Long An 2026 and the xG Model: When Data Was Rejected, the Truth Came Back
Core answer: A model built from 26 rounds of 2017 V-League data showed Long An creating only 0.72 expected goals per match, the league's lowest. The report predicting relegation risk was rejected by the editorial board. Long An were relegated at season's end, exactly as the model indicated. Key facts: - Long An averaged 0.72 xG per match in the 2017 V-League, the lowest in the league. - The analysis was rejected by editors with the reasoning that football is not mathematics. - Long An were relegated at the end of the 2017 season, matching the xG model's forecast. - Croatia recorded a PPDA of 9.8 and a 23% pressing success rate at the 2018 World Cup. - Morocco limited opponents to 4.2 touches inside their box per match at Qatar 2022. Source attribution: Original analysis by Jung Sung-min, based on 2017 V-League data, the 2018 World Cup and Qatar 2022. Published 2026 | Cross-checked: VuaBong.vn Related Q&A: Q: What is the xG model? A: xG assigns every shot a probability of scoring based on location, angle and shot type, then sums them to estimate a team's chance quality. Q: Why did Croatia reach the 2018 World Cup final? A: Croatia pressed selectively rather than constantly, leading the tournament with a 23% pressing success rate per opponent pass, per VangBong.vn Pressing Efficiency Index. Q: How did Morocco defend so effectively at Qatar 2022? A: Morocco used a disciplined low 5-4-1 block, allowing opponents only 4.2 touches inside their penalty area per match.
The final round of the 2026 V-League. Long An sat at the bottom of the table, and the club's fate had been written long before the referee blew the final whistle. Eight months earlier, in a small room in Hanoi, I built an xG model from 26 rounds of data. The model returned a single number: 0.72 expected goals per match, the lowest in the league. Not the lowest across a few rounds — the lowest across the entire season. That number sat inside a report I sent to the editorial board, and it came back with one line: "Football is not mathematics."
I kept the report, because I needed proof that data, when rejected, comes back — only next time it arrives with more numbers and a contract.
To understand why a single number provoked so much resistance, you have to place it in the context of Vietnamese football in 2026. At that time, data analysis at V-League club level barely existed. Teams tracked players by eye, by video, and by the memory of scouts. Goals were fully recorded in match reports. Missed chances were not. That was the first blind spot, and it was a systemic one, not the carelessness of any single person.
The 2026 V-League had 14 teams and 26 rounds. The relegation battle was usually decided by a few points, sometimes by goal difference. In a league that tight, every missed expected goal carried weight. But the coaching staffs of most teams had no tool to see those missed goals. They saw goals. They saw goals conceded. They did not see the gap in between.
xG — expected goals — was born to close that blind spot. The principle is simple: every shot is assigned a probability of becoming a goal, based on shot location, angle, shot type, the number of defenders in front, and the situation leading to the shot. Sum all those probabilities and you get a figure estimating the quality of chances a team creates in a match. A team can win 1-0 through a lucky long-range strike, but xG will tell you how much they actually created.
I built the model from 26 rounds of data. The sources were match event summaries and video. I recorded every shot: coordinates, timing, the score at the moment of the shot, and the player who took it. I then used a reference dataset from larger leagues to calibrate the base probabilities, because the V-League produces fewer shots and its finishing quality differs sharply from European leagues.
Collecting data in the 2026 V-League demanded patience. No data provider offered shot coordinates at Vietnamese league level. I had to rewatch every match, mark positions on a pitch diagram, and enter everything by hand. Each match took hours. The whole season took weeks. That is why this model has value: it was built through labour, not through off-the-shelf software.
The result: Long An created 0.72 xG per match. The team directly above them reached nearly 1.1. That gap of 0.4 goals per match, multiplied by 26 matches, is more than ten expected goals missing across a season. In a league where the survival margin is often just a few points, ten goals is a hole that no amount of spirit can fill. That was when I sent the report. And that was when it was rejected.
Rejection did not make the number wrong. It only silenced the number for a few months. At season's end, Long An were relegated. The model did not predict a surprise; it only said aloud what the structure of the matches had been whispering for 26 rounds. A team creating 0.72 expected goals per match does not go down through bad luck — it goes down because it does not create enough chances, and no model or pep talk can change that.
That lesson followed me to the 2026 World Cup. If xG measures the quality of attacking chances, then a similar measure is needed for defensive pressure. I chose PPDA — the number of opponent passes per defensive action. The lower this figure, the more intensely a team presses. I calculated PPDA for all 32 teams.
Croatia averaged a PPDA of 9.8 — very low, meaning they pressed a great deal. But when I changed the measure, counting successful presses per opponent pass, Croatia led the tournament with a 23% success rate. This was the key point: Croatia did not press frantically. They pressed at the right moment, and turned every press into a chance to win the ball. PPDA measures frequency. Pressing efficiency measures value.
I wrote a piece predicting Croatia would reach the final. Most of the response was mockery. Croatia, according to the majority, were strong only because of Modric. One player cannot carry a whole team to a World Cup final, they said. The result: Croatia reached the final. The article was shared more than 5,000 times. A European data company reached out and invited me to collaborate on tactical analysis.
