Swimming and the Data Voids: From Rome 2026 to the Empty Pool of Tokyo

Summer 2026, at the Foro Italico in Rome, a scoreboard lit up with a string of numbers that silenced the stands. By the end of the session, organizers counted dozens of world records erased within days. A men's 200m freestyle record that experts believed would stand for at least a decade was pushed down after a few heats. I sat in front of a screen, logging every split, every reaction time, every underwater distance after the start, and asked myself a question that would haunt my entire analytical career: was that the achievement of a human being, or the achievement of a suit? That question has no clean answer. And precisely because it has no clean answer, it became the first lesson in how I read swimming to this day. When data is empty, when a model no longer holds, the analyst must choose: either invent a neat conclusion, or admit that he stands before an unreadable zone. I chose the second path, and I chose to say so plainly. I once treated models as scripture. Now they are only a compass — but without one, we get lost. Swimming is the sport where every hundredth of a second can be reduced to a calculation, and also the sport where that same hundredth regularly betrays the calculation. This article does not aim to praise data, nor to dismiss it. It aims to dissect a paradox: we have more numbers about the lane than ever before, and we understand the lane less than we think. CONTEXT: WHEN THE LANE BECOMES A DATA PROBLEM Swimming enjoys the greatest analytical advantage of any indirectly contested sport. No opponent touches you, no referee penalizes you for contact, no complex defensive tactics. You swim in still water, lane length is measured to the centimeter, and time is recorded to the hundredth. Everything seems designed to produce perfect data. But that very cleanliness creates a trap. Because swimming data looks so tidy, people easily believe it says everything. In reality, a race time is only the endpoint of a chain of decisions unfolding over tens of seconds to minutes, and most of that chain never appears on the results sheet. Take a basic example. An athlete swims 100m freestyle in 48 seconds. That number can be split into layers: reaction time at the start, the distance and speed of the underwater glide, stroke count, stroke rate, distance per stroke, turn time, post-turn time, and speed over the final 15 meters. Each layer responds to a different factor. A high stroke rate can signal endurance, or it can signal panic when fading. When I began reading swimming seriously, I had to admit something: the results sheet tells me what happened, not why. This is why I value splits above final times. An athlete who swims 400m with a faster back half than front half tells a completely different story from one who posts the same time but collapses over the last 50 meters. Two identical medals on the board, two entirely different bodies. After the 2026 World Cup, I learned something I carried into swimming: numbers do not lie, but people always find ways to lie with numbers. In swimming, the most subtle way to lie with numbers is to compare things that cannot be compared. A record set in a 25m short-course pool cannot be placed beside a 50m long-course record as if they were the same. Short-course racing involves roughly twice as many turns, and every turn is a powerful push-off — meaning the speed advantage comes from turning technique, not from swimming strength. Placing two numbers side by side without specifying the pool is a kind of lying by layout. I usually divide swimming data into three layers. The first is time — public, visible, easy to compare, and easiest to misread. The second is race structure — 50m splits, stroke rate, cycle length — which only emerges when you bother to rewatch video and cross-check it against the official timing sheet. The third is the submerged layer: physical condition, competition schedule, unhealed injuries, psychological pressure. The third layer is almost never fully disclosed, and it decides most of what happens in the first. In that third layer, I stand before an unreadable zone of data. And I learned that admitting I cannot read it is part of the job, not a failure of the job. CORE ANALYSIS: THE TECH-SUIT ERA AND THE LESSON OF ROME 2026 To understand why Rome 2026 is a landmark, we must recall the technical context of that era. In the late 2000s, manufacturers released full-body suits made from polyurethane and elastic polymers. These suits compressed the body, reduced drag, and increased buoyancy. Mechanically, this was a direct intervention into performance: you did not need to swim better, you only needed to be less impeded by water. The consequence? In a short period, a wave of world records fell with margins far beyond swimming's natural rate of progress. Records that should have lasted ten years were erased in months. When FINA issued its ban on tech suits, effective from the start of 2026, the sport entered what is commonly called