Home Advantage in V.League: 156 Empty-Stadium Matches, One Adjustment Coefficient, and Models That Refuse to Update

Opening: A Shot With 0.06 xG

In the 88th minute at Hang Day Stadium, a diagonal ball from the left flank found the foot of the away team's striker. The side-footed finish went into the far corner. The match ended 1-0 in favour of the side that was not playing at home.

I logged the numbers for that moment: 0.06 xG. Across 90 minutes, the home side accumulated 1.94 xG, the away side 0.71. The actual goals were 0 and 1. Read the scoreboard and the story is "the away team got lucky." Read the data chain and the story is entirely different.

I have covered V.League since 2026. Over those seven years I have learned that most arguments in the stands are not about football, but about who gets to define the truth. People argue with feelings. I argue with the tracking logs of 22 players per match. When the press room laughs at xG, I know I am reading the right book — the one they have not opened.

The question I took home that night was specific: if home advantage is shrinking, what exactly is shrinking?

Method: What I Measure and How

In football, "home advantage" has long been treated as a constant. Classic prediction models, from Elo to Poisson, assign the home team an additive coefficient, typically floating between 0.3 and 0.4 expected goals. That coefficient is handed down from one generation of models to the next, largely based on European data from the 1990s and 2000s, when stadiums were full and travel was difficult.

V.League does not operate like the Premier League. Distances between clubs are shorter, travel infrastructure differs, and pitch conditions — a systematically undervalued variable — vary enormously. A model imported wholesale from Europe is measuring the wrong thing while believing it is measuring the right one.

To me, home advantage has three components of different natures: crowd noise influencing referees, pitch familiarity influencing individual technique, and travel cost influencing fitness. When one of the three disappears, the total coefficient must shift. In 2026 all three vanished at once for a stretch of time, and that was a rare experimental window that Vietnamese football handed over by accident.

I collect data in three layers. The first is match events: goals, shots, shot locations, from which I calculate xG. The second is process data: passes, PPDA, penalty-box entries. The third is decision data: cards, penalties, and the referee's position on each phase. Every layer is cross-checked against at least two independent sources before it enters any conclusion of mine.

The statistical definition of home advantage is a trap in itself. Many studies calculate it as the difference between the league-wide home win rate and away win rate. That approach folds team quality into a single number. A league in which the strong teams happen to play more home games in the first half of a season will produce a fictitious home advantage. I always strip out team quality before measuring, even though it makes the sample smaller and the confidence interval wider.

The Evidence Chain: 156 Matches and an Eight-Point Break

The 2026 season was suspended and then resumed in empty stadiums. I took all 156 V.League matches from that period, normalised by venue, by fixture congestion and by opponent quality. The home win rate fell from 46% to 38%. That shift had never appeared in the V.League data I archive.

Eight percentage points sounds small. It is not. An average home team plays about 13 home games per season. That drop is worth roughly 1.0 to 1.5 points per season — enough to reverse the standings of a club competing for a continental slot. Across two consecutive seasons, that error can change who goes to AFC competition.

I pushed the data one step further. If the crowd is the real variable, the drop should cluster at stadiums with large capacities and high average attendance. That is exactly what happened: the three stadiums with the highest average attendance before 2026 contributed most of the eight-point drop. Smaller venues changed far less. That is a discriminating test, not a general correlation.

I then extended the dataset across the 2026 to 2026 seasons, when crowds returned gradually and unevenly. The results did not snap back immediately. The home win rate recovered in an L-shape: rising fast at fully reopened venues, flat at those with capacity limits. The variable is not the season. It is the number of people actually present in the stands.

The Break in Process Data

The break did not stop at results. It appeared in process data before it appeared on the table.

I measured PPDA — passes allowed per defensive action — for each away team during the empty-stadium period. The index fell by an average of 1.4 units compared with the same period with crowds. In other words, away teams pressed harder, higher, earlier.

The mechanism is simple. Crowd pressure, the thing that makes away players hesitate in duels, vanished. With no tens of thousands screaming behind them, away players dared to push up to the very edge of the home team's defensive line. Empty stadiums do not erase the truth. They strip away the fog that 40,000 shouting voices once created.

That fog concealed an uncomfortable conclusion: most home advantage in V.League does not come from the home team playing better, but from the away team playing below its own ceiling. Home advantage, in most cases, is an advantage the opponent voluntarily concedes.

I also checked away teams' completed passes into the final third. That figure rose during the empty-stadium period, with the largest increase belonging to top-tier away sides visiting mid-table hosts. That is a sign they were being suppressed by crowds in ordinary matches, not a sign they suddenly improved.

Reverse-Testing With a Model

If home advantage is shrinking, old models must fail in a systematic direction. I took the classic Poisson model, re-ran every home fixture from 2026 onwards, and compared it with actual results. Forecast error rose markedly in matches between mid-table hosts and top-tier visitors. Exactly where the +0.35 goal coefficient does the most damage.

A single number can lie, but a model validated across 10,000 matches has no reason to pretend. The problem is that my empty-stadium dataset contains only 156 matches. Too few to assert, enough to suspect. I always draw that line for myself before publishing anything.