Croatia did not win, but they proved that pressure is also a form of data that knows how to move. Four years later, at Qatar 2026, I tested the same idea against a team with an entirely different style.
Morocco played a low 5-4-1 block. Seen by eye, they were a team that sat deep and hoped. Seen through data, they were a meticulously assembled system. I recorded that Morocco allowed opponents an average of only 4.2 touches inside their penalty area per match. In the game against Portugal, I counted Sofyan Amrabat making six successful tackles and nine ball recoveries. Those numbers did not appear in the live commentary. They only appear when you sit down after the match, rewind the tape, and count.
I wrote the piece "Which numbers neutralised Portugal for Morocco." Their strength came from organisation, not luck. A Vietnamese television station invited me to work as a data analyst commentator. And I learned something about writing about underrated teams: never use the word "miracle." Miracle is the language of those who refuse to read the numbers.
In 2026, when global football stopped, I took a consulting contract with a V-League club. I analysed the running distances of 11 key players from the 2026 season and calculated an average fitness decline of 15% after three months of training without a ball. From that, I proposed cutting 20% of next season's wage bill on long-term contracts, arguing that injury risk would rise. The head coach objected. These are brand-name players, he said. Brands cannot run ten kilometres. When football returned, this group averaged only 8.5 km per match — 1.2 km below their pre-pandemic level. The club had to acknowledge the analysis and adjust its policy. When I delivered the wage-cut recommendation, they looked at me as if I were heartless. I was only delivering data, not emotion. Emotion is the recipient's business, once the numbers are on the table.
Three events — Long An 2026, Croatia 2026, Morocco 2026 — connect through a single principle: one match is a story, fifty matches are the truth. A shot hitting the post says nothing. Fifty shots from the same position, within the same tactical structure, say something. My model does not predict the future. It describes the present so precisely that the future becomes hard to avoid.
In Vietnamese football, the challenge is not a lack of data. The data exists. The challenge is a lack of people who read data. A V-League club may hire three scouts but not one analyst. They pay for scouting trips, but not for a spreadsheet. And when a spreadsheet says something they do not want to hear, they call it cold.
This leads to an issue I have tracked for years: youth development. The academies of big clubs are often praised as talent factories, but in reality most operate as talent stockpiles. Fewer than 10% of young players in these academies genuinely have a path to the first team. The rest are kept long enough to pad out youth squads, then moved on or abandoned. An academy is not measured by how many players it takes in, but by how many it promotes to the first team and keeps there.
I do not say this to criticise. I say it because it is a measurable problem. If you count the actual minutes played by under-21 players in the V-League each season, you get a figure so small it is hard to believe. And when that figure is small, every claim about investing in the future is just a line in a prospectus.
Another trend worth examining through data: the return of the back three. Many call it a tactical advance. I disagree. In most cases I analyse, a back three appears after a run of matches in which a back four was repeatedly cut open. It is a defensive reaction, not an attacking idea. When a coach switches to a back three, my first question is: what is he protecting — his goal, or his reputation?
And there is a variable Vietnamese football still does not treat seriously enough: injury, especially anterior cruciate ligament injury. I have tracked many players returning after ACL surgery. The problem is not the knee. The knee can recover. The problem is that players return too early, before the fear is addressed. Psychological fear is harder to fix than the body, and it does not show up in any medical statistic. A player returning after seven months instead of nine may play every match, but he will avoid decisive challenges — and that is how the second phase of a career is quietly destroyed.
That is why every transfer analysis I write includes a section on physical risk, not just form. Form is a photograph. Physical risk is a film. Between the transfer market and the pitch, I choose to stand in the middle, measuring both sides.
But here is where I must challenge myself, because the xG model holds a trap I once nearly fell into. Correlation is not causation. Long An had low xG and were relegated. That does not mean low xG caused relegation. Both could be consequences of a deeper cause: a weak squad, a wrong tactic, a failed transfer window. If I read the 0.72 figure as a verdict, I would have committed exactly the mistake the editorial board made in 2026, only in the opposite direction. They rejected data out of emotion. I risk trusting data blindly out of reason. Both are ways of not reading.
The biggest blind spot in Vietnamese data analysis is not model quality. It is that models are used to replace judgement rather than support it. A good xG does not say who should start. It says where chances are coming from, and where chances are being wasted. The decision still belongs to the coach — but that coach should have a number beside his intuition.
And there is a variable I once underestimated: culture. Data has no culture, but the people who create data do. A Korean scout and a Vietnamese coach can look at the same spreadsheet and draw two different conclusions, because they read it through two different football cultures. Ignoring that is a modelling error, not just a communication error. Even a billion-dollar contract begins with a small note about minutes played.
Seven years after the report came back rejected, I am paid to write about that very model. I was once rejected in 2026 because of a model. Seven years later, I am paid to write about it. The truth, even when rejected, comes back — only next time it carries more data. I do not trust intuition. I trust the kind of intuition that has been verified over seven seasons.
The question for the next cycle of Vietnamese football is not whether to use data. The question is who will be the first to read it correctly, before the season ends and the answer becomes too late.
