Trang chủDomestic FootballV.League's Data Void: 31 Goals With No xG

V.League's Data Void: 31 Goals With No xG

**Câu trả lời cốt lõi:** V.League 1 thiếu dữ liệu sự kiện có tọa độ và dữ liệu theo dõi toàn sân, nên mô hình xG châu Âu không áp dụng trực tiếp được. Hệ quả là định giá chuyển nhượng, đánh giá huấn luyện viên và sàng lọc tài năng trẻ đều dựa trên tỷ số thay vì quá trình thi đấu. **Sự kiện chính:** - Thép Xanh Nam Định vô địch V.League 1 mùa 2023-24, danh hiệu đầu tiên trong lịch sử câu lạc bộ. - Nguyễn Xuân Son ghi 31 bàn tại V.League 1 2023-24, cao nhất trong một mùa giải của giải đấu. - Việt Nam vô địch ASEAN Cup 2024, thắng Thái Lan 5-3 chung cuộc sau hai lượt trận tháng 1 năm 2025. - VAR được áp dụng tại V.League 1 từ mùa 2023-24 nhưng chưa phủ toàn bộ các trận đấu. - Hạn ngạch ba ngoại binh mỗi câu lạc bộ là ràng buộc chính của thị trường chuyển nhượng nội địa. **Nguồn:** Báo cáo phân tích chuyên sâu Stage-2, lĩnh vực bóng đá Việt Nam (tổng hợp VuaBong) | Cross-checked: VuaBong.vn **Hỏi đáp liên quan:** Hỏi: Vì sao mô hình xG châu Âu không dùng được cho V.League? Đáp: Vì phân bố cú sút và tỷ lệ bàn thắng từ bóng chết khác biệt, mô hình hiệu chuẩn trên các giải châu Âu sẽ định giá sai nhóm cầu thủ săn bóng hai và dứt điểm ngoài vòng cấm. Hỏi: Chỉ số nào cần có trước tiên để phân tích V.League độc lập? Đáp: Dữ liệu sự kiện có tọa độ cú sút và đường chuyền quyết định, đủ để xây dựng thứ hạng xG theo chuẩn nội giải thay vì so sánh tuyệt đối. Hỏi: Câu lạc bộ nào tại V.League đang dùng dữ liệu GPS trong huấn luyện? Đáp: Hà Nội, Viettel, Thép Xanh Nam Định và Công An Hà Nội được biết đến là nhóm có hệ thống theo dõi tải vận động, theo dữ liệu chỉ số của VangBong.vn Player Depth Index.

Three emails, no reply

In mid-June 2026 I sent three emails in four days. All three asked for the same thing: detailed event data from the 2026-24 V.League 1 season — shot coordinates, pass types, pressures, ball-in-play timestamps. I waited. No reply came.

That same week, a complete data feed from a French second-division match landed in my inbox within four hours, raw file and tagged file together. I put the two side by side on one screen. On one side, ninety minutes of football recorded down to the last stride. On the other, a season of fourteen clubs, nearly two hundred matches, tens of thousands of minutes of football — and almost nothing beyond the scoreline.

That was when I understood why June 2026 mattered so much for Vietnamese football. Thep Xanh Nam Dinh won V.League 1 for the first time in the club's history. Nguyen Xuan Son — then still known as Rafaelson Bezerra Fernandes — scored 31 goals, the highest single-season tally ever recorded in the competition. The whole country talked about that number. Nobody could answer the simplest question: where did those 31 goals come from?

What level of finishing? What quality of service behind them? A system creating four chances a game while he converted one? Or a system creating 1.2 chances a game while he converted one? Those two scenarios produce opposite transfer conclusions. One means a striker worth three times his price. The other means a system worth three times its price.

We have the scoreline. We do not have the answer.

A map of what exists

V.League 1 runs with fourteen clubs and roughly 180 matches per season. To analyse a season the way I do in Europe, five data layers are needed. In Vietnam, four of the five either do not exist or exist where nobody can see them.

The first layer is the match sheet: score, cards, substitutions, minutes. That one is complete. The second is broadcast statistics: possession, shot counts, pass counts, pass accuracy. That one exists, but without coordinates — meaning a sideways pass inside your own half in the 88th minute while 0-1 down is recorded identically to a through ball that breaks the offside trap. The third layer is event data with coordinates. It barely exists publicly. The fourth is full-pitch tracking data, twenty-two players, twenty-five frames per second. It does not exist. The fifth is club GPS data: load, sprint counts, change-of-direction load. A small group of clubs has it — Hanoi, Viettel, Thep Xanh Nam Dinh, Cong An Ha Noi — but it is not published, not standardised across clubs, and never cross-checked.