the textile era. This is the point where data analysis must be extremely careful, and where I believe many readers of record tables have misunderstood. You cannot treat a record set in 2026 and a record set in 2026 as the same kind of evidence. They were born under two different technical regimes, two different rule frameworks, and two different psychological expectations. Placing them side by side without a footnote is a methodological error, even if it looks fair. I remember once arguing with a colleague. He said simply: time is time, faster is faster. I countered with a question: if tomorrow the rules allowed fins, would an athlete breaking a world record mean the lane had improved? He went quiet. That is exactly the problem of Rome 2026, except the technology looked more like a suit than like fins. After 2026, analysts had to rebuild the entire frame of reference. Every new record needed an era label. Every cross-era comparison needed a note on conditions. And most importantly: an athlete's value cannot be measured solely by the number of records broken, because records are a quantity dependent on the rules of the moment in which they swim. From the Rome lesson, I drew a working principle applicable to any sport: before comparing two numbers, confirm they were produced under the same rule, the same measuring tool, and the same definition of performance. If any of the three is missing, the number is only raw data, not yet evidence. Also from Rome, I began to distrust a common habit of sports media: calling every broken record "history." A record broken under changed equipment is the history of the equipment first, and the history of the person second. That order matters. RACE STRUCTURE: WHERE DATA SPEAKS MORE TRULY THAN TIME If final times easily mislead, race structure is where data becomes more honest. I spend most of my analytical time on 50m splits and technical indicators, because that is where an athlete reveals how they win or lose. Start with the start. In swimming, the start and the underwater glide matter far more than spectators feel. In freestyle and backstroke, the rules cap underwater travel at 15 meters. That means an athlete can cover nearly a third of the pool without surfacing to swim. Over that distance, they move on the push of the legs after the start, and speed in this phase is usually higher than normal swimming speed. This creates an optimization problem. A longer glide gains on drag, but if you glide too long, you lose momentum and must resume from a slow state. A shorter glide preserves rhythm but costs more energy over the swim. There is no universal formula. Each athlete has an optimum depending on leg drive, height, and breath-holding ability. And here video analysis becomes more important than the timing sheet: you must see at which meter the athlete surfaces to understand the strategy chosen. I often log the surfacing points of the same athlete across meets, and that change usually tells a story about injury or training adjustment. An athlete who once glided to 14 meters suddenly surfacing at 11 meters for months is a signal. Perhaps back pain. Perhaps shoulder pain. Perhaps the coaching staff changed strategy due to a dense schedule. The timing sheet does not tell you that. Video does. Next come the turn and the touch. In a 50m pool, a 1500m swimmer turns 29 times before the final touch. Every turn is a chance to lose or gain hundredths. Multiplied by 29, the gap can reach seconds. An athlete who turns half a hundredth slower than a rival each time loses more than a second over the race, and one second over 1500m is the difference between gold and nothing. This is why I always separate "swimming" performance from "turning" performance. The two require different skills, different training, and often reflect different problems. An athlete who swims fast but turns poorly has clear technical potential. One who swims slowly but turns well has good conditioning and spatial awareness, and is missing strength. On the timing sheet, these two can look alike. On splits, they are worlds apart. Finally, the front-half and back-half structure. I classify athletes into three types: steady over both halves, fast in the front half, and fast in the back half. The third type — the negative split — is often praised, but be careful. An athlete who is faster in the back half may be a good controller, or may be a slow starter benefiting from the acceleration of rivals. To tell them apart, you must cross-check against the field's pace and against their own previous-round top speed. Here, swimming analysis meets a limitation worth noting. Swimming is a sport where you race parallel to rivals, not directly against them. Your speed is independent of your rival's speed, at least in theory. That makes reading tactics from outside harder than in contact sports. You do not see an athlete "defending" or "attacking" in any clear sense. You only see a sequence of speeds over time, and must infer the intent behind that sequence. Based on my experience