Nine Layers of a Professional Basketball Report, Seen Through an Empty Analysis

The first page reads "analysis subject: insufficient information." The second page, the tactical section, says the same. By page twelve, every data cell — from true shooting percentage to usage rate, from salary structure to contention window — carries one identical line: insufficient data to conclude. No player names. No scoreline. Not a single number.

The young analyst who prepared that document sat still, waiting to be escorted out of the room. The head of basketball operations read all twelve pages, closed the folder, and said something no one in that department has forgotten: this is the most honest document this office has produced in six months. I have sat in rooms like that, in two different time zones, and I believe the sentence.

Professional basketball entered this decade with near-absolute faith in data. Every team in a major league runs its own analytics department, hires modellers, installs optical cameras that track every footfall and every release angle. The volume of raw data multiplies each season. The quality of decisions does not rise at the same rate — and that is the point few people dare to state aloud at industry conferences.

That empty report, despite having no content at all, accidentally exposed the correct scaffolding. It listed all nine layers a professional report must pass through, and at each layer it said plainly: where data is missing, say it is missing. Read closely, it is a checklist rather than an analysis. In my trade, a correct checklist is worth more than ten incorrect analyses.

The tactical layer opens with four questions. How advanced is the offensive system. How well is it executed. Do the personnel fit it. And which numbers support all three previous answers. Offensive rating, defensive rating and pace are the minimum three axes any report must carry. The most important axis, though, is rarely written as its own line: transferability to the postseason. A system earns trust only when it survives four playoff rounds.

That data does not live in the final box score. It lives in possessions after timeouts, in closing lineups, in the way a team deliberately fouls when down three with forty seconds left. I once spent an entire season logging timeout situations and closing lineups for eight teams. What I learned: the final minutes do not create tactical identity, they merely expose it.

One case forced me to rewrite my entire approach to spatial metrics. Stephen Curry forces opposing defences to stretch beyond the three-point line even when he does not have the ball, and every teammate's offensive efficiency number must be re-read under that condition. A spreadsheet that cannot describe that gravitational pull will reach the wrong conclusion about all five players on the floor, even if every individual cell is correct.

Player data has four rungs, and the fourth is usually dropped. The basic rung is points, rebounds, assists. The efficiency rung is true shooting percentage and overall efficiency rating. The impact rung is plus-minus and estimation models of contribution. The final rung is usage rate — how much offensive resource a player consumes. Drop the fourth and you misread the third entirely: twenty points at very high usage is not the same as twenty points at average usage.

The age curve is the most neglected section. A twenty-nine-year-old and a thirty-three-year-old can post identical numbers this season, yet their market values differ by a gap no figure on the page can explain. Decline risk must be written as its own line, placed beside injury history and position. Those three lines together produce a valuation; any one of them alone produces a guess.

There is a test I run before trusting any table: is this real data or filler data? A player scoring heavily in decided games, posting clean numbers once his team is down twenty — the phenomenon is real and measurable. The only way to catch it is to isolate possessions while the game is still live, then compare against everything else. Numbers do not lie, but people can teach them to lean.

Salary structure splits into four occupying groups. Maximum contracts. The mid-level tier. The surplus generated by rookie-scale deals. And the luxury tax threshold together with the hard ceilings that cost a team its roster-building exceptions once crossed. Rookie surplus is where the largest returns in modern basketball are generated, because a young player contributing at a max-level rate occupies only a fraction of the cap. Teams that understand this build more durable contention windows than teams that only know how to sign big contracts.

The trap sits in the panic premium — the price a team pays out of desperation rather than calculation. The trade market reprices that desperation every February, when teams chasing playoff spots are forced to act. My faith does not rest on chance, it rests on large denominators: the team that accumulates future first-round picks and keeps a flexible salary band wins over four years, not four weeks.

The league landscape stacks into four clear tiers: title contenders, playoff teams, play-in teams, and teams deliberately losing for draft position. The contention window is the product of three variables — the age of the core, the contract length of the pillars, and cap flexibility. Miss one variable and the window shuts a season earlier than planned. And one season in professional basketball is the entire peak of a handful of human careers.

Rules and governance are load-bearing walls, not an appendix. The collective bargaining agreement specifies how to sign, how to extend, how to cross thresholds, how to penalise, how to manage workload. Some teams build an entire art around exploiting the gaps in those clauses. Others lose a playoff berth because they misread one line about a trade deadline. In a decent report, every tactical conclusion must be re-tested once through the lens of the rulebook.