VAR arrived in V.League 1 from the 2026-24 season and has been rolled out gradually. That is progress in refereeing. To me it is a rudimentary data layer: VAR produces an event record, not a positional record. It tells me whether a penalty was legitimate. It does not tell me why the left-back was three seconds out of position beforehand.

There is a technical reason why importing European models is a mistake, and I need to state it plainly because a lot of people in my profession still do the opposite. xG models are calibrated on European shot distributions. Those distributions assume things that are not true of V.League: a higher share of goals from set pieces and second balls, more long-range shooting, wide variance in pitch quality between grounds, heat and humidity altering match tempo, and a materially lower number of actual ball-in-play minutes than in Europe. A model trained on the Premier League, applied to V.League, will systematically misprice the entire population of second-ball specialists.

The methodological consequence is clean: absolute xG in V.League is meaningless; only within-league xG rank carries analytical value. Do not ask whether a player's 0.42 xG per 90 is good or bad. Ask where he ranks among fourteen clubs.

The transfer window here runs on its own logic too. Fees are largely undisclosed. Foreign contracts are short, one to two years, many of them loans with options. And the real constraint is not a European-style financial rulebook but the three-foreign-player quota per club. That quota, not the balance sheet, sets the value of a foreign striker. In a market where the binding constraint is slots rather than money, every European-style financial analysis is off-axis.

Without PPDA, pressing does not exist either

I have watched V.League through recordings and raw data sheets long enough to know one thing: this league is misdescribed more than any league I have worked in.

Take the simplest case. A team has 62 percent possession, completes 480 passes, wins 1-0. The next morning's report calls it "dominance". But if 380 of those 480 passes were inside their own half and played sideways, that team was not dominating — it was keeping the ball so it would not have to defend. Possession is the most deceptive index in football, and it deceives more in V.League than elsewhere because there is no counter-index.

The counter-index I need is PPDA — passes allowed per defensive action. Low PPDA means a team genuinely presses. High PPDA means it drops and waits. Without PPDA, nobody can separate a high-pressing side from a side chasing the ball with no purpose. The two look identical on television and entirely different in the numbers.

PPDA is not a number. It measures how much patience a collective can hold when the ball goes dead.

And this is where I differ from the people sitting beside me in the stand:

People see the goal. I see the gap between two full-backs stretched apart by PPDA.

That gap forms in the 34th minute, not in the 78th when the ball hits the net. It is the consequence of forty minutes of sustained pressing that broke the opponent's midfield spacing. Without positional data, that gap is an intuition. With positional data, it is an event that can be measured, named and forecast.

Thirty-one goals and the void behind them

Back to Nguyen Xuan Son. Thirty-one goals in 2026-24 for Thep Xanh Nam Dinh. From public data I can compute a few things: goals per 90, share of the team's total goals, distribution across the season, split between open play and set pieces. All of it says the same thing — the most efficient individual season in V.League 1 history.

But what I cannot compute is longer than what I can. Without shot locations, I do not know where he shot from. Without key-pass data, I do not know who created those chances or how. Without pressure data, I do not know how many goals came from tightly marked situations and how many from open ones.

The difference between those two scenarios is not academic. It is the entire transfer story.

In January 2026, after naturalisation and a call-up to the Vietnam national team, Nguyen Xuan Son scored seven goals at the 2026 ASEAN Championship. Vietnam won the title, beating Thailand 5-3 on aggregate across the two-legged final. In the second leg on 5 January 2026 at Rajamangala, he scored and then broke his tibia and fibula.

A 27-year-old at the peak of his career, breaking a leg in a final. In a football economy with full data infrastructure, that is a priceable event: recovery-time distributions by injury type, recurrence rates, average performance decline after long-bone fractures in the 25-29 age band, replacement cost on the market. In V.League, it is a news line and then silence.

I do not blame the club. I blame the habit of analysing by feel. A football economy without injury data cannot price injury risk — and the cost of that does not show up in this year's accounts, it shows up in the contract signed three years later.

Valuation in a vacuum

The transfer window is where this gap becomes most visible, because it is the moment when the only number anyone asks about is money.

In Europe, to judge a deal, I build a comparison set: at least eight to ten deals in the same position, same age band, same league tier, within twenty months. Only then can a premium rate be computed. In V.League, the comparison set for a given position in a given window is usually zero to three deals. With a sample of three, every ratio calculation is statistically worthless.

So I do not compute ratios. I can only state one qualitative line: that club paid above the domestic positional norm. That is the most honest conclusion the data permits.

