelerations and heart rate. Wide-angle cameras capture every off-ball movement. Event-data providers sell clubs indicator tables so detailed they are nearly impossible to verify with the naked eye.

That sounds like good news. Yet on my two most recent trips home, I counted four clubs with their own analysis rooms, and only one of them had a person who actually sat down with the players to review video the following day. The other three sent spreadsheets through an internal messaging group at 11 p.m., with a single line attached: “Have a look, lads.”

Professional football in Southeast Asia has entered a phase in which data has become a form of jewellery. Owning it does not guarantee being better, but lacking it makes it hard to answer to sponsors. The regular season runs across many months, and every matchday pushes another mountain of charts towards the coaching staff. The real question is not how many more metrics can be gathered, but who reads them, when, and to what end.

A heat map records position, not assignment

A heat map shows only where a player stood during a match. It does not show what he was asked to do while standing there. A full-back instructed to step inside as a third midfielder will produce a footprint almost identical to a full-back told to hold the flank and advance only when the ball travels to the opposite side. Two near-identical maps. Two completely different jobs.

The heat map has become the tarot card of modern football: it answers every question with a beautiful picture and takes responsibility for none of the answers.

The consequences of misreading are concrete. A misinterpreted data sample leads to a wrong conclusion about a player's quality, and that wrong conclusion then lives inside the transfer meeting room for months, sometimes for seasons. A player labelled “poor at finding space” because his heat map clusters on one flank may simply be the most obedient man in the squad.

In October 2026, at the Workers' Stadium, I sat directly behind the technical area as Beijing Guoan lost 1–2 to Shanghai SIPG in a match the hosts dominated with 63 percent possession. What caught my attention was not the possession figure. The right-back was withdrawn in the 25th minute. Head coach Roger Schmidt stood by the touchline with his arms folded and said nothing for the following seven minutes.

Read only the statistics and you see a team controlling the game and losing to bad luck. I stayed up until 2 a.m. cross-checking three sources: the pass map, the location of duels, and the pitch temperature. The result showed the hosts' passing flow tilting heavily to the left, where the defensive corridor had been abandoned very early. The 63 percent was the visible part. The submerged part was an empty corridor — and a substitution in the 25th minute was the only remaining way for the coaching staff to plug it.

The pitch does not lie — but people do.

I wrote that line into my professional notebook after that night, and it still holds nearly a decade later.

PPDA has been decoded, and mid-table teams have learned to exploit it

PPDA — the number of opponent passes allowed per defensive action — was once a reliable proxy for pressing intensity. For several seasons, the team with the lowest PPDA was considered the highest-pressing side in the league. That reading is now obsolete.

The technical weakness is that PPDA measures defensive actions, and a defensive action can arise from very different causes: a committed tackle inside an organised structure, a tactical foul to cut off a counter-attack, or a player chasing the ball on empty lungs. All three produce the same value and carry three contradictory meanings.

I tracked a mid-table V.League club across last season, recording their PPDA round by round. Over the first ten rounds the figure sat very low, and pundits described them as the most aggressive pressing side in the division. When I sat down with the video of each match, a different picture emerged: most defensive actions occurred in midfield after short losses of possession, and very few of them led to regaining the ball within five seconds.

A low PPDA sometimes means only that a team runs more, not that it understands more. Mid-table football is turning endurance into athletics with a ball attached.

The tactical consequence is specific. When a side presses with fitness rather than with structure, it collapses between the 60th and 75th minutes, and it collapses along a repeating pattern: the midfield loses vertical spacing, the back line is pulled too high for its own pace, and the gap between the two centre-backs widens. In the V.League, where fixture density is high and pitch quality uneven, that window is a recurring point of death. Anyone who watches enough will see it before the scoreboard does.

Measuring defence by what cannot be seen

One metric I consider more useful than PPDA in the V.League context: the number of players behind the ball within the first five seconds after losing possession. This measures collective defensive awareness rather than individual effort, and it cannot be dressed up by running a few meaningless extra hundred metres.

Two teams can record identical average distance covered per match. One reorganises its shape within five seconds, the other needs twelve. In that seven-second gap, a fast counter-attack has already travelled three-quarters of the pitch. The post-match statistics will record the shot and the conceded goal. They will not record the seven seconds that were lost, nor who was responsible for them.

I once built a private tracking sheet for this metric across twelve consecutive rounds, simply by rewatching footage and using a stopwatch. The results were unkind to the sides rated highest for physicality: two of them ranked near the bottom for speed of reorganisation, despite average distance covered placing them near the top of the league.

The signal sits in the substitution minute, not in the goal

One habit I have kept since I began travelling with a squad is recording the substitution timings of both teams, round by round, across the whole season. The sheet is dry enough that a colleague once asked what I kept it for. The answer lies in this: substitution timing is the clearest trace of when a coaching staff actually saw the problem.

When a right-back is withdrawn in the 25th minute, that is a stronger signal than any goal. It says the problem had been identified beforehand, that the original match plan did not hold, and that the staff were forced to break the structure they had spent a week building. It is the most analytically rich moment in a match, and the least written about.

