The Athletics Dossier With No Data: Five Red Flags and the Limits of Every Analysis Sheet

Two Files on the Same Screen

In June 2026, at Yanmar Stadium Nagai in Osaka, Ryota Yamagata ran the 100 metres in 9.95 seconds, equalling the Japanese national record. I was sitting on the eastern stand with a paper notebook and a tablet. The first thing I recorded after the electronic clock stopped was not the time. It was the wind reading.

That afternoon the wind was within the legal limit set by World Athletics. The results sheet was clean: time, wind speed, competition name, date, athlete name, distance, placing. A complete dossier makes analysis boringly simple, and that is precisely why I remember it.

Three years later, a different kind of report arrived in my work inbox. Nine major sections, each with sub-tables and a risk checklist. Every field carried the same three characters: N/A. No athlete name, no distance, no wind reading, no date, no source. The closing note was a single sentence: insufficient information, cannot assess.

I put the two files side by side on the screen. One was the clean dossier of an afternoon in Osaka. The other was an empty skeleton with perfect headings and no content. The distance between those two files is my entire profession.

Data does not create stories; it strips the stories of others bare. When there is nothing in hand, the first professional reflex of an analyst is to say there is nothing. But the story does not end there. The repeated N/A is itself information. It tells you the source does not publish wind readings, does not publish shoe specifications, does not publish split times, does not publish competition dates. In an athletics dossier, an empty space always means something. The question is what.

Where an Athletics Dossier Is Actually Made

Athletics generates data through a supply chain, and that chain has many joints that can snap. A race begins with the starter's pistol, passes through the electronic timing system, the wind gauge, the panel of officials, and only then becomes a results sheet. From that sheet, data moves to the national federation, and from the national federation to the World Athletics database. Every handover is a chance to lose a detail.

The wind reading is the first detail to vanish. Anemometers are only used in certain events: the 100 metres, the 200 metres, the hurdles, the long jump, and the triple jump. In the 100 metres the device measures for ten seconds from the gun. In the 200 metres it measures for ten seconds from the moment the leading athlete enters the home straight. In the horizontal jumps it measures for five seconds of the approach run. Outside those events, wind does not exist in the dossier, even when it exists on the track.

Splits are the second detail to vanish. Modern timing systems can record 50- or 100-metre segments in the sprints, 200- or 400-metre segments in the middle distances, and kilometre segments in the long distances. But most small meets never publish that data. They publish the final time, because the final time is the only thing spectators read.

Equipment specifications are the third. Since January 2026, World Athletics has regulated sole thickness on the track, capped at twenty millimetres for events up to 800 metres and twenty-five millimetres for longer events, with limits on the number of rigid plates inside the sole. On the roads the cap is forty millimetres. But who inspects, when, and whether the inspection result is ever published is another matter entirely.

And then there is the date. A report with no date cannot tell you where an athlete sits in a training cycle, how many weeks into injury recovery they are, or how high above sea level the race took place.

Based on my experience covering domestic Japanese meets, the corporate league system here publishes generously: splits, wind readings, coach names, training camp schedules. When I turn to competitions in Southeast Asia, the level of disclosure drops sharply. Most domestic meets in the region publish only a final time and a placing. No wind, no splits, no equipment, no official names.

That gap is not a gap in athlete quality. It is a gap in recording infrastructure. And it produces a very concrete professional consequence: an analyst working with Southeast Asian data must learn to reason from absences more than from presences.

Five Red Flags Inside an Athletics Dossier

When a dossier reaches me containing nothing but a final time and an athlete name, I run a five-item checklist. These five items are not disqualifying conditions. They are five questions that tell me what I am missing.

Red flag one: wind and altitude

World Athletics sets the record-eligibility threshold at a tailwind no greater than two metres per second. A sub-ten-second 100 metres run with a three-metres-per-second tailwind is still a fast run, but it does not measure the same quantity as a 9.95 in still air.