watching matches and swim meets, I find that most errors in swimming prediction models come from ignoring the psychological variable. An athlete who swims a superb morning heat but fades in the evening final may simply be unable to hold psychological energy for the moment of greatest value. The model cannot measure that. The coach at the pool's edge can roughly gauge it with the eye, and is often more right than the model. CAREER SHAPE: THE CURVE DATA NEVER FINISHES DRAWING One of the most common mistakes when reading a swimmer is to place them on a single performance curve and forecast the future from its slope. In reality, a swimmer's career usually follows different curves for each event, and those curves can run in opposite directions. Think of a female swimmer racing from 200m to 800m. In her teens, she may shine early in shorter events thanks to strength that outpaces her peers. When puberty arrives, the body changes in proportion, fat, and center of mass, making previously automatic skills suddenly no longer automatic. This is the puberty barrier. For many female swimmers, this is the most dangerous phase of a young career, and also the phase where "next big thing" prophecies fail most. Over long distances, the picture is different. Performance at 800m and 1500m often keeps improving into adulthood because it depends more on an aerobic base than on pure strength. A distance swimmer may peak years later than a sprinter. This means a meet where distance swimmers succeed is often a meet with a different age structure than a sprint meet. This is why I am extremely cautious when someone asks me to predict a young athlete. I usually reply with three questions before a conclusion. First, which event do they swim, and how does that event respond to age? Second, have they passed the growth phase of puberty, or is it ahead? Third, have they competed at the highest level often enough to expose their psychological limits? If any of the three is missing, any prediction is blind probability. I have heard many prophecies like "this athlete will break the record within two years." Looking back, the hit rate of such prophecies is far lower than the general feeling. Not because prediction is hard, but because people predict with one curve instead of three overlapping curves. And there is one variable data almost never captures: the decision to quit. Some athletes have every indicator pointing to the top and still stop — because of injury, money, psychology, or a life beyond the water. You cannot predict such a decision with a model. You can only remember that a person is not the sum of their indicators. When the stands are empty, every model collapses. I rebuild from the burnt data. THE CONTRARIAN ANGLE: WHEN CORRELATION IS NOT CAUSATION This is the part I consider the most important of this entire piece, and also the part most easily skipped when people get excited about data. In swimming, there are countless beautiful correlations. Faster swimmers often have a certain stroke rate. Champions often have above-average height. Swimmers who hold back-half speed often have lower lactate levels. All of this can be verified with numbers, and all of it risks being misread as a causal relationship. Multiplying a correlation does not mean finding a cause. Some fast swimmers having high stroke rates does not mean raising stroke rate will make you swim faster. It may be that high swim speed causes stroke rate to rise, not the reverse. It may be that both are the result of a third factor — say, a strength base — that you have not measured. This is the classic error of sports analysis, and swimming is not immune. I have seen training programs built on a single correlation. A champion has a certain metric X, so young athletes are pushed to focus on metric X. The result is usually young athletes achieving metric X and still swimming slower, because metric X is only a trace of a larger base, not the base itself. At a deeper level, I believe swimming needs a more humble reading of data: treat data as a tool for scoping hypotheses, not as a tool for issuing conclusions. Data tells you where to look. It does not tell you what you will see. The looking must still be done by humans. There is another type of error worth mentioning: using a small sample to conclude a large rule. In swimming, elite meets are infrequent, and each athlete appears at the top only a few times a year. So your sample is very small. An athlete who swims twice at the top and wins both proves nothing about their mental strength. It only proves you have two data points. In my own writing, I try to state the sample size whenever I make a claim about a trend. If the sample is small, I use the word "hypothesis." If it is large enough and cross-checked across meets, I use the word "trend." This is language discipline, and I believe it matters no less than numeric discipline. And here the story returns to the starting point of this piece. There are times when the data sheet is entirely empty. Not because there