Vietnamese Swimming and the Data We Haven't Read Yet

In the men's 200m freestyle final at the national championships, I sat timing every 50m. The winner opened almost half a second faster than the regional standard for his age group, then lost nearly two and a half seconds in the final split. The third-place swimmer started seven-tenths slower, but touched only two-tenths behind the winner in the closing stretch. On the final results sheet, both are filed under one category. Split the race into four parts, and they become two different stories. I've kept that habit since 2026, when I was writing about swimming for Thanh Nien newspaper. Back then I timed by hand, wrote the numbers in a notebook, and got called back by my editor whenever a figure disagreed with the electronic board. Twenty years later, I still believe most of the story of a race lives in the splits the results sheet never prints. After leaving the newsroom for sports data analysis in Binh Duong, I began tracking swimming the way I track a market. I record pressure metrics in football matches, then ask whether swimming has an equivalent. It does: split structure, stroke rate, stroke efficiency, and what I call the debt-repayment split of the third 50. Context: a sport changing its yardstick Vietnamese swimming enters the 2026 Asian Games cycle with a paradox. National records fall more often each year than in the previous period, yet the number of continental finalists barely moves. Infrastructure has clearly improved: Ho Chi Minh City, Hanoi, and Da Nang have regulation 50m pools, a few centers in Binh Duong have begun investing in electronic timing, and more athletes are sent abroad for training camps. The country's distance tradition, with Nguyen Huy Hoang's SEA Games appearances, shows that aerobic base is not the biggest weakness. So where is the remaining gap? I don't think it's in absolute speed. It's in race structure: how speed is distributed across splits, how breathing is managed as lactate rises, and how you read opponents in a lane where you can see almost nothing but the line on the bottom. To test this, I applied a simple frame to three years of timing data I collected myself: break each race into small splits, compare against the curve of regional finalists, and find the split with the largest deviation. The result repeats so consistently it becomes boring. The biggest deviation isn't in the opening split. It's in the third. The core: reading a race in four parts In the 200m, four 50m splits form an almost fixed structure. The classic pacing of elite swimming is a negative curve: the second split faster than the first, the third nearly flat, and the fourth surging. If you measure the average speed of each split among Asian finalists, you'll find that curve is usually fairly flat at the third split before climbing in the fourth. Among the Vietnamese swimmers I track, the curve is an inverted, shifted V: the first split high, the second still high, the third collapsing, the fourth struggling to recover. The peak is in the wrong place. You cannot swim a 200m by spending all your speed in the opening split and hoping the last one saves you. The pool doesn't lend on those terms: every second spent early is collected back with interest, and the interest rate at the third split is higher than at any other. What's telling is that the third split is not hard to measure. It is uncomfortable, not mysterious. To know where you're losing, you need one disciplined timer and four boxes for numbers. Yet for years, domestic meets printed a results sheet with a single final number, then argued about it as if it were the whole story. I once treated the model as scripture. Now it is only a compass, but without it, we are lost. The four 50m splits are the cheapest compass any coach can use. You don't need a motion-analysis camera system costing hundreds of millions. You need a mind willing to ask: "What was your third split today?" Alongside split structure, two other metrics matter. The first is stroke rate in the second half of the race versus the first. Swimmers losing speed tend either to let stroke rate drop or to spike it. Dropping means losing propulsion; spiking means the battery is nearly empty. Elite swimmers hold stroke rate nearly constant from the second split to the fourth; speed comes from distance per stroke, not from flailing faster in desperation. The second is breathing count per split. In the 200m, breathing distribution is a tactical signal. A swimmer breathing little in the first split may be conserving, but if breathing spikes abruptly in the third, their race is falling apart metabolically. I have watched many races where the third split was only half a second down, yet breathing count rose by nearly half again, and I knew the fourth split would be a fight for survival. This is where we reach what I call the unread data. In Vietnamese swimming we have plenty of final results but very little process data. No age-group-standardized split database. No stroke-rate profiles logged across years. No per-split breathing records. We have many "how many seconds" questions but few "by what means" questions. Working with