I spent a long time with the case of Abdul Hakim Sani Brown, who reached 9.97 seconds in Austin in 2026. At the time, Japanese analysts split into two camps. The first treated it as proof that Japan had entered the sub-ten era. The second asked one question: what was the wind. That question carries no scepticism. It is a mandatory technical question.

Altitude is wind's sibling variable. Thin air at elevation reduces drag. Bob Beamon's 8.90 metres in Mexico City in 2026 came at more than two thousand two hundred metres above sea level. It is one of the greatest moments in athletics history and, simultaneously, a performance that benefited from physical environment. Both things are true.

What makes this red flag dangerous is that it leaves no trace when it is skipped. A results sheet missing a wind reading still looks tidy, still looks professional, still looks complete. The reader does not know what is missing, so has no reason to doubt.

Red flag two: the equipment dividend

Over the past three decades, running equipment has become a measurable variable. Eliud Kipchoge's marathon world record of 2:01:09 in Berlin in 2026 and Kelvin Kiptum's 2:00:35 in Chicago in 2026 were both produced in a generation of thick-soled shoes with rigid plates. Tracks changed too. The Tokyo 2026 Olympic surface was engineered to return more energy to athletes, and it produced a wave of personal bests across multiple events within days.

This creates a concrete accounting problem. When an athlete runs half a second faster than their own previous best, I need to separate how much came from the legs and how much came from the material underneath. If I cannot, I am comparing two different eras and calling them by the same name.

On the track, shoe rules are tighter than on the roads, precisely because administrators understand that an equipment advantage can be large enough to reorder a final. But rules only work when someone inspects and someone publishes the inspection. At meets that do not publish equipment data, a performance can still be ratified; it is simply ratified with a level of uncertainty recorded nowhere.

Red flag three: small sample

A single mark is not a level. This is the sentence I have to remind myself of most often, because a single mark is always the most impressive thing on the page.

Suppose an athlete runs the 400 metres two seconds faster than their previous best once, then runs two seconds slower on the next three attempts. Statistically, the fast mark is an outlier outside the distribution. In media terms, it is a headline. The distance between those two views is the distance between data and story.

My approach is to count repetitions under identical conditions. Same event, same phase of season, same surface, same opponents. If a dossier does not support that count, I write explicitly that the sample is too small. Writing that does not weaken the analysis. It makes it honest.

At the 32nd SEA Games in Phnom Penh in 2026, Nguyen Thi Oanh won four gold medals across four different events, some of them scheduled very close together. This is a case where the small-sample red flag arrives together with a competition-conditions red flag. An athlete contesting multiple events in a short window does not mean her marks are worth less. It means each mark must be read alongside the physical cost already spent, and that cost is almost never entered into the dossier.

Red flag four: unratified training marks

There is a category of information that appears regularly in sports pages and never in official databases: marks achieved in training. An athlete is said to have run sub-ten in practice, jumped over eight metres in practice, run a kilometre faster than the national record in practice.

Those numbers cannot be verified. No wind gauge, no officials, no certified timing, no published measurement conditions. Technically they belong to a different frame of reference from competition marks. Psychologically they carry enormous weight, because they suggest what is about to happen.

I do not throw this category away. I classify it. In my notes, training marks sit in the soft-signal bucket, alongside internal injury whispers and coaching changes. Soft signals can point a direction, but they cannot be used to compute probabilities. Once a soft signal slips into a spreadsheet and is treated as hard data, every model behind it loses value.

Red flag five: missing splits

The final time is a sum. A sum does not reveal internal structure. Two athletes who both run 1:46 for the 800 metres may be telling completely different stories: one went out fast and faded, one went out slow and accelerated over the final two hundred. To a coach, those are two different diagnoses. To an analyst, two different probability distributions for the next race.

When I built my own dataset on Cerezo Osaka's pressing behaviour in the 2026 season, no official data existed on how many passes opponents completed before being closed down. I watched video and hand-counted one thousand two hundred and forty pressing situations. That work took hundreds of hours, and it taught me something no classroom could: when data does not exist, there are two options, either create it yourself or admit you do not have it.

