V.League 1: Set Pieces, PPDA and the Hole in the Third Column

At Thien Truong, minute 88, a corner from the left. The ball drops into the second contact zone, the home centre-back rises above everyone, and the header arrows down toward the far corner. The stand erupts. From the press seats I do not cheer. I open the spreadsheet and type another row: the fourteenth goal of the season from a set piece, out of thirty-one the club had scored by that round. Forty-five point two percent. The league average over the same span, drawn from a dataset I hand-labelled across 546 matches in three seasons, is twenty-four point seven percent.

People watch goals and roar. I watch a seventeen-minute probability chain to understand why it happened.

To reach that corner, the home side had strung together eleven cross-field passes in four minutes, dragging the defensive block hard to the right, then pushing the ball to the left flank exactly as the opposite full-back lost his footing. A set piece is the end of a sequence, not a gift falling out of the sky.

Nine years ago, aged twenty-four, I had a manuscript thrown back at me with one line: what would a woman know about tactics. I did not argue. I sat down, rewatched the footage, annotated every dead-ball phase and built a three-page methodology appendix. That piece became the first time the concept of expected goals appeared on a Korean football platform. Since then, every claim I make carries its raw dataset so anyone can check it.

The foundation: fourteen clubs, twenty-six rounds, and almost no public data

V.League 1 runs with fourteen clubs and twenty-six rounds, which is 182 matches per season. That is a far smaller sample than any European league, so statistical noise runs much higher. Three straight wins lift a side four or five places; two defeats drop it into the danger zone. Any conclusion built on a five-match sample here is fragile.

The second problem is heavier. The league has no public standard data feed. No real-time shot-location provider, no PPDA, no complete passing maps. To analyse anything, I had to build my own source. Six years ago, when stadiums shut and my newsroom lost seventy percent of its revenue, I refused to write speculative pieces about what might have happened without the pandemic. I sat down and hand-labelled 632 old matches, shot by shot, dead ball by dead ball, active defensive action by active defensive action. The whole world stopped spinning, but my ghost football database kept breathing. Today that set has grown to 546 V.League 1 matches across three seasons, cross-checked against four other regional leagues.

The league's financial structure explains the data scarcity. Most club budgets come from a single sponsor or a single owner; broadcast and commercial revenue distributed to each club sits very low. When income concentrates in one person, investment decisions concentrate in one person too. The analytics department is the first line cut, and it is also the cheapest line that generates an edge.

Three columns that separate champions from relegated sides

Across three seasons and 546 matches, I filtered out the three variable groups with the strongest classifying power.

The first is the share of goals from set pieces. The four sides finishing in the leading group averaged between thirty and thirty-three percent. Mid-table clusters sat around twenty-four percent. The bottom four managed only eighteen to nineteen percent. The gap between the top group and the bottom group reaches fifteen percentage points, close to a third of a season's total goals. In a league where open-play chance quality is very low, dead balls are the cheapest remaining source of goals.

But reading that far only reads the first column. The second is efficiency: goals per hundred corners. Clubs with a dedicated set-piece coach averaged 4.9 goals per hundred corners; the rest averaged 2.6. That is nearly double, and the cost is one job title on the coaching staff plus roughly forty hours of video a month.

The third column is the decisive one. Here I do not count goals, I count chances created per corner, meaning the repeatability of the process. One side can score seven goals from eighteen corners in a season; another scores seven from seventy. The league table looks identical, the real value is nothing alike. Their point of failure is not in the dressing room. It sits in the third column of the spreadsheet I filter.

The second variable group is PPDA, the passes an opponent is allowed before each active defensive action. The lower the number, the higher the press. In my dataset, the top four sides averaged 10.4; the bottom four averaged 17.1. The bottom group let opponents make nearly seven extra passes before they touched the ball.

What caught my attention was not the metric itself, but its correlation with points won. I calculated roughly 0.34. For set-piece goal share, the equivalent figure is 0.71. In V.League 1 conditions, how many passes you allow matters less to your results than what you do with the ten most distinctive minutes of the match. Correlation is not causation, but the distance between 0.34 and 0.71 is enough to direct where resources go.

The third group concerns personnel and minute allocation. Foreign players take 15.2 percent of total minutes yet score 38.6 percent of the goals. For lower-half clubs alone, that share passes 46 percent. This is structural dependency, not a tactical choice. When a club has only one foreign striker capable of scoring, every attacking plan has to run through his feet, and opposing defenders know it.

Minutes are distributed just as unevenly. Last season I counted close to thirty domestic players above two thousand two hundred minutes, effectively playing almost every match. That group showed a muscle-injury rate in the final three rounds more than sixty percent higher than players under one thousand eight hundred minutes. In a league with a congested calendar and long travel legs, this is a debt written against tomorrow.

One further layer sits in the timing of goals. Sides with genuine squad depth scored 61 percent of their goals in the final thirty minutes; thin squads scored 44 percent. Substitution timing exposes the same gap: the leading group made its first change on average in the 54th minute, the bottom group in the 68th. For forty minutes before that first change, they played the same eleven that had already been read. Early-substituting teams won 0.34 more points per game, which across twenty-six rounds is roughly nine points.

There is one warning signal I always keep separate: teams that outperform expected goals by more than eight in a single season tend to collapse the following year. Finishing above the baseline does not repeat. The table records the result, the database records the process, and the two do not always travel together.

The problem is not talent, it is the semi-finished-goods mechanism

The standard media story when V.League 1 hits a crisis is a shortage of talent, money and crowds. I do not see the data supporting that reading. Names like Nguyen Quang Hai, Nguyen Hoang Duc and Nguyen Tien Linh all came through domestic academies, and the next cohort keeps being produced. What is missing is a mechanism to keep them long enough for the developing club to capture matching value.

The pattern repeats: a mid-table club finds an eighteen-year-old, gives him around one thousand eight hundred minutes, shapes him across two seasons, then sells him to a bigger club. Compensation in many cases does not cover the development cost and carries no downstream profit-sharing clause. Small clubs become finishing schools for the big ones, and so they never have an incentive to invest long term in themselves.

The loan-with-obligation-to-buy structure I have tracked in several markets makes the same problem worse: it lets big clubs defer cost recognition while pushing small clubs to build financial plans around money that may never belong to them. At that point the final decision no longer sits with the football people. It sits in the third column of the accountant's sheet.

In the other direction, I have to be sceptical of my own data too. A club with a high set-piece goal share is not automatically well coached on dead balls; it may simply be impotent in open play and forced to live on corners. A high number has two opposite causes, and only the third column separates them. Data practice is not prophecy. It is so you are never fooled twice by the same lie.

What to watch next round

At thirty-three, I believe every number is a witness that never lies under oath, but a witness only answers the question you actually ask. Next round, two signals are worth watching. The group whose set-piece goal share is above thirty percent while chances per corner have fallen for three straight rounds carries the highest probability of fading. The group holding results while PPDA rises steadily each round is spending down its own fitness reserve.

If you coach a lower-half club, the question is not whether to buy another foreign striker. The question is who on your staff will spend forty hours a month turning your ten minutes of dead-ball football into a real advantage. Vietnamese football does not lack talent. It lacks people willing to read the third column.