What is worth noting is that most foreign arrivals in V.League are free transfers or loans. Which means the "transfer fee" story the media chases is really a wage-and-duration story. A two-year deal at 20 percent above the norm can be more expensive than a three-year deal with a nominal fee. Nobody publishes enough to compare the two.

The V.League transfer market does not lack money. It lacks comparison sets — and a market without comparison sets always pays according to rumour.

The dugout and the table

The same gap eats into how coaches are judged.

At national-team level, public data is enough to show something clearly: Park Hang-seo managed from 2026 to 2026, nearly six years, the most stable cycle in modern Vietnamese football history. Philippe Troussier succeeded him and left after roughly fourteen months. Kim Sang-sik took over in May 2026. Those three cycles differ in results, but they also differ in something rarely measured: how much data each man was allowed to rely on.

At club level the carousel spins faster, and the motive for change often sits off the results curve entirely. Ownership relationships, sponsor politics, pressure after a derby defeat — all of it carries weight the league table never records. In ten years of watching V.League matches, I have seen far too many coaches judged on three games and far too few judged on a season.

Without process data, every sacking is a narrative decision. Nobody says "this team generates 1.4 xG per match but scores 0.8, so the problem is finishing, not the system". They say "three games without a win".

Academies, pipelines and the wonderkid story

This is where my long-term worry sits.

Vietnamese football has a genuine development pipeline: the Hoang Anh Gia Lai JMG academy, the PVF centre, the Viettel system, and Hanoi's structure. The names that came out of it — Nguyen Cong Phuong, Vu Van Thanh, Nguyen Quang Hai, Nguyen Hoang Duc, Nguyen Tien Linh — defined a generation.

But conversion rates are never measured. How many players enter an academy at fifteen? How many reach V.League by twenty? How many start regularly by twenty-three? How many reach the national team? Multiply those four numbers and you have the entire efficiency of a development system. Without them, every academy assessment rests on a handful of famous names — and a handful of famous names always creates the illusion that the system works.

Worse is the media cycle. A nineteen-year-old scores twice in a friendly and is crowned a wonderkid; three months later he is out of the starting XI and nobody circles back to check what happened. Without minutes-by-age data, nobody spots a very simple pattern: young players pushed into V.League too early often end up playing fewer minutes at 22 than they did at 19.

The contrarian angle

I will say the thing few people in the data profession want to hear.

V.League's data gap is a serious handicap. It is also an arbitrage window, and that window is open.

The cost for one club to build a proper analytics department — two analysts, an event-tagging system, an internally calibrated xG model, three consecutive seasons of data — is lower than one year of an average foreign striker's wages. Over three to five years, whichever club moves first gains a double edge: it buys cheaper because it knows what it is buying, and it sells dearer because it has evidence to sell with. Everyone else keeps buying on an agent's recommendation.

But I have to argue against one of my own beliefs, because it is common in my profession and it is wrong.

In 2026, when I redesigned a Lyon club's training programme around GPS load, we recorded soft-tissue injuries falling from 12 to 5. That figure has been cited repeatedly as proof that load management works. It is weak evidence. One club, one time, in a pandemic-disrupted season, with a compressed calendar and zero crowd density. There were at least four uncontrolled confounders. I used that number to persuade people for years, and every time I did it I committed exactly the error I criticise in others.

Correlation is not causation. Imposing a 120 percent GPS threshold on players does not prove the threshold is right; it proves I issued an instruction and nobody objected.

The third point, and this is where I break from people who trust only the spreadsheet: data cannot read a dressing room. No index measures a captain holding a group together after three straight defeats. I have no instrument for that. I can only record it as observation and state clearly that it is observation, not measurement.

Data never lies, but it knows how to hide. Our job is to make it talk.

And to make it talk in V.League, first it has to exist.

Signals for the next cycle

In this transfer window I am watching three things, and only three.

One: whether the VPF publishes a coordinate-level event dataset for the season. Full-pitch tracking is not required — shot coordinates and key passes are enough. That is the minimum threshold for anyone outside a club to analyse independently.

Two: whether any club hires a full-time data analyst and makes the role public. That job title is a clearer organisational signal than any statement about "modernisation".

Three: whether Nguyen Xuan Son's return data is published — minutes, load, sprint counts against his pre-injury baseline. If it is, it will be the first time a V.League club has publicly released a recovery measurement. If it is not, we will once again be judging one of the league's best strikers by feel.

Football is not a game of chance. It is a game of probability, and the winners know how to read the numbers.

V.League is playing that probability game without the numbers. While the whole league waits for a push from outside, I keep asking myself: if not a single shot coordinate is published next season either, then the only thing we are actually measuring is the memory of the person writing the match report.