Vietnamese Athletics Is Missing Its Most Important Data Column Ahead of the New Season

Eleven at night, I reopened the results file of a domestic athletics meet that had just closed. The file had four columns: name, unit, mark, place. Enough to hand out medals. Not enough to do anything else. I normally work with data for a football team, so I am used to opening a match and finding hundreds of rows: expected goals, passes allowed per defensive action, distance covered split into 15-minute blocks. On the track it is the reverse. A 400m hurdles race is recorded as a single number. That number does not tell me the rhythm of the first 200 metres, does not tell me what share of top speed was lost at the eighth hurdle, and does not tell me which way the wind blew. When numbers speak, I only have to listen. What I was holding was a number with its tongue cut out. That gap is the subject of this article. Vietnam does not lack achievements to report. At the 31st SEA Games in Hanoi in 2026, Nguyen Thi Oanh won four gold medals, spanning the 1,500m, the 5,000m, the 3,000m steeplechase and the mixed 4x400m relay. Four years earlier, at the 2026 Asian Games, Quach Thi Lan won the women's 400m hurdles in under 56 seconds, and Bui Thi Thu Thao took the women's long jump. Race walking has also carried Vietnamese athletes to podiums at continental and regional level. But what remains after moments like those is mostly memory plus a few short news lines. A modern athletics system runs on a multi-layered data structure. The lowest layer is the raw result. Above it sit the condition variables: wind speed, venue altitude, temperature, humidity, footwear specification. Above that sits the split data, segmented every 100 metres. The top layer is the year-by-year personal-best series, long enough to be drawn as a curve. Vietnamese athletics stands firmly on the lowest layer and is close to empty on the other three. That is why serious analysis of domestic running ends with the same helpless line: not enough data to conclude. What matters is that this shortfall is systemic, not the accident of a few organising committees. A wind column barely exists in the official protocol. World Athletics rules require sprint and jump marks to carry a wind reading, and a mark can only be ratified as a record when the tailwind does not exceed 2.0 metres per second. That threshold sounds small, yet its effect is not: the gap between a 1.5 headwind and a 1.5 tailwind over 100 metres can approach three tenths of a second, which is the entire distance between gold and fourth at a SEA Games. At many domestic meets the wind column is either not measured or measured and never published. The result is a ranking table blind to weather. Comparing marks across two meets without wind adjustment is the worst kind of analytical error: the silent kind. Split data is what turns a race into a story. An 800m runner can go out too fast and collapse over the final 200 metres, or run evenly and kick over the last 150. Both produce the same finishing time, and only split data separates them. Without it, the analyst is forced to speculate, and speculation about running is the shortest route to a wrong conclusion. Distance never lies; we are simply not patient enough to listen. The personal-best progression curve is the most valuable lens available and the most absent. An athlete who improves steadily across five years, gaining a few tenths each season, tells a completely different story from one who suddenly improves by two seconds in a single season. In athletics analysis, a jump of roughly three times the historical annual gain is a signal to investigate, not a verdict. To run that check, I need the year-by-year series. Our record tables usually keep only the personal best, the single highest point of a career. A personal best is one pixel; the whole climb to the summit disappears. The dual consequence matters here: missing data does not equal clean. In athletics, an empty return is non-information, not a certificate. No one is exonerated by the silence of a database, and no one is convicted by it. Both directions require numbers. Competition density and peaking cycles also vanish from the record. The SEA Games fall in May, the national championships are scattered around it, and an athlete may have raced five or seven times before the heats. That count, the gaps between races, the number of taper days before the peak, all are variables that act directly on the final result. Without them I cannot separate a genuine peak from a lucky run on a fine day. I do not believe in luck; I believe in what has been repeated enough times. The final layer is metadata about the training environment. Where the athlete trains, under which training centre, whether they have been at altitude, whether they run on asphalt or a synthetic track, whether the shoes carry a carbon plate. A carbon plate delivers a measurable speed dividend, and so does a freshly laid track. Unless that dividend is subtracted, we place a 2026 mark beside a 2026 mark as if they shared one frame of reference, while the track and the shoes have changed under the athlete's feet. Apply this to one concrete case. Nguyen Thi Oanh's four gold medals at the 31st SEA Games were among the most impressive performances by any Vietnamese athlete in the past decade. But ask why she withstood that density while most rivals could not, and the public data cannot answer. There are no per-lap splits, no wind readings per event, no notes on training load across the eight weeks before the meet. We know the result and very little about the cause. It is an uncomfortable paradox: the more medals are won, the wider the data gap becomes. At this point the natural reflex is to call for collecting much more data. I think that framing is wrong. The bottleneck in Vietnamese athletics is standardisation, not volume. A meet that measures wind and a meet that does not produce two datasets that cannot talk to each other. Splits available in a final and absent in the heats make every comparison between rounds meaningless. In statistics, non-uniform data is worse than scarce data, because it manufactures false confidence. Forced to choose, I would rather have one minimal standard enforced at every meet than ten meets with ten different formats. It is also worth stating plainly where correlation ends. An athlete who changes training centre and then runs faster has not proved that the new centre produced the performance. Many other things changed at once: age, base fitness, opponent quality, competition conditions, luck included. Every number is a confession the race cannot deny, but that confession is only trustworthy when we know the conditions in which it was recorded. Mistaking correlation for causation is the most common error in sports data work, and it is especially dangerous in athletics, where every conclusion rests on a short time series. The way forward is fairly clear and does not require much money. Three columns added to every domestic result protocol: wind speed, split times, and the track name with weather conditions. A multi-year performance database for the priority athlete group, updated after each season. And a simple convention for publishing training data at a raw level, with no need for detailed session plans. Do those three things and every future analysis changes in kind. I still keep the habit of rereading a result protocol before sleeping, even when it holds only four columns. That night I stopped at the last sheet sent in by a young coach. He had written out his athlete's lap times from the final training session by hand, in ballpoint, on the back of a photocopy. Cruder than any electronic board, and more complete than everything published across the entire season. Sometimes the best data in Vietnamese athletics sits in the pocket of a man with a pen, waiting for someone to read it seriously.