The locker room holds the questions no spreadsheet contains. How long will the owner stay patient. At what level does the front office actually operate. How secure is the head coach's chair. And can two stars share one ball. The last question is the most expensive, because it determines the rest of the document, and it is almost never answered with quantitative data.

Risk must be written as a matrix, not a list of feelings. Six categories: competitive, contractual and financial, personnel, regulatory, public opinion, systemic. Each needs a level, a probability, an impact, and a mitigation. A report without a risk matrix is a report that only knows how to praise — and that kind of report tends to surface precisely when an organisation most needs a warning.

Media narrative and expectation is the layer analysts like to exempt themselves from. The market always carries an expectation level, and the analyst's job is to measure the gap between that expectation and reality across three separate domains: team record, player form, award outlook. Add a source-tiering pass — a trade rumour from a tier-one source, a tier-two source, or merely the echo of an agent mid-negotiation represents three entirely different levels of reliability.

Industry ripple is the final layer of the map. A decision at team level reaches the shoe manufacturer, the broadcast schedule, the regional market, the agency ecosystem, and the derivative markets. In the United States, broadcast rights and on-court footwear revenue are large enough that a single player departure can reshape an entire season's schedule. Skip this layer and the report remains technically correct and strategically useless.

And there is a tenth layer, never listed but decisive: handling the void. When an analytical dimension has no data, the report must state that there is no data, rather than filling the space with speculation. It sounds simple. In practice, the pressure to fill empty cells vastly exceeds the pressure to be honest. No one is reprimanded for adding an unsupported claim; people are reprimanded for submitting a blank page.

I learned that lesson from an unremarkable match. In the 2026 MLS season, New England beat Atlanta United 2-1, but my notes showed Atlanta generated 2.8 expected goals against the hosts' 1.1. I wrote that the visitors were merely unlucky. An entire forum called me a daydreaming bookworm. I did not argue. I kept collecting until Atlanta's season average settled at 1.87 expected goals per match, and they reached the playoffs.

The lesson was not that I was right. The lesson was that I was right only because I had data to count. Had I lacked expected-goals metrics that day, the only honest option was to write that I did not know. And had I written that I did not know, no one would have called me a bookworm — and no one would have remembered the piece. That incentive structure pushes an entire industry toward unsupported conclusions, and it operates so quietly that almost nobody names it.

A year later, in the round of sixteen at a World Cup, Spain held 74 percent possession and lost to Russia on penalties. My pressing data gave Russia a figure of 7.8 — meaning they deliberately surrendered the flanks and sealed the central lane, inviting sideways passes that never went forward. My piece argued Russia had every basis to eliminate a stronger opponent. When it happened, a well-known German coach shared the article with one line: data does not lie. I keep that line, and I still think it is true — provided the person reading the data knows when to stay silent.

Here is where I break with the crowd. The industry measures report quality by fullness: how many metrics, how many charts, how many pages. Fullness is not a measure of accuracy. A report with forty filled cells and three fabricated ones is far worse than one with twenty filled and twenty blank. A blank cell is a signal. A fabricated cell is a crack you cannot see until the whole wall comes down.

Every system cracks if you look long enough. Then you find the order sitting inside the wreckage. In basketball the crack usually appears in exactly three places: a metric computed from too small a sample, a correlation read as causation, and a conclusion retained only because it was once published.

Correlation is not causation — everyone knows the sentence, almost nobody applies it. Teams that shoot threes better win more, but that does not mean shooting threes makes them win. Both may be the product of generating more open space. If the report cannot separate those two possibilities, it is selling you a belief rather than an analysis.

There is another test I apply to every model I build: reverse the hypothesis and see whether the data still holds. A model that works in only one direction is an unvalidated model — it has merely been decorated. On the pitch or in the virtual arena, entropy behaves exactly the same way, and entropy never yields to a handsome chart.

This is what I want to say to those entering the trade: the hardest skill is not producing numbers, it is daring to leave a cell empty. I do not guess, I count. And then one day the gem surfaces from the raw data — but it surfaces only for the person patient enough not to fill every gap with a hypothesis. Those who fill will never see the gem, because their eyes have grown used to artificial light.

I enter data the way others enter meditation. Every number is a breath of the game, and every blank cell is a breath held — not from forgetfulness, but because the moment has not arrived. How a data department treats its blank cells says more about whether they truly understand the sport than any performance table ever could.

The next cycle of basketball analytics will not be decided by who collects more data. Cameras are everywhere, sensors are cheap, and raw volume outgrew human reading capacity several seasons ago. It will be decided by who dares to publish their blank cells first. Numbers stay silent, but the story never does — even when the story is about a page that has not yet been written.