The Empty Column in VBA Box Scores and the Cost of Deciding by Gut Feeling

The Empty Column in VBA Box Scores and the Cost of Deciding by Gut Feeling

Three minutes and forty-seven seconds nobody recorded

Three minutes and forty-seven seconds. In the play-by-play file of a VBA game at the Quan Khu 5 arena in Da Nang, there is a gap exactly that long. No shot, no foul, no rebound, no substitution. The clock jumps from 6:12 of the third quarter to 1:35, and in between is white space. The league's stat system goes silent, as if someone cut a segment out of the film.

I stayed until nearly midnight with my laptop, cross-checking the paper scoresheet handed to reporters, my own handwritten notes, and a phone video shot from the fifth row. The away team led by nine when the white space began. When it ended, the lead was three. Six points evaporated inside a stretch the official system says never happened.

Nobody complained. Neither coaching staff asked for a correction. Reporters filed their stories off the final score. Fans left believing they had watched a complete game. Only I sat there with a punctured log file, thinking about a line I use with young coaches: numbers do not lie, but they cannot tell a story either. And when the number disappears, people still tell the story. They just tell it wrong.

The white space was not a mystery. It was an entry error. But how Vietnamese basketball handles that error is what matters.

A nine-year-old league, a three-column ledger

The VBA tipped off in 2026 with six teams. Nearly a decade later, the data the league publishes is essentially the original frame: points, rebounds, assists, fouls, minutes, shooting percentages. That is everything an analyst has at 11 p.m.

There is no publicly released play-by-play. No shot coordinates. No deflection data, no matchup data, no time of possession, no on/off impact. What a major-league club treats as the most basic layer of information simply does not exist here.

I am not saying this to sneer. I have tracked this market since 2026, when I was a third-year student in Da Nang blogging xG analysis for a V-League club. Back then I published the raw dataset, the spreadsheet, and the collection method just to prove one simple thing: a striker can score steadily while still wasting chances. A young coach posted online that a girl knows nothing about tactics. Twelve games later his team had nine points from a possible thirty-six, exactly as the sheet predicted. He apologised publicly.

A professional league has sponsors, player contracts, transfer fees, ticket revenue, broadcast partners. All of that money flows through a decision system whose primary input is the human eye. The eye is not bad. It has one fatal flaw: it remembers the spectacular play, not the repeated one. Basketball is a sport of repeated plays.

Even a raw box score yields pace

People say you cannot analyse anything without advanced data. Technically, that is wrong. From the very box score handed out after each game, you can estimate possessions using a formula Western analysts have used for twenty years: possessions equals field goal attempts plus 0.44 times free throw attempts, minus offensive rebounds, plus turnovers.

The 0.44 is not a magic constant. It reflects the fact that a two-shot trip does not consume two possessions. That is the kind of detail a tired data-entry clerk at 10 p.m. never thinks about, and it decides the entire picture.

Take an example from my tracking notes this season. Team A wins 82-79. On the scoreboard, Team A was three points better. But the sheet shows Team A took 68 field goals, 20 free throws, grabbed 11 offensive rebounds and turned it over 13 times. Their possessions: 68 + 8.8 - 11 + 13, roughly 78.8. Their offensive rating: 82 divided by 78.8 times 100, or 104.1 points per 100 possessions.

Team B took 74 field goals, 12 free throws, grabbed 8 offensive rebounds and turned it over 18 times. Their possessions: 74 + 5.28 - 8 + 18, roughly 89.3. Their offensive rating: 79 divided by 89.3 times 100, or 88.5.

Read those again. Team A won by three, yet their offensive rating was 15.6 points per 100 possessions better. The three-point margin is an illusion created by Team B manufacturing more than ten extra possessions and missing almost all of them. Replay that game with the same two performances and Team A wins by double digits. Team B's coach can keep everything the same because the loss was only three points. That is the trap.

There is one more thing hidden in the formula: pace. The two teams averaged about 84 possessions each in forty minutes. That is a slow game. A team that claims to play fast but finishes the season at 78 possessions is not playing fast; it is shooting early. Those are different things, and only possessions can tell them apart.

