Nine Dimensions of a Basketball Game: When the Data Is Empty, the Conclusion Must Be Empty Too

The first page of the report carried a single line. The forty pages behind it were already scaffolded: nine sections, each with a table, each table with empty cells waiting for numbers. I wrote that first line at two in the morning, after opening the data file for the eleventh time: "Insufficient data to reach a conclusion."

Then I hit send.

The recipient was an analytics director for an Eastern Conference team. Four hours later he replied with one sentence: "Thank you for not making it up." Three months on, when that team entered the stretch run with two hamstring injuries inside four weeks, he called back and asked whether I could build the nine-dimension framework for his entire roster.

The door opened from the empty space, not from the filled one.

I tell this story not to advertise professional virtue. I tell it because across a decade of reading basketball numbers, I have chosen the opposite many times. I once held a correct finding and let it rot in a drawer because I wanted it to be more perfect. I once held a correct report and wrote it so thick that nobody finished it. I once held a correct number and lacked the nerve to say it louder.

Those three failures taught me what no classroom did: the hard part of this job is not finding the truth. The hard part is knowing how much truth you actually hold, and saying exactly that much.

The forty-page scaffold was built out of three mistakes. It has nine dimensions, and each one exists because I once left it out.

In 2026 I was twenty-four, newly hired at a basketball analytics blog in Los Angeles. That summer I sat at NBA Summer League in the cheapest seat, notebook ruled into columns, and I watched an undrafted free agent named Dillon Brooks. Nobody was paying attention. He defended with his hip, with his elbow, with a lateral step taken before the opponent put the ball down. Across five games his defensive rating was 98.3. The player competing for his position, Troy Williams, posted 104.2.

I spent three weeks building a probability model before publishing. Three weeks. I checked the distribution, re-ran the bootstrap, rewrote the limitations section three times. The day I hit publish, a rival blog had already run a piece celebrating Brooks three days earlier. Mine went unread. Not because it was wrong. Because it was late.

The 2026 lesson was not "write faster." It was this: a correct finding delivered late is worth the same as a wrong finding delivered on time. I learned to define "good enough" per piece: draft inside forty-eight hours, final twenty-four hours reserved purely for fact-checking, no chasing infinite perfection.

In 2026 I was twenty-five and built an early-signal framework for football: xG differential plus pressing intensity toward the box. The World Cup in Russia kicked off. Croatia were not winning on luck. They held 74 percent of possession in the middle third, and Luka Modrić created 12 key passes across knockout rounds. I published "Croatians Are Not Lucky" right after the group stage. It was buried because my name was too small. When Croatia reached the final, the piece was shared three thousand times in one night.

Every finding needs a moment before it becomes a fact. Croatia did not reach the final by accident. They were guided by people who knew how to read numbers.

In 2026 I was twenty-seven, a mid-level analyst at a sports data consultancy. The NBA suspended play. I spent four months studying the history of injuries after long layoffs. The finding: Kawhi Leonard carried roughly 1.6 times the hamstring re-injury risk if played at high density immediately after a stoppage. I wrote a forty-page report and sent it to the LA Clippers medical staff. Nobody replied. Forty pages, no executive summary, no recommendation on page one. In August, Kawhi went down exactly as the model predicted. The Clippers were eliminated in the second round.

Nobody read the report on Kawhi's knee. The market only reads after the sound of the snap.

In 2026 I was twenty-nine, hired by a brokerage to evaluate South American talent. I applied the early-signal framework refined in 2026. Enzo Fernandez at Benfica was producing 11.4 progressive passes per ninety minutes, with a 78 percent success rate under pressure — the best among under-23 midfielders at the Qatar World Cup. I sent a two-page report to a Premier League sporting director recommending a move at 30 million euros. In January 2026, Chelsea paid 120 million euros. My report leaked onto a data forum.

All four share one structure: I was right about the content and wrong about the delivery. And on the fourth, I was right about both and let slip something I was not permitted to let slip.