In athletics, creating split data by hand is far harder than in football, because the speeds exceed what the eye can read. That does not mean ignoring it. It means that when a dossier lacks splits, every tactical conclusion must be downgraded to a hypothesis.

The N/A space as a signal

Back to the report with nine N/A lines.

After reading it a third time, I realised it revealed more than it appeared to. A document with a complete structure, nine sections and six risk categories, shows that whoever built it had a process. That process exists, is named, is graded. But the input data for that process is entirely empty.

In my trade, a complete process running on empty data is more common than people assume. It happens when a federation publishes later than the moment analysis is needed, when event names fail to match across sources, when athlete names are transliterated differently everywhere, when a meet is postponed and schedules are not updated across every database.

My way of handling this is to record the gap itself as its own column. That column is called: reason for missing data. If the reason is that the organiser has not published yet, I wait. If the reason is that the event never publishes that class of data, I adjust the research question. If the reason is that nobody measured it, I write it into the limitations section.

I collect mistakes, classify them, and then I know where the team is heading. For athletics, that becomes: I collect gaps, classify them, and then I know how far I am permitted to conclude.

Track, Stadium, and the Unmeasurable Noise

There is one variable no analysis sheet captures, and it is present in every race: noise.

In 2026, when the pandemic suspended the J-League for four months, I sat in Osaka and re-analysed old Cerezo matches. I built a model from pressing data and predicted the team would drop in form when the league resumed, because the absence of home crowds would reduce pressing intensity. At season's end Cerezo finished fourth, below my predicted second.

My model was wrong, and I could not blame luck. I traced the inputs and found I had ignored a variable: crowd effects do not run only in one direction. Empty stadiums also reduce the intensity of away teams, reduce psychological pressure, and reduce referee error in favour of the home side. Net of everything, the effect was far smaller than I assumed.

An empty stadium, yet the numbers are still full of noise. Pressure, expectation, officiating error, the roar of a stand: none of it appears on a results sheet, and none of it is noise to be discarded. It is a variable to be modelled.

In athletics this variable takes another form. An athlete competing at home, before a full stand, tends to start faster over the first twenty metres and pay for it over the last twenty. In an empty stadium, that tendency disappears. In a dossier containing only a final time, the two situations are indistinguishable.

I sat in Tokyo during the 2026 Olympics and watched one of the strangest athletics meetings in history: a vast stadium, a fast track, and almost no human sound. Multiple world records fell in that atmosphere. When I analysed them, I had to attach a caveat to every model from that season: no-crowd context, not directly transferable to a meet with spectators.

The Counter-Intuitive Angle: Excessive Data Discipline Is Its Own Error

The five red flags above sound entirely reasonable, which is exactly why I have to write the defence for the opposite direction.

If I applied those five flags strictly, I would eliminate nearly all athletics data from Southeast Asia. No wind readings, no splits, no equipment data, no reliable dates in some sources. Push data discipline to its extreme and the conclusion is always the same: insufficient information, cannot assess.

An analyst who only says that has stopped being an analyst. They have become a file checker.

The real value of a screening process lies in distinguishing degrees of uncertainty, not in erasing everything imperfect. A mark without a wind reading is a mark of medium confidence. A mark without a wind reading and without splits is low confidence. A mark without wind, splits and source should not enter any model at all. Those three tiers are different, and collapsing them into one is an error with practical consequences.

There is a second, more dangerous trap: manufacturing counter-intuitiveness. When a piece earns praise for concluding against the crowd, the writer acquires an incentive to hunt for the contrarian conclusion first and then look for data to support it. I know that incentive well, because I once wrote that way.