No VBA team publishes its pace. No press conference in this league has ever asked about it. But it is free. It sits in the sheet handed out after every game, waiting for someone to spend ten minutes with a calculator.

Hot hand, cold hand, and standard deviation

There is a phrase I hear in every arena in the country: he has a hot hand tonight. It usually follows a made shot and precedes a decision, like keeping a player on the floor four extra minutes or giving him the ball on the decisive possession.

Take a guard from my notes. Over five straight games he took ten three-pointers each and made 5, 4, 6, 3 and 5. That is an average of 4.6, or 46 percent. The commentator says he is in rhythm, his hand is hot.

Now compute the standard deviation. The deviations from 4.6 are 0.4, 0.6, 1.4, 1.6 and 0.4. The standard deviation is about 1.14. If this were pure randomness, a 46 percent shooter's expected standard deviation on ten attempts would be the square root of ten times 0.46 times 0.54, about 1.58.

The observed deviation is lower than the expected one. The week described as a hot streak was actually an unusually stable week. Nothing was hot. A player shot exactly the way he always shoots, and a coaching staff looked at a five-game sample and called it a trend.

One more calculation for those who still trust the feeling. For a 46 percent shooter, the probability of a night of six makes or better out of ten is roughly 28 percent. That means roughly once every three or four games, he will have such a night without changing anything in his body. Over a twenty-game season, that is five or six nights, scheduled by probability itself. People will call it a star's moment. Analysts call it the binomial distribution doing its job.

Every coach talks about feeling. I do not have feelings, I have standard deviation.

I understand why the hot-hand story is seductive. It turns ten players into one hero. It makes a hard decision easy to explain in a press conference. But the cost is real: a player shooting 2-of-9 who gets the ball on the decisive possession because he went 7-of-10 the night before is receiving a privilege no data supports.

The market pays for points, not efficiency

This is where white space turns into real invoices.

Compare two players whose scoring averages give them market value. Player A scores 20 points on 25 field goal attempts and 6 free throws. Player B scores 14 points on 10 attempts and 4 free throws.

Player A's true shooting percentage is 20 divided by two times 25 plus 2.64, or 20 over 55.28: about 36.2 percent. Player B's is 14 divided by two times 10 plus 1.76, or 14 over 23.52: about 59.5 percent.

The gap is more than twenty-three percentage points. To reach 20 points at the same efficiency, B needs about 14 attempts instead of 25. In other words, B saves his team more than ten possessions a game, and ten possessions in an eighty-possession league is over ten percent of a team's total opportunities.

The transfer market will pay A twice what it pays B. I have seen this repeat, and not only in basketball. In 2026, while doing data analysis for a sports consultancy in Hanoi, I sent a report to a bottom-of-the-table club. The first thing the coaching staff asked me was not about efficiency. It was which player scored the most goals.

Points per game is a stat that depends on how often you are allowed to shoot. A high scorer on a weak team where he takes twenty-five shots is not necessarily better than a lower scorer on a strong team that shares the ball. The box score does not say that. The box score only says what the clerk typed.

Every transfer window, part of the VBA's money goes to low-efficiency players simply because nobody in the room calculated their true shooting percentage. That money is not lost to anyone. It simply never exists as points on a scoreboard.

Empty arenas and the lesson of context

There is a category of data this league does not have, and it has nothing to do with basketball: data about the environment around the game.

In 2026, when world football shut down, I collected numbers from three hundred matches across eight European leagues played without crowds and found something simple: home win rates fell from 45 percent to 38 percent. Forty-five percent was a figure everyone treated as obvious, the home advantage, the cradle of football. It lost seven percentage points because the stands were empty.

I sent that result to a V-League club fighting relegation, proposing they press high from the opening minutes in away games, because opponents no longer had a crowd to lean on. The head coach was sceptical at first. After testing it in the second half of the season, the team took 12 of 15 points from five away games, having taken only 6 of 15 in the first half of the season.

Vietnamese basketball has never published a home-court advantage figure. In my tracking notes, VBA home win rates typically sit between 54 and 58 percent, well above what a league without draws would produce. Where does that edge come from? Partly travel. The VBA has flights and long bus rides a major league does not impose. Partly floors, lighting, rims. And partly, perhaps more than anyone admits, the crowd and the human decisions on the floor.