The nine-dimension framework grew out of that. It is not a forecasting model. It is a list of questions that must be answered, or explicitly marked as unanswerable.

Dimension one: tactics and technique. How does this team create advantage, and does that advantage survive four playoff rounds.

Dimension two: player data. Volume, efficiency, impact, and usage. Is this player posting pretty numbers in an unsustainable role.

Dimension three: operations and the salary cap. Contract structure, the tax line, apron position, remaining flexibility.

Dimension four: league landscape and team positioning. Which tier is this team in, and how long does the contention window stay open.

Dimension five: rules and governance. Which provisions are binding, and is anyone gaming them.

Dimension six: coaching staff and locker room. Who actually holds the voice, and does that voice survive a losing streak.

Dimension seven: risk. Six categories, each with a level, a probability, and a mitigation.

Dimension eight: media narrative and expectation. What is the story, and where is it in its heat cycle.

Dimension nine: industry ripple. How a decision on the floor reaches sneakers, broadcast, regional markets, the agency ecosystem, and derivative markets.

Nine dimensions, one rule: any cell without evidence gets marked as having no evidence, and is never filled with confident prose.

That is why the first page of that report held a single line.

Dimension one is the easiest to be fooled by, because it lets the writer sound like a coach. I usually start with a narrow question: where does this team create its first advantage. In pick-and-roll, the answer sits in the coverage choice. If the defense drops — big man retreating, on-ball defender chasing over the screen — the advantage belongs to the ball handler with a reliable mid-range shot. If the defense switches everything, the advantage shifts to the big who can post. If the defense hedges early, the advantage falls to the roller.

Three choices, three consequences, and one common error: reading a team by the name of its system rather than by how the system gets broken. I once wrote a piece praising a switch-everything defense, citing regular-season defensive rating. In the playoffs, opponents dragged that team's smallest defender into screen after screen, and the scheme collapsed across six quarters. The regular-season number was not wrong. It simply said nothing about transferability.

So I apply one test to every tactical claim: does it survive when the opponent has seven games to study it. A five-out lineup with five credible shooters generates enormous space in the regular season, because no team has time to prepare for every opponent. In a seven-game series, the opponent will pick the coverage that hurts you most and repeat it forty times.

An example I use when teaching young analysts in Vietnam: two bigs on the floor together. Twin towers looks wonderful on paper. But if neither can shoot from beyond the arc, the defense simply stands in the paint and you lose all your spacing. The data will tell you that — but only if you look at spacing instead of points per game.

Based on my experience watching games, there is one small signal almost nobody records: the breathing of the primary ball handler in the fourth quarter. When an offense leans too heavily on one man, that man starts breathing through his mouth around the thirty-eighth minute. It is an earlier indicator than any box score. And it belongs in the descriptive section, not the statistical one.

Dimension two is the easiest to misread, because basketball has too many metrics and too few honest ones. Four layers I always separate: volume, efficiency, impact, and usage.

Volume is points, rebounds, assists. The easiest layer to read and the easiest to be fooled by. A player averaging 22 points on 20 shots is not a good player. He is a player who is permitted to shoot a lot.

Efficiency is true shooting percentage and effective field goal percentage. This is where I check first. TS% puts threes and free throws over a common denominator, so it compares a two-point scorer and a three-point shooter fairly.

Impact is plus-minus and aggregate metrics such as EPM. This is the most dangerous layer, because raw plus-minus depends on who stands beside you. A good bench player sharing the floor with four other bench players will post a negative plus-minus, and that says nothing about him.

Usage rate is the layer I use to detect empty stats. If a player has high USG%, low TS%, and negative plus-minus, his pretty volume is a loan, not an asset.

There is one more check I learned from my own mistakes: playoff shrinkage. Plenty of players hold their efficiency through the regular season and drop ten percentage points of TS% the moment a series begins, because the opponent picks one thing to take away from them. If your report lacks that column, it is not a report yet.

