The Report With No Player in It: Tennis Analytics and the Discipline of a Null Result

Last week a tennis analysis report arrived on my desk. Nine sections, more than four thousand words, every table carefully built. In every cell the same line came back: “insufficient information.” The first read made it look like a blank document. The second read made it look like the most honest thing I have handled this season.

Tennis owns a moment that never reaches the scoreboard — the line judge's hand that stays down. The ball lands out, nobody speaks, the point continues. Call for the review and the evidence shows the call was wrong. The call that never came carried the largest piece of information on the court. This report sits in the same relationship to that moment.

The pipeline runs in two tiers. Stage-1 pulls names, dates, numbers, entities and information points out of the raw text. Stage-2 lays a nine-dimension framework over those points: technique and tactics, data and form, tournament system, tour landscape, rules and governance, management, risk, media narrative, industry transmission. Stage-2 validity rests on Stage-1 completeness. That is not a new policy; it is the arithmetic of the pipeline.

Then came the audit. Stage-1 returned almost nothing. No title, no source, no article type, an empty one-sentence summary, an empty list of information points, not a single name in the “entities involved” field, time sensitivity unassessed, source quality ungraded. One field survived — the domain label: tennis.

That single field was enough to raise the question. Tennis is the sport where fifty-two-week rolling ranking points, serve and return data, break-point conversion and winner-to-error ratios all sit in public view. Tennis is easier to audit than any team sport, because every decision leaves a number behind. If a sport this exposed produces a report with not one name in it, the problem is not in the sport. The problem is in the pipeline.

A null result is a legitimate result, and it is a different object from “no risk found.” The most important thing this report did was refuse to issue an overall risk rating. Writing “low risk” over an empty input set would have produced a false negative — an all-clear where nothing had been examined. Sports analytics commits that error constantly, because an empty cell looks like failure and failure is uncomfortable to display.

The framework's most valuable check could not run at all: the points-defense cliff. Inside the fifty-two-week rolling cycle, a large block of points can expire in a short window, and the ranking collapses if those points are not re-earned. A defending champion carries the sharpest version of that exposure. Running the check needs a ranking and a year of results history. Both were missing. The report therefore said nothing about any player's form, risk profile or commercial value, and it did not pretend to.

Coding all 169 goals of Russia 2026 taught me a line I still use — every goal is a data point until you watch all 169. On day one a studio producer handed me a coffee order. I handed back a one-page brief showing that more than forty percent of group-stage goals came from set pieces or second phases, while the teleprompter already carried the “counter-attacking World Cup” line. He read the numbers on air. He did not name me. That day produced my rule: no framework of mine reaches print without a named source, myself included. The rule has a second face — when nothing is known, that also gets written down with a name on it.

In 2026, at the NWSL Challenge Cup in Herriman, Utah, there were no spectators. An empty stadium is not empty information; the pitch microphones caught more than four hundred audible coaching cues and goalkeeper organizing calls. Boston gave me velocity; Utah gave me the pause between signals. There I had to build the instrument myself, because no off-the-shelf tool was built to hear that. Here the instrument is intact. The raw signal simply never arrived. The difference looks small and behaves enormously.

In 2026 I coded 48 races off public split sheets from a Boston dorm room, because a ticket to London was not in the budget. The “Split/Second” series was born from one belief: I built the pipeline before I trusted the pattern. In 2026, before Tokyo, I published a falsifiable prediction — in a spectator-less stadium the record most likely to fall was the men's 400m hurdles, because its rhythm is internal rather than crowd-fed. Karsten Warholm ran 45.94. I also flagged Elaine Thompson-Herah's 10.61 in the 100m. Pre-registration does not run one way. The prediction goes out before the event; the audit of what I got wrong goes out after it. By the same logic, the announcement that the information was not there belongs in print early, not late.

The encouraging part is that the failure is cheap and fixable. The framework is fully intact; Stage-1 needs a re-run. The two highest-yield fields are information points and entities involved. Recover those and Dimensions 1, 2, 3, 4 and 7 switch on, and that is where most of the analytical value sits. Source quality and time sensitivity come next. Without a source tier, confidence tags cannot be calibrated — a tabloid doping rumour and an ITIA statement do not deserve the same reading, yet both land in a database as “doping news.” Without a date or a season phase, surface-switch risk and points-defense windows cannot be projected at all.

The industry rewards volume. Nine sections, star ratings, filled tables — the shape of work completed. Four dimensions rated one star, because the content is absent, not because the analysis is weak. That is where the temptation arrives: place a name in the empty cell to be helpful — a top player, a points-defense cliff, a believable story. That is the framework's most dangerous trap. Invented information is never less harmful than wrong information, because it dresses itself as truth and closes the door on correction.

The empty report is itself a signal. A tennis article with no named entity, no stance and no purpose is not a normal specimen of tennis journalism. The likelier explanation is a failure inside extraction — encoding loss, or a truncated source. The report says nothing about sport; it says something about the pipeline. Holding that distinction matters, because a null result that leaks into the aggregation layer gets read as “no risk found,” which is far more dangerous than “not examined.”

Before the arena roars, someone has to map the noise. Sometimes the map reports that no signal is present here. That is the only certain statement in this report. The question turns back on my own desk — how many of our published conclusions rest on an extraction that failed quietly, with nobody noticing?