Nobody measures which part is which, because nobody records it. Context is the data most easily dismissed as decoration, until it becomes the variable that decides a season.

Local players locked in the corner

There is one metric I want in the VBA more than any other: individual usage rate.

A quality import or naturalised player typically consumes about a third of a team's possessions while on the floor. That is reasonable for a star. The problem is the consequence: when one player takes a third of the balls, thirty to forty percent of the remaining possessions are distributed to his teammates in a very specific way, standing in the corner waiting for a kick-out and shooting when left open.

Local players are not technically inferior. They are placed in a role that permits them to develop exactly one skill. Three seasons in the corner produce a corner specialist, not a ball handler. And when the national team needs a ball handler, someone goes looking for an overseas Vietnamese player again.

Cases like Dinh Thanh Tam, Justin Young or Nguyen Huynh Phu Vinh are exceptions, and precisely because they are exceptions they get mentioned for years. On/off impact would answer this question within a single season if anyone bothered to record it. It needs no motion-tracking cameras, no sensors in jerseys, no ten-person department. It needs one person sitting near the bench with a stopwatch every time a substitution happens.

Until then, every argument about whether local players are good enough remains an argument between two people who have no record.

White space is not a finding

Back to the three minutes and forty-seven seconds.

One thing I want to state clearly, because this is where people in my line of work fool themselves: white space is not a finding. It is a gap. The two are different, and telling them apart is the measure of the profession.

I have seen enough data reports to recognise a common error. When a dataset is empty, some people write that no risks were detected. In a report to a coaching staff, that sentence gets read as: we are safe next game. But in reality, the biggest risk is not that a risk was missed. It is that nobody went looking. Silence gets read as reassurance. That is the worst failure an information system can produce.

This can happen to any VBA team this season. An unclassified injury, an unrecorded lineup change, a training week cut short by travel. All of them fall into the category of things nobody types into a machine. A week later someone looks at an empty table and says everything is fine.

I have one rule: question empty data, never paper over it. If a team's offensive rebound column is blank for three straight games, do not write that they failed to improve on the offensive glass. Write that nobody recorded it. That honesty looks small, but it is the difference between a league that learns from its mistakes and one that repeats them in silence.

Data is a monastery: the less noise, the more clearly you hear something trying to speak.

I also have to warn myself in the other direction. In recent years I have noticed how easily I am drawn to counter-intuitive numbers, to the point of nearly choosing them because they are counter-intuitive. A week ago I found a stat that looked impressive: a team had won eight of its last ten. I almost wrote about their strength. Then I checked the opponents. Six of those eight wins came against teams at least four places below them. The number was right. The story was wrong. I closed the file and wrote nothing.

Three traps I remind myself of every time I open a dataset: picking rare numbers for drama, predicting confidently as if I already saw the answer, and talking about Vietnamese basketball as a flat board I have finished reading. All three are the illusions of a long career.

Signals for the next round

In a regular season, the biggest stories never sit in the final standings. They sit in the columns nobody bothers to fill.

Four signals I will track until the season ends. First, free throw attempts divided by field goal attempts for the top teams, because it reveals who actually attacks the rim and who simply shoots from distance. Second, opponents' offensive rebound rate in away games, because it is a direct read on fatigue after long travel. Third, true shooting percentage of perimeter shooters over the mid-season stretch, the final six games and the playoffs, because sample shrinkage in the postseason is where heroic narratives get exposed. Fourth, pace, because it tells me who is truly running and who is merely shooting early.

If three months from now I still cannot answer these four questions, the answer is not that Vietnamese basketball is hard to analyse. The answer is that nobody bothered to write it down.

The return on record-keeping is very concrete. A team that appoints a full-time data analyst with a voice in the room, and gives that person two stable seasons, will most likely win one or two extra games a season compared with its own baseline. In an eighteen-to-twenty-game season, two wins is the difference between a playoff berth and a summer at home. That person costs less than a bench player. But the work never produces a number on the final scoreboard, so it never appears in any meeting minutes.

The last time a young coach told me a computer cannot understand Vietnamese basketball, I just smiled. I touch the future with a keyboard.