was no race, but because data was not recorded, not disclosed, or lost during collection. When that happens, the correct response is not to fill the gap with speculation, but to mark the gap and say plainly: this zone is unreadable. In the analytical profession, the hardest skill is not reading a full table. The hardest skill is looking at an empty table and not inventing content for it. VIETNAMESE SWIMMING: WHAT CAN BE READ FROM WHAT EXISTS Bringing those principles into the Vietnamese context, the picture becomes interesting in a different way. Vietnamese swimming has a few athletes who made their mark at the regional level and gradually pushed toward the continent, but the public database on them is far thinner than in swimming powers. Nguyen Thi Anh Vien is the most analytically interesting case of her generation. Born in 2026 in Can Tho, she emerged very early and became the backbone of Vietnamese swimming at SEA Games across several editions. What stands out about her is not only the medal count, but the way she was trained in a period focused on the physical base and multi-event technique. She raced many events, from individual medley to butterfly and freestyle, something very few elite swimmers dare to do given the physical and technical load. When analyzing the career of a multi-event swimmer like that, the three-curve principle helps. You must separate performance in each event, place them on separate curves, and only then speak about the whole. If you merge them all, you get a meaningless average curve that reflects no event at all. Nguyen Huy Hoang is another case, belonging to the distance group. Born in 2026 in Quang Binh, he rose in middle- and long-distance freestyle, collected results at the regional level, and gradually reached continental and world stages. For a distance swimmer, the development window often stays open longer, which means predictions about him must be placed in a longer timeframe than a sprinter's. Here I must say plainly what I consider honest: public data on Vietnamese swimmers at the level of splits and technical metrics remains very limited. We have results tables and final times, but lack detailed split data, force data, and published physiological data. That makes analysis at the second and third layers almost dependent on direct observation and interviews, rather than on numbers alone. This is why I recommend building an internal data system for Vietnamese swimming, logging 50m splits, stroke rate, cycle length, turn times, and related metrics at every meet. No need for expensive technology at first. It needs a disciplined, consistent, internally open record-keeping process. Over ten years, such a system could create a competitive advantage greater than any short-term training camp. I once told some coaches that Vietnamese swimming data currently resembles a map with a clear coastline but an almost empty interior. You know where you are on the coast, but you know nothing about the land inside. The only way to know is to walk in and record. A FEW NOTES ON THE COMPETITION SYSTEM AND ITS ASYMMETRY One factor readers often overlook is that the competition system shapes how we read data far more than we think. An athlete swimming at a domestic, regional, and world meet faces three completely different structures: number of rounds, rest between rounds, event density, and level of competition. At major meets, an athlete often swims the morning heats, then semifinals and finals in the evening, sometimes on the same day. So energy must be allocated across multiple starts, and that directly affects the ability to hold speed in the final. An athlete who swims heats too fast to "send a message" may pay in the final. One who swims heats just enough may save energy for the most important moment. At regional meets like the SEA Games, the structure is often different. Rounds may be fewer, event density may be heavier for multi-event athletes, and competition in some events may be significantly lower than at continental level. This means a very good SEA Games swim does not necessarily translate directly into an equivalent continental or world result. You must convert through the competitive context before comparing. This is a very common way of lying with numbers: comparing times across meets without adjusting for competition level. An athlete who swims slower at a highly competitive meet may actually have swum better than one who swam faster at a low-competition meet, if we account for the pushing effect rivals create. In swimming, rival pace has real influence. Racing beside someone faster can pull you up, especially in short distances. Conversely, swimming alone in the middle lane in a final where you dominate too early can cause you to decelerate over the final 15 meters because the competitive drive is gone. This is why some athletes tend to post better marks in meets with strong rivals. Data does not record this. It only records the final number. And so the final number, however accurate to the hundredth, can still