football data, I learned that expected-goals metrics aren't wrong, only that football is irrational, and after 2026 I learned to count the irrationality too. Swimming is different. The pool is far less irrational than grass: no shifting wind, no long turf, no penalty kicks. So when a swimming model is wrong, the fault mostly lies in the input data, not the sport itself. If you don't record the third split, you have nothing to fix. One period tested this frame harshly. The 2026 pandemic froze competition almost completely. When pools reopened, conditions were different: empty stands, compressed schedules, athletes lacking opponents at their level. When the stands are empty, every model collapses. I rebuilt from the half-burnt data, and realized the third-split curve, which I had assumed was a constant, also shifts with context. What I thought were physiological laws turned out to depend far more on the competitive environment than I had believed. The contrarian angle: more water doesn't mean more speed This is where I have to say something not everyone in the industry wants to hear: training more doesn't necessarily make you faster, and going abroad for camps doesn't necessarily fix the third split. A common habit is to use training hours, weekly kilometers, or sessions per day as the measure of effort. Those numbers are easy to count, easy to report, easy to impress sponsors with. But they barely correlate with how speed is distributed across splits. A swimmer doing 60km a week with a skewed split structure will still lose to one doing 45km with the right structure, at the same fitness level. Correlation is not causation. The fact that many high-performing swimmers train a lot leads us to assume volume is the cause. In many cases, training a lot is a result of already having a good base, not the source of the performance. Reversing the variables is a classic error, and swimming is the sport most prone to it, because training volume is so easy to count. The same goes for overseas camps. Swimming beside stronger opponents is a valuable experience, but it converts into faster times only when you have a measurement frame telling you which split to fix. Without that frame, you bring home a pile of sensations rather than a model. The feel of the water, the atmosphere of the arena, the rhythm of the best swimmer, all are useful, but only when attached to a structure you can verify. I don't deny the value of going abroad. I'm only saying that if you take the same wrong measurement abroad, you'll get the same wrong result, with a larger bill. Blind spots worth naming There are two blind spots I want to name clearly, because they are often conflated. The first is the data blind spot. We measure outcomes but not processes. An athlete knows how many seconds they finished in, but not which split dragged them down. A coach knows a pupil lost, but has no evidence pointing to where to fix it. The second is the coaching blind spot. Lacking process data, coaching is forced to rely on experience and feel. Experience is a valuable asset, but it cannot replace evidence when results have moved in the wrong direction for years. A training program built entirely on feel will produce swimmers who look beautiful when everything goes smoothly, and who hit a wall when change is needed. In the football transfer market, people often say the market is the only place where you pay for expectation rather than present value. Swimming has no transfer window, but it has a similar shadow market: camp slots, investment slots, and media attention. There too, expectation is bought and sold with pretty numbers, mostly final numbers rather than process numbers. Singapore has long been the benchmark for Southeast Asian swimming, with Joseph Schooling its symbol after the 2026 Olympic 100m butterfly gold. Looking at how they built their system, most of the success came from a talent pipeline that is continuously recorded and evaluated, not merely from one exceptional individual. A foundation to build on If I had to choose one thing to do in the next cycle, I wouldn't start with swimming faster. I'd start with recording seriously. A simple spreadsheet with four splits, stroke rate, breathing count, and the date of each measurement is enough to build a data set over two to three years. Once that set is thick enough, every debate changes, because people will argue with evidence instead of impressions. Two major milestones lie ahead: the 2026 SEA Games hosted by Thailand, and the 2026 Asian Games in Aichi-Nagoya, Japan. Both are chances to test the third-split hypothesis, or to bury it if the data says I'm wrong. I accept both possibilities. Reputation is only a name. What remains is always how you read the race. I believe that about swimming more than any other sport, because in the lane you have no one to blame but the structure you built. On August 13, 2026, looking back at three years of timing data on my desk, I still see an empty box in the third split of many races. That empty box belongs to us, not to the athletes. It is the data we haven't read yet, and also the easiest to fill if we stop trusting feel and start trusting the record.