The N/A Report and 23 Lost Days: Vietnam Swimming Injury-Analysis System Failure 2035

On March 12, 2035, at the PVF Sports Medicine Centre, I opened a 47-page PDF and stopped at page 14. It was the injury-analysis report for Vietnam's national swimming team, preparing for the 2035 SEA Games in Bangkok. The report was generated by AquaTrack 3.0, complete with a digital signature, an integrity-check code, heat maps, and colour-coded tables. On page 14, under "Shoulder Load Index", eight of twelve athletes were marked "N/A". No numbers. No red flags. No error notes. Just empty cells, carefully formatted, right-aligned, in a uniform font. Three weeks later, Nguyen Thi Mai Linh, 19, a 200m butterfly swimmer, stopped training with shoulder tendinitis. Four days after that, Tran Van Duc, 21, a 400m freestyle swimmer, abandoned a fitness test with knee pain. Two injuries. Twenty-three lost training days. And a question no one on the coaching staff asked when they read the report: what does "N/A" mean? In January 2035, the Vietnam Swimming Federation signed a contract with a domestic sports-technology company to deploy AquaTrack 3.0 for the national team. The system consisted of 36 pressure sensors worn on shoulders, knees, and wrists, eight infrared cameras mounted on the pool wall, and an AI module analysing training load in real time. The goal was clear: cut injuries by 30% in the 2035 SEA Games cycle. Budget: 4.7 billion dong. Deployment time: six weeks. The team had 12 athletes, five of them first-time SEA Games competitors. The federation target was eight swimming golds. Performance pressure pushed the coaching staff to increase training volume in March, the final sprint before the Games. I was brought in as an independent expert on April 2, after Mai Linh and Van Duc left the pool deck. Across the 47 pages, I found three problems. First, the system had no rule for rejecting empty data. When a sensor lost signal for more than 45 seconds, AquaTrack recorded a "null" value instead of raising an error. The AI module, instead of halting the analysis, kept generating the report with fully formatted "N/A" cells. Second, the user interface did not distinguish between "no data" and "normal data". Both were pale green. Only by clicking each cell would a reader see the small "N/A" in the lower corner. Third, the coaching staff had not been trained to read technical reports. They had been told that "green means safe". And the eight "N/A" cells were all green. This was a pipeline failure, not a data failure. The system did exactly what it was programmed to do: receive signals, process them, generate a report. But it lacked a safety valve, a hard rule requiring the report to be withheld if any core data field was empty. In injury analysis, "no data" does not mean "no risk". It means "risk not assessed". And at the elite level, "not assessed" is more dangerous than "at risk", because it triggers no preventive response at all. I cross-checked all 12 athlete files. Eight had empty reports. Of those eight, six had increased training volume by more than 20% in the preceding two weeks, a standard warning threshold in sports medicine. The other two had prior shoulder or knee injuries. Mai Linh was in the first group. She had raised her butterfly volume from 4,200m to 5,100m per day in the second week of March. Her report showed "N/A" on all three indices: shoulder load, wrist load, and recovery score. Three empty cells. No warning. Van Duc was in the second group. He had a history of knee pain dating to 2033, recorded in his medical file. The system did not read medical files; that was a separate module, not yet connected. His report showed "N/A" on the knee-load index. No warning. Hoang Thi Thu Ha, 17, a 100m backstroke swimmer, was also among the eight with empty reports. She was lucky not to be injured, but her training volume had risen 18% in two weeks, below the warning threshold, yet still a signal worth tracking. The figure of 23 lost days is not large compared with mass-injury episodes. In 2026, when I started at Hai Phong FC, I logged 127 injuries across 43 players in a single season. At Lach Tray, I learned to read injuries from the first numbers. In 2026, after the pandemic, hamstring injuries in the V.League rose 40%. In 2026, at the World Cup, there were 31 muscle injuries across 48 group-stage matches. But these 23 days had a different feature: they were predictable. And they were not predicted. In injury analysis there are three kinds of failure. The first is failure from missing data: no sensors, no reports, nothing to read. The second is failure from wrong data: broken sensors, input errors, miscalculation. The third is failure from empty data presented as complete data, and this is the most dangerous, because it creates a false sense of safety. AquaTrack 3.0 belonged to the third kind. I reviewed the system logs. On March 8 at 14:23, Mai Linh's shoulder sensor lost signal for 6 minutes 12 seconds. The system logged "null". The AI module, instead of stopping, interpolated from the previous day's data and assigned "N/A" to the entire series. The report was still issued at 18:00 that day. On March 9, Van Duc's knee sensor lost signal for 4 minutes 48 seconds. Same mechanism. Same result. The coaching staff read the report at 19:30. They saw green. They did not click each cell. They kept training. This is the blind spot of sports digitalisation: the more automated the system, the less humans check. And when the system fails silently, humans fail with it. Numbers stay silent, but their sequence always knows how to tell a story. I compared two similar systems. The US Swimming Federation system, deployed in 2031, has a two-layer rule: the first layer checks data integrity, the second checks the plausibility of the result. If either layer fails, the report is withheld. Australia's system adds a step: a human expert must sign off before the report reaches the coaching staff. Both systems have empty-report rates below 0.5%. AquaTrack 3.0 had a rate of 66.7% in March 2035. The comparison is not meant to criticise Vietnamese technology. It is meant to show that a simple validation rule can prevent a complex chain of failure. And in sport, where every lost training day is worth a competition slot, 23 days is not a small number. In my report, I proposed three changes. First, add a hard validation rule: if any core field is empty, the system must refuse to issue the report and send an alert to the person in charge. Second, change the interface: "N/A" must be red or amber, not pale green. Third, retrain the coaching staff in reading technical reports, with an exit test. The Vietnam Swimming Federation accepted all three. AquaTrack 3.1 will be deployed in June 2035, four months before the SEA Games. But there is a larger question: how many other empty reports exist in the system? No one knows. Because the system has no cross-report checking function. Each report is an independent file. And no one thought to check. The coaching staff's first reaction was to blame the technology. "The system failed," they said. "We are not engineers." But the technology did not fail. It did exactly what it was programmed to do. The failure was in the process: a system without a safety valve, an interface that did not distinguish risk, and insufficient training. In nineteen years of watching the industry, I have seen a recurring pattern: whenever something goes wrong, people blame the tool. In 2026, when hamstring injuries rose 40% after the pandemic, they blamed the schedule. In 2026, when 31 muscle injuries struck the World Cup, they blamed the weather. In 2035, when eight reports came back empty, they blamed AI. But the tool is never the root cause. The root cause is always human: system design without checks, report reading without challenge, decision-making without cross-referencing. The body is a closed system, but data is the key that opens it. More worrying is the public reaction. When news of the incident spread, many said: "See, AI can't be trusted." That is the wrong conclusion. AI can be trusted if designed correctly. The problem is that we trusted it too early, without testing it thoroughly enough. In sport there is an unwritten principle: never trust a single metric. Always have at least two data sources. Always have at least one cross-check. Always have at least one person asking questions. AquaTrack 3.0 violated all three. On April 15, 2035, Mai Linh returned to the pool deck with her shoulder strapped. She will make the SEA Games. Van Duc needs two more weeks. In the final report to the federation, I wrote a single sentence in the conclusion: "A system that cannot say 'I don't know' is a dangerous system." The question for those building sports-analysis systems: have you designed the safety valve? And if not, will you know when it is needed?