The Empty Analytics Sheet and the False-Precision Trap in Basketball

At eleven at night I opened a data file and found a skeleton already built for me: the header row spelled out offensive rating, defensive rating, pace, effective field goal percentage, impact metrics. Below it, every cell was blank. No number, no player name, no game.

For someone paid to advise basketball teams on data, that is the worst imaginable outcome. Three minutes later I changed my mind. A blank sheet cannot lie to me. A filled-in one can. If somebody types 112.3 into the offensive-rating cell without a source, I will read it with exactly the same trust I give every other number, and then carry it into a real team's decision meeting.

The most dangerous thing in this trade is not missing data. It is manufactured data presented neatly.

I entered the profession through a personal blog. In 2026, aged twenty and sitting in Da Nang, I wrote about expected goals for SHB Da Nang. The striker Gaston Merlo averaged 0.8 expected goals per match but scored only 0.4. A young coach from another club commented publicly that a girl knows nothing about tactics and should stop reading numbers and guessing. I did not argue. I published the raw data for the next twelve matches: shot counts, shot locations, shot timings. The club took 9 points from 36, exactly as the model calculated. He apologised publicly.

What I kept from that episode was not that the model was right. It was that I had to open the raw data so others could check me. Transparency is the only fence against educated fabrication.

Thirteen years later, I work with basketball teams. Vietnamese basketball is at a stage where data departments appear inside clubs faster than the people to run them can be trained. A season lasts a few months, three games a week, and the coaching staff needs answers the same evening. Nobody has time to wait for a verified table. That pressure breeds a habit: fill the template first, check the source later.

Two stray instructions inside a data cell

The sheet I received that night was not merely blank. It carried two instruction lines nested inside the data cells: identify from the information points above, and judge from the source fields of the points. Both sentences are meaningless, because there were no information points above to identify anything from.

In basketball, this is the equivalent of a scouting report that says: evaluate by the player's advanced metrics. Which metric, from which game, over how many minutes, against what kind of defence? Nobody answers. The report looks professional, with a title, a table, a conclusions section. It is missing exactly one thing: the truth.

Numbers do not lie, but they cannot tell a story either. A clean column of figures can make an entire meeting nod, even when that column was generated out of an empty cell.

The lesson from Germany

In 2026 the whole world mourned the German national team. I quietly re-read the model's log file. Germany's PPDA in qualifying was 12.5, against an average of 9.8 for the five most recent World Cup winners. Their average distance covered was only 98 kilometres per match. I published a prediction that Germany would exit in the group stage. Colleagues called me a laboratory scientist. Germany finished bottom of Group F, lost 0-2 to South Korea, and went home.

The point is not the result. The point is that in that article I also published the condition under which I would be wrong: if Germany kept PPDA below 10 in their first two matches, my model collapses. A prediction without a falsifying condition is a guess wearing a statistical jersey.

Hot hand and standard deviation

Every coach talks about feel. I have no feel. I have standard deviation.

In basketball that feel is called the hot hand. A player hits four of five threes in a quarter, the arena stands up, the staff designs the next play for him alone, and the whole team believes it is witnessing a special state. Research from the 1980s onward keeps showing that such shooting streaks appear at roughly the frequency a random model predicts. The hot-hand effect exists; it is simply far smaller than the feeling inside the gym.

If those four of five threes came from Stephen Curry, I still would not invent a new law. I ask: how big is the sample, across how many games, did the defence change its coverage. A streak and a belief differ in one respect, and that is whether they can be tested.

Empty stadiums

In 2026, when competitions stalled, I collected data from 300 matches across 8 European leagues played without crowds and found the home win rate fell from 45 percent to 38 percent. I sent a report to a team sitting near the bottom of its table, recommending a high press from the first whistle in away games, since opponents had lost their crowd. The head coach was sceptical at first. After testing it in the second half of the season, the team took 12 of 15 away points, having taken only 6 of 15 before.

Basketball went through a comparable test in the season played at a single venue, when home advantage all but vanished because there were no crowds and no travel. Data is a monastery: the less noise, the more clearly you hear something trying to speak. Those seven percentage points were a context signal, not a law.

The blank sheet is the safest document in the room

Sports data people are addicted to shocking numbers. I was too. Seven years ago I would have presented a rare correlation with enthusiasm and called it a discovery. Now I ask myself before writing: does this number change how I set the lineup tomorrow night? If the answer is no, it is decoration.

The biggest risk facing a basketball analytics department is not a shortage of people, machines or data. It is false precision. Once a template is built, people tend to fill it in for symmetry, because an empty cell looks like laziness. But an empty cell tells the truth that we do not yet know. A cell containing 112.3 with no source lies that we do.

In the empty-stadium report I stated plainly that the seven percentage points were a correlation, not a cause. Compressed schedules, reduced travel and a shortened season could all contribute. An honest analyst says so before anyone thinks to ask.

Once a young coach told me: just give me one number. I laughed. I touch the future with a keyboard. I gave him a range, not a point: if opponents switch to a zone, this metric drops; if they stay man-to-man, it rises. He did not like it. Three weeks later he came back with the same request, and this time he understood why I never hand over a single number.

The signal for the next cycle

The strongest lineup is never five beautiful names. It is five equations sounding in harmony. And an equation is only trustworthy when every variable in it traces back to a real game, a real quarter, a real shot location.

Based on my experience tracking games, next season will not be decided by which team holds more advanced metrics, but by which team dares to publish the raw source of those metrics. When you open a scouting report, look for the source column before you read the number column. If the source column is empty, every figure behind it is wallpaper.

The question for the next cycle is not what data your team is missing. It is what that data sheet is hiding.