The Counterintuitive Angle: Correlation Is Not Causation

This is where most Vietnamese football data analysis goes wrong.

The popular narrative runs: empty stadiums, lost home advantage, crowds decide results. The chain sounds tidy, but it skips a more important intervening variable: the referee.

During the empty-stadium period, I cross-referenced cards and penalties awarded to home teams. Penalties awarded to hosts fell, but the decline was larger than the decline in penalty-box entries created by hosts. Home teams created fewer situations, and they also benefited less per situation. The variable that changed was not in the players' feet. It was in the whistle.

This leads to a conclusion opposite to crowd intuition: most of the home advantage we worship is a form of decision bias, not a competitive quality. Decision bias is something process can remove, without a single spectator. A neutral VAR assistant does the same job at a lower cost than 40,000 tickets.

I must state the limit clearly. My data does not prove referees deliberately favour home teams. It only shows that the degree of bias falls when there is no crowd. Those are two different propositions, and I refuse to merge them for a prettier headline.

The Overlooked Variable: The Pitch

In V.League, the pitch is a variable as powerful as the crowd, yet almost nobody writes it into a model.

Poor-quality grass inadvertently advantages the home side, because the hosts are already used to the ball's irregular bounce. A long pass on a bumpy surface does not follow the trajectory the visiting defender calculates. When the calendar is compressed and matches are moved to neutral venues, that advantage evaporates. This is why I always check pitch quality before drawing any conclusion about form.

I learned this lesson after nearly being fooled by a match in central Vietnam. The home team won 2-0 with only 0.8 xG. I nearly wrote that they got lucky. A day later I rewatched the footage and saw the away defence slip three times in the same spot on the pitch. That was not luck. That was pitch data I had not yet collected.

In 2026, in the press room after SHB Da Nang faced Hanoi FC, I asked the head coach about his side's 0.4 xG despite a 1-0 win. A male reporter cut in loudly, saying women know nothing about football and just invent numbers. I did not argue. That night I published a long analysis using tracking data from all 22 players to show the win came more from luck than from control. Since then I have always led with raw numbers before judgement, and always cross-checked at least two sources.

Major Tournament Season and the National-Team Mechanism

We are in a major tournament season, and at national-team level home advantage is amplified by flag pressure and head-to-head history. But the underlying mechanism is identical: noise affects referees, the pitch affects technique, and a little luck affects everything else.

I once predicted Croatia would reach the 2026 World Cup final using exactly this logic. Their PPDA of 8.2 was the highest pressing figure in Europe at the time, and their final-third pass completion sat in the top three. Many colleagues called me insane. When Croatia did reach the final, a few apologised. I turned down a television analyst role because I wanted to stay where I could dig deep into data rather than speak in short bursts on camera.

Crowds may remember a goal forever. I remember the third pass before it, where the real decision was made.

Over the past four years I have also observed another trend: young squads with little home experience benefit more from the loss of crowds. They have not yet built a habit of relying on the stands, so when the stands vanish, they lose nothing to replace. This suggests a recruitment strategy: smaller clubs should prioritise young players for away fixtures, since they are less affected by opposition crowds than veteran players.

The Transfer Market Paradox

There is a paradox in V.League that the data exposes clearly. The clubs spending the most money are buying the contracts with the highest brand value, not the highest tactical value. A well-known foreign striker in a regional league is paid three times the wage of a domestic midfielder who has proven his progressive-pass numbers per 90 minutes.

Every transfer contract is an equation with many unknowns. Most reporters only read the coefficient before the equals sign. They read the fee and the wage, then conclude the club got stronger. They do not read the age distribution, the injury history, and above all the fit between a player's profile and the team's pressing model.

The transfer race among V.League's big clubs remains an arms race of prestige. The real value sits with smaller clubs, those that buy one player who fits the system and pay 40% below the market price assigned to him. Data can create a competitive edge here, yet this is precisely where the fewest clubs invest in data.

A note on integrity, because it bears directly on how we read numbers. In esports, betting is eroding competitive integrity faster than in traditional sport because regulation lags behind market growth. Football is not immune. When a number is published by a party with an interest in it being believed, check the provenance before checking the meaning.

Closing: The Signal for the Next Round

I am not delivering a verdict on one match. I am proposing a technical adjustment: prediction models used for V.League should split the home coefficient into two parts, one adjusting for crowd presence and one tied to pitch characteristics. When the two run separately, forecast error in mid-table matches falls markedly across my historical dataset.

The error margin I accept is around 0.15 goals per match, provided pitch data has been updated within 48 hours of kickoff. If that condition is not met, I reset the coefficient to the previous season's default. Better to be wrong from excessive caution than wrong from excessive confidence.

The next round is the first test. Watch the mid-table home sides when they host top-tier visitors. If they win with a goal worth under 0.3 xG, do not call it home courage. Call it data repeating itself. And if someone tells you home advantage no longer exists, ask them what they measured it with, across how many matches, and under what conditions.