In 2026, at seventeen, I logged every Japan match at the World Cup in Russia. Against Belgium, Japan held around fifty-five percent possession but touched the ball inside the opponent's penalty area only seven times, against twenty-one for Belgium. I wrote a blog post, grounded in those figures, arguing that pushing the defensive line high in the closing minutes was a tactical error. A group of supporters attacked the piece hard.

On that night in Russia in 2026, I watched data shatter in front of me. The lesson was not that I should stay silent. The lesson was that I had presented a probabilistic proposition as a moral one. Saying a decision carries a higher probability of loss is not the same as saying the decision-maker was wrong. Those are different statements, and in a piece written for a mass audience, the difference usually gets erased.

The third trap concerns my own position directly: a Vietnamese analyst based in Japan, writing about athletics for both markets. The data standards of the Japanese corporate league system were built in an environment with high measurement infrastructure, large budgets and a long record-keeping tradition. Applying that full standard to a meet in Vietnam produces two effects at once: it devalues genuinely good performances, and it obscures the problems that genuinely need addressing.

The real problem in Vietnamese athletics is not that athletes run slower than world standards. The problem is that a good performance leaves too few traces to be analysed, compared or replicated. Nguyen Thi Oanh ran four events and won four gold medals, and most of what survives in the public record is a time and a placing. No splits, no pacing strategy, no recovery data between events. An analyst trying to learn from that medal can only learn half of it.

The fourth trap is cultural. Southeast Asian fans follow sport through a different frame of reference than the one Western models assume. They care about the athlete as a person carrying a story, and factors that look irrational to a model, loyalty, social pressure, family expectation, carry real weight in competitive and transfer decisions. A model that cannot measure those things should not behave as though they do not exist.

I once wrote about set pieces after Euro 2026, when Denmark scored four of their six goals from designed routines, against a tournament average around twenty-eight percent. I compared that with RB Leipzig's data in the 2026-21 Bundesliga, during Julian Nagelsmann's tenure, where set-piece drills were designed from running-position and ball-landing data. The piece was republished by a small football site and drew around fifteen thousand reads.

What I did not write then, and now regret, was a paragraph on the limits of the comparison. A set-piece routine that works in the Bundesliga does not automatically transfer to a league with lower average height, worse pitches and fewer training sessions. Cross-league comparison is only valid when the writer states how many other variables are being ignored.

What Remains When the Spreadsheet Closes

Across nine years of watching tracks and matches, the thing I believe most firmly is that every model will eventually be wrong. Every probability hides a shock; my job is only to make sure it does not repeat. The only way to secure that is to record in full why the model failed, rather than to record how the failure felt.

An empty athletics dossier does not frighten me. What frightens me is a dossier that looks full. A sheet with an athlete name, a time, a placing, twelve numeric columns and not a single footnote about measurement conditions. That kind of dossier is far more dangerous than a page reading N/A, because it manufactures a false sense of safety.

My working rules now fit in three sentences. Each analysis may use only three primary metrics, and each must answer one specific question. Every conclusion must carry its uncertainty level. And every dossier with missing data must be logged as its own line, even when that line exists only to say the thing you needed does not exist.

With the regular season entering its closing stretch, I am tracking three signals. The first is the reappearance of wind readings in regional reporting, because that indicates measurement infrastructure is improving. The second is the publication frequency of splits at domestic Southeast Asian meets, because that is the earliest indicator that analytical quality will change within two to three years. The third is the number of times a training mark reaches the sports pages without measurement conditions, because that indicates the market is missing a standard.

None of those three signals appears on any race results sheet. They sit at the margins, in small footnotes, in the columns an organiser chooses not to fill. My job, from the beginning, has been to read the margins before the middle.

The report with nine N/A lines is still in my folder called empty records. I have not deleted it. It reminds me that in this trade the most honest answer is sometimes three characters long, and that daring to write those three characters instead of inventing a beautiful story is the entire difference between an analyst and a storyteller.

As for the season now underway, I will keep recording the wind reading before the time. Not because the wind matters more than the time, but because the time always has someone to record it, and the wind does not.