And an age-curve check. A basketball career is asymmetric. The physical peak arrives earlier than the skill peak, and for players who play with their minds, the skill peak can extend four years past the physical one. Nikola Jokić won three MVP awards not because he jumped higher than anyone. He won because the one thing age does not take away is the ability to read ahead.

Dimension three is the one Vietnamese fans skip most often, and that is a real loss, because in modern basketball the salary cap is tactics at another layer.

Four structural tiers: max contracts, the mid-level tier, rookie-contract surplus, and position against the tax line and the aprons. The first three decide how good you are. The fourth decides how long you are allowed to be good.

Cross the second apron and a team loses almost every roster-building tool: no mid-level exception, no salary aggregation in trades, no receiving players via sign-and-trade. That is why champions routinely choose between holding the roster and owning a future.

When I evaluate a trade, I compute three numbers: price paid, fair value, and the premium ratio. But the fourth number matters most: the panic premium. When a team has just lost four straight and the front office needs a headline, they pay 20 to 30 percent above fair value. That premium never shows on the cap sheet. It shows up three years later, when the contract becomes an obstacle.

One lesson I repeat: a young player with two years left on a rookie deal is the cheapest asset in the league. A team holding three such assets can acquire a star without breaking its structure. A team holding none must pay in draft picks, and draft picks are the one thing money cannot buy back.

Dimension four demands the most bluntness, because it forces you to place a team in a tier. Four tiers: contender, playoff, play-in, and draft accumulation.

Tiering is not hard. Accepting the result is hard. Some teams consider themselves playoff teams while their age structure and contract structure say they belong in accumulation. The error is not ambition. The error is using ambition to buy a 32-year-old on a four-year deal, then getting stuck precisely when the young core matures.

The contention window is a three-variable equation: the age structure of the core, the contract length of the core, and the flexibility of the cap. When all three align, the window widens. When two align and one does not, the window narrows a little each year, and nobody notices until it slams shut.

Dimension five is the least read and the most surprising. Load management rules, rest-game provisions, extension clauses, and disciplinary mechanisms create constraints that box scores never display.

I always add a rule-gaming simulation: not to encourage gaming, but to predict it. If a provision lets a team retain a player below market cost, the team will use it. If a rest rule lets a star sit four games without losing salary, the team will sit him. Predicting behaviour does not require morality. It requires reading the rule correctly.

Dimension six is the dimension data never reaches, which is exactly why it matters. An impatient ownership will fire a coach after fifteen games, and every tactical analysis you wrote becomes waste paper. A locker room where two stars do not speak will lose games the box score says they should win.

The signals I track are not press-conference quotes. They are how a player reacts when pulled mid-quarter. Who stands off the bench first when a teammate falls. Whether the coach makes eye contact with the twelfth man during a film session.

None of that is measurable, and it predicts better than most of the metrics I have tried.

Dimension seven I build as a matrix, because risk is not a number but a set. Six categories: competitive, contract and financial, personnel, rules, public opinion, and systemic.

Each needs three inputs: level, probability, and mitigation. The category I always rate higher than others is systemic — risk that does not come from one bad decision but from the structure of the whole league. Schedule density is the clearest example. No medical staff saves a team from two games a week across ten months. That is a systemic risk, not an individual one, and it does not vanish by hiring another specialist.

Dimension eight is where Croatia taught me most. What is the story being told, where is it in the heat cycle, and do the fundamentals support it.

There is one ratio I always compute: media heat over fundamentals. When that ratio crosses a threshold, I know a correction is coming. A player on the front page for three straight weeks after four good games, while his season sits at league average, is a sell signal, not a buy signal.

And one sourcing principle I apply strictly: rank sources by structure, not by fame. A reporter with direct access to a front office is not the same as an aggregator account. Leak motive is not the same as informational motive. A leak that surfaces exactly when a team needs to inflate a player's trade value is not news. It is a move.

Data is like a book. The crowd looks at the cover; the wise read every page.

Dimension nine is the dimension most Vietnamese basketball writing skips, and I regret that, because it is the one that connects the floor to ordinary life.