Three sources, or don't write

The principle I set for myself in 2026 is simple: a claim is only written when at least three independent sources point in the same direction. For match data, those three sources are the event feed, the video, and notes taken on site.

The event feed tells you who passed to whom. The video tells you why that pass was played. The on-site notes tell you whether the pitch was wet or dry, which way the wind blew, and how far the receiving player had to run before touching the ball. These three layers frequently contradict one another, and the contradiction is exactly where the answer lives.

In one match on a surface measured at 32 degrees Celsius, I recorded that the ball rolled noticeably slower than in the first half. Short passes increased, runs off the ball decreased. Using the event feed alone, you would conclude the team switched deliberately to possession football. The reality was simpler: the hot surface slowed the ball, and the players adjusted automatically without any instruction from the technical area.

That is why I always place data beside the scene rather than reading it in an air-conditioned room. Data does not lie. The reader of data is the one who can.

xG needs two columns, not one value

When I introduced expected goals — xG — to colleagues in the newsroom, I always presented it in two columns: the value the model produces, and the list of possible errors attached to it. The second column is longer than the first, and that is deliberate.

The xG model does not know who is taking the shot. It does not know the player had already run 11 kilometres before striking. It does not know his stronger foot has been sore since the previous round. It does not know the opposing goalkeeper read the direction early. A single value cannot contain those variables, and forcing it to try is a methodological error, not a technological one.

On June 27, 2026, in Kazan, Germany lost 0–2 to South Korea and were eliminated in the World Cup group stage. Kim Young-gwon scored in the 90+2nd minute and Son Heung-min sealed it in the 90+6th.

Few know that in the days before, at a training centre in Moscow, I was shown — through a connection with a German technical assistant — how the Germany staff analysed opponents with an xG model. Their charts identified a clear dead zone: the space between the two centre-backs stretched whenever the team lost the ball in the opponent's half, and high-speed counter-attacking sides exploited exactly that area. The warning was on the table. It was ignored.

I wrote that night with sweat soaking my shirt, not from emotion but because the office air conditioning had failed. The feeling was strange: a professional belief confirmed by someone else's failure. The data had been right. The decision-maker had not read it.

Since then, every metric I present comes with two columns. Colleagues call me a dry writer, producing copy that reads like a statement to an audit committee. I do not argue. But young coaches in China began asking me how to read a data sheet, and in Vietnam the number of people asking the same question has grown over the past two years. That is the only practical reward I need.

The contrarian angle: Vietnam's problem is not a shortage of data

Outside media habitually frame Southeast Asian football through a familiar template: these leagues are behind on technology, and they need more devices, more software, more foreign experts.

I think that reading puts the emphasis in the wrong place. A V.League club today holds a data volume that many European sides of the previous decade never dreamed of. The problem sits behind the data: there are too few people able to translate a table of metrics into a specific decision on the training pitch on a Tuesday afternoon.

Vietnamese football has a surplus of data and a shortage of translators. That is a training gap, not a budget gap.

This gap exists in every football nation, including the richest ones. The difference is that in major leagues a club can hire three analysts for a 25-man squad. In the V.League, an assistant coach often handles event tagging, video editing and opponent reports, all in the same afternoon.

I have lived alongside a squad to understand why they lose. And the most valuable thing I learned did not come from a spreadsheet, but from watching who opens the spreadsheet, at what hour, and for how long.

Physical detail as a witness to time

There is one thing statistics can never capture, and I always try to keep it in my writing.

That thing is the stone bench in the technical area. After a season of continuous work, the surface sinks at exactly one spot, where the head coach stood through every training session. The following season the club changes coach, and the hollow remains. No metric measures that hollow. But it describes more precisely than any chart how a collective was worn down, month after month.

An empty season, yet the stone bench still carries the hollow of a seat.

Inside a team hotel, silence is also an official statement. I learned this on long away trips, when the team dining room grew unusually quiet on the evening before a major match, and nobody on the coaching staff wanted to discuss the quiet.

They remember the goals — I remember the sigh after the whistle.

Data keeps the rhythm — emotion is the one who sings.

Signals to track through the coming stretch

For the remainder of the season, I will not follow the table in the conventional way. Three signals go into my notebook before every round: first, the timing of each side's first substitution in the opening half; second, the number of players behind the ball within five seconds of losing possession; third, the window between the 60th and 75th minutes, when sides that press with fitness begin to pay.

None of these signals appears on the league table. None appears in a match summary. But each appears before a defeat becomes a result, and that is the entire reason I stay behind after the whistle, writing numbers into a paper notebook, in a row of seats nobody looks at.

If a V.League club hires exactly one person capable of reading data and patient enough to sit through video with the players, the gap between that club and the leading group will close faster than through any single transfer. Every contract is a promise with an expiry date. A correct method of reading numbers has no expiry date — it only waits for the next person to open the notebook.