The Empty Stadium of Data: When Athletics Analysis Must Say 'Unrated'

In the summer of 2026, when the pandemic emptied the stands of the Bundesliga and the J-League, I sat in my Osaka apartment and collected data from 200 matches. The result: home win rates in the Bundesliga fell from 47% to 38%, while the J-League dropped to 35%. I wrote 'Home Advantage Is an Illusion' and sent it to Football Analytics JP. They published it immediately. That was the first time I understood that data can overturn a collective assumption.

But there is another lesson from that period I rarely tell. Once I received a completely empty data file — no athlete name, no discipline, no date, no competition, no source. Only a single label remained: 'athletics'. That was when I understood something the sports analytics world often forgets: the absence of data is not the same as data being zero. The gap between 'unrated' and 'low risk' is the gap between an honest analyst and a fabrication machine.

'An empty stadium does not kill football; it strips football's mask off.' Today I want to talk about another kind of empty stadium — an empty stadium of data, where the most forceful claim becomes a lie without evidence.

Context: athletics is a sport of anchored numbers

In athletics, every judgment must be anchored to a specific data point. Unlike football, where one can argue about 'team spirit' or 'pressing tactics', athletics is a sport where every conclusion can be reduced to a number: personal best (PB), season's best (SB), tailwind reading, venue elevation, weather conditions.

The deep analysis I am studying comes from a nine-dimension framework. It begins with an input-integrity check — and stops right there. The field-by-field audit shows: title N/A, source N/A, article type 'unclassified', information points empty list, core viewpoints blank, entities unresolvable. Diagnosis: the analytical pipeline returned a null payload.

This is a data-integrity failure, not a content judgement. One cannot say the article is weak, strong, accurate or inaccurate — only that its deconstruction is absent. Conflating the two commits exactly the error the framework is built to prevent: treating a missing measurement as a measured value.

Core analysis: the nine dimensions and what it takes to activate them

Dimension one — event and performance. Without a discipline, without a mark, one cannot classify a result as an official competition mark, wind-assisted, altitude-assisted, indoor, or an unratified training mark. The most important missing number here is the wind reading: a sprint mark counts for records only with a tailwind of no more than +2.0 m/s. And 'super shoes' — racing shoes with a rigid plate and supercritical foam midsole — deliver a systematic dividend that must be deducted when comparing across eras.

Dimension two — athlete condition. Without a birth date and a discipline, an athlete cannot be positioned on the peak-age curve. The 'abnormal PB explosion' screen — the key cross-check into the anti-doping dimension — is inoperable. A one-year gain of three times the historical average improvement requires a year-by-year PB series; here there is none. Injury risk must be read as 'unrated', never as 'low'.

Dimension three — competition structure and qualification mechanisms. Athletics has two entry pathways: the qualifying standard and the World Ranking points list. Their relative weight differs by event and championship cycle. Without the competition and the year, neither channel can be assessed. The national-selection variable — the highest-variance element in the whole system — is also blank. A one-race-decides-everything trials system and a comprehensive-evaluation system produce entirely different risk profiles for the same athlete.

Dimension four — event landscape and national strength. No dominance pattern can be assigned. Whether the event is a single-ruler discipline, a two-horse race, or in generational transition is a function of the top-10 season list and head-to-head records, none of which are present.

Dimension five — rules and anti-doping. No anti-doping risk can be scored, in either direction. The absence of a red flag here means 'not assessed', not 'clean'. Three missed tests within twelve months, Athlete Biological Passport anomalies, or association with sanctioned coaches — all require named facts. The DSD (Differences of Sex Development) rules affecting the women's 400m–1500m cannot be checked because the event and the athlete's category are both unknown.

Dimension six — team and training system. A coaching change immediately before a major championship is one of the highest-risk indicators, and it is invisible in a null input.

Dimension seven — risk landscape. This is the most dangerous point of the entire process. The overall risk rating reads N/A — cannot be rated. Because risk scoring here is the product of identified facts × event-specific hazard base rates. With zero facts, no product exists. The correct reading is 'unrated', explicitly not 'low'. Reporting this as low-risk would be the single most dangerous output this pipeline could produce.

Dimension eight — public narrative and expectation. The 'prodigy filter' cannot be applied. The historical fulfilment rate of 'next great champion' labels is low, and youth-record holders convert to senior elite status at a modest rate — that is a prior, not a finding about this article.

Dimension nine — athletics industry transmission. No transmission path can be traced, because transmission analysis requires an originating shock: a mark, a signing, a rule change, a commercial deal, a retirement. There is none.

Contrarian angle: the value of saying 'unrated'

This is where I want to argue with my own industry. Sports media, especially in Vietnam and Japan, has a habit of filling in blanks. When information is missing, we speculate; when we speculate, we assert; when we assert, we turn it into a headline. Each step is small, but together they create an information system where truth and conjecture are hard to distinguish.

There is a hidden technical problem here. When a fast athletics mark appears in an article without a wind reading, without venue elevation, without a competition round, and detached from its original context, any later re-use of that mark as evidence of true ability is methodologically invalid. This holds even if the mark is real and valid in its original context.

I saw this in the Japan–Belgium match in 2026 — when the whole country blamed fitness after the 94th-minute goal, while the real problem lay in coach Nishino withdrawing Inui and Kagawa, pushing the formation into a 6-3-1 that broke the passing chain and became an invitation for Belgium to push up. 'Fighting spirit' cannot beat 'effective pressing metrics'. But to say that, I needed slow-motion video and formation diagrams, not just emotion.

'Public opinion hates the contrarian, but history feeds it with time.' Saying 'I don't have enough data to conclude' is a contrarian act against the sports media industry. It generates no views. It generates no debate. But it protects the integrity of information, and over the long run that is the only thing that keeps this profession meaningful.

Takeaway: the riskiest tool is the confidently fabricating one

There is a paradox in sports analysis: the riskiest tool is not the one that makes mistakes, but the one that confidently fills in blanks. A poor prediction model will lose — and we notice its error. A fabricating model produces plausible-sounding stories — and we swallow them, repeat them, build narratives on them, until no one remembers there was never any source data.

So I tell every young athletics analyst: when you receive an empty file, say 'empty file'. When you lack a wind reading, say 'missing wind reading'. When you don't know who the athlete is, what the event is, what the competition is — write exactly that. The absence of data is not data being zero, and 'unrated' is not 'low risk'. That is not the weakness of the writer; it is the discipline of the analyst — the thing that separates us from the fabricators.