A signing reaches the sneaker market. A star changing teams reaches the broadcast schedule. A league expanding reaches regional markets. An academy opening in a province reaches the agency ecosystem ten years later. And a national team's winning run reaches derivative markets in ways nobody forecast.

For Vietnamese basketball, this dimension is embryonic but moving fast. Once a domestic league has sponsors, home arenas, and paying spectators, analytics stops being a blogger's game. It becomes a scouting tool. And at that stage, a data error is no longer an error in an article. It is an error about a person.

That is why I tell young people who want this job: learning to read metrics matters, but learning to say "I don't know yet" matters more.

Now the hardest part — the part where the nine dimensions betray themselves.

A report with nine sections looks more authoritative than a report with one. A table with twelve rows looks more authoritative than a table with three. A forty-page document looks more authoritative than a two-page one. And that is the trap: completeness of form manufactures the feeling of completeness of content.

My Kawhi report ran forty pages and nobody read it. My Enzo report ran two pages and leaked because people read it. The difference was not the quantity of information. It was whether the information could be used.

But there is a deeper layer. If I build nine sections and seven of them read "insufficient information," my document looks like a complete document with holes, not like an honest one. The reader asks: then what exactly do you do for a living. That is the moment the profession tests you.

I chose the wrong answer once. In 2026 a client I badly wanted to keep sent me a data file containing a domain label and nothing else. Two options: tell them I needed data, or build an analysis from what I already knew about the team from outside. I chose the second. The analysis read beautifully. It was also wrong on three of four conclusions, because what I knew from outside was six months stale.

The client never noticed. I noticed. And I spent the next six months repairing something that should have cost one email.

That is why I am writing this. This industry rewards the people who fill the gaps. A headline with numbers always beats a headline with a question mark. A decisive prediction always gets shared more than a conditional one. And because the reward sits on the filling side, most basketball analysis you read today is empty space filled with confident prose.

I am not saying that to disparage anyone. I am saying it because I understand the pressure. When you are a Vietnamese writer covering the American market, you always feel you must prove you know more than the next person. That feeling is the number one enemy of accuracy.

Here is an uncomfortable truth: readers do not need you to be right. They need you to be confident. Those are not the same thing.

Reading early is not the same as reading correctly. I called Croatia from the group stage and was right, but I was also wrong at least four times in the following four years, and three of those four had the same cause: I trusted my model more than I trusted the fact that my model lacked data.

So since 2026, every report I write opens with an information-status line. Three levels: confirmed, hypothesis, insufficient data. That line does not make the report prettier. It makes it more honest.

And if you are wondering why an article about basketball spends four paragraphs on data honesty, the answer is here: basketball is a sport where one wrong number can change a person's career. A misread defensive rating can leave a free agent unsigned. An ignored injury report can end a season. A leaked two-page evaluation can quadruple a player's price.

We are talking about things that cannot be fixed with a published correction.

So I set three public verification markers for myself, so readers can come back and check.

Marker one, end of the 2026-27 season: the nine-dimension framework must be applied in full to at least twelve teams, with an information-status line on page one of every report. If I fail, this framework is an article, not a system.

Marker two, before the 2027 playoffs: at least one of my predictions must be wrong, and I must write about it within seven days — not to apologize, but to identify where the model fell short. An analyst with no wrong predictions is an analyst who predicts nothing.

Marker three, longer term: within three years, at least one Vietnamese team or league must use a serious analytics framework for player recruitment, not merely for media content.

What I write today may be forgotten. The system it builds will not be.

Correct data that goes unread is not data — it is a debt owed by the people who refused to read it.

And the largest debt in this profession is not the numbers that went unread. It is the gaps that got filled with prose.

If you have read this far and feel uneasy that the article offers no grand prediction, that is precisely the point I want you to carry out. Sometimes the most correct answer is the one that has not arrived yet. A good analyst is not the person who always has an answer. It is the person who knows exactly what is missing, and says so before anyone else notices.