mislead a reader about the true value of a performance. ON RULES AND THE GREY ZONES NO ONE WANTS TO DISCUSS Swimming has a detailed and strict rule system, governed internationally by World Aquatics, once known as FINA. The rules here define not only how to swim, but how to measure, how to assign lanes, how to handle rare situations. A few rules directly affect tactics. The 15-meter underwater cap in freestyle and backstroke forces swimmers to choose the optimal surfacing point, as noted earlier. In breaststroke, the single butterfly kick after each start and turn was permitted from a certain point in time, creating a technical opportunity that elite swimmers exploit. In butterfly and breaststroke, rules on the continuity and synchronicity of movement are strictly checked, and a small error can lead to disqualification. There are grey zones outsiders tend to miss. For example, the video review procedure at touches, the way underwater distance is determined, and the handling of situations where a swimmer is obstructed by a neighbor. These are rare but can affect the final result, and they are why reading only the results sheet without considering the officiating process can lead to wrong conclusions. One more point must be stated clearly: swimming is among the sports with the strictest anti-doping oversight, with rigorous in- and out-of-competition testing. Any information related to doping must be handled with a clear rule: separate a confirmed positive test, a contamination dispute, a procedural violation, and online speculation. These four categories have very different levels of certainty and must be presented differently. In every case, the silence of data is not evidence of a violation, nor is it evidence of compliance. Those are two symmetric errors I try to avoid. I have seen short commentaries concluding about an athlete based on unverified information. As someone working with data, I consider this the most serious error in the profession, because it is wrong not only in numbers but in ethics. A wrong number can be fixed by a right one. A damaged reputation is far harder to repair. WHAT REMAINS WHEN THE LANE IS EMPTY During the pandemic, many swim meets took place without spectators. This was a rare natural laboratory for testing the effect of crowds on performance — a variable that normally cannot be isolated from the others. In swimming, this effect may be even clearer than in other sports, because it is a sport where crowd sound can be altered by water noise and arena structure. In a pool with spectators, the roar of the crowd can change a swimmer's perception of time and of the distance to rivals. In an empty pool, that signal is gone. The swimmer races in a different kind of solitude, and their body responds to that solitude in ways pure time cannot measure. When the stands are empty, every model collapses. I rebuild from the burnt data. During that period, I began logging more about swimmers' subjective feelings after racing, because I believed most answers to questions about the empty-crowd effect would lie there, not in the data columns. Data can tell me whether an athlete swam faster or slower. It cannot tell me what they felt while doing so. This is the fundamental limit of every sports analysis model. The best model still only captures what can be measured. The unreadable part of the data — what happens inside a person when they stand on the starting block — remains beyond the reach of every algorithm I have ever known. I do not say this to celebrate mystery. I say this to remind that an analyst must clearly define the map of his "unreadable data." What percentage of my conclusion rests on evidence? What percentage rests on assumption? If I cannot answer that, I am not doing analysis; I am doing storytelling decorated with numbers. CLOSING Back to Rome 2026. When records fell at breakneck speed, people had two reactions. The first was to call it a peak era of humanity. The second was to distrust every number and give up on analysis. Both are flawed, and both are common. The right path lies in between, and it is not easy: keep the data, but redraw the frame of reference. Add an era label to every number. Separate swimming performance from turning performance. Separate the athlete from the equipment. And above all, remember that every number in the lane is the final result of a chain unfolding in zones the instruments cannot reach. For Vietnamese swimming, I think the future lies in building a patient, honest, long-memoried data system. Not a system to predict medals, but a system to understand where we stand. Knowing clearly where you stand is the first condition for going further. Reputation is only a name. What remains is always how you read the race — and in swimming, that race is run against time itself. I choose to read it slowly, with cross-checking, without rushing to conclusions. If there is one thing I want to leave after this piece, it is this: honesty toward the gap is the foundation of every true understanding of the lane.