The 1500m Lane and the Pacing Equation: Re-reading the SEA Games Medal Table Through SWOLF

In the last four lanes swum over 1500m freestyle by a leading Vietnamese long-distance swimmer, the gap between two adjacent 100m splits narrowed from 2.8 seconds to 0.9 seconds. That number does not appear on any results board. Nobody awards a medal for it, and almost no newspaper prints it. But in my tracking sheet, it is the clearest signal that a long-distance swimmer is entering a mature pacing phase rather than a peak-speed phase. I sit far from the pool to see the race more clearly than the referee. And what I see is not a record, but a curve flattening out. The amateur question here is simple: when a swimmer finishes 1.5 seconds slower than a personal best, do we call that decline or maturation? Both can be true, depending on which metric we read. If we only read the finish time, we see a decline. If we read the split structure, we see a newly built stability. These two readings yield contradictory conclusions, and in Vietnamese swim coaching circles, the first reading still dominates — simply because it is easier. I began watching swimming in 2026, when I was a journalist on the swimming beat at a major Saigon newspaper. Twenty years later, I moved into data analysis for football teams, but a professional habit did not change: whenever I see a contested result, my first reflex is to open the numbers before opening my mouth. Swimming is the sport I come from, and also the sport most easily misread, because it hands the entire narrative to a single string of numbers: the finish time. A little context on how swimming metrics are built, so the reader does not have to trust me blindly. In the 1500m freestyle, the entire race is divided into 30 splits of 50m. Each split can be described by three parameters: time, stroke rate (SR, cycles per minute), and stroke length (SL, metres per cycle). The product of SR and SL is speed. There is a composite index called SWOLF, calculated as stroke count plus time to complete each 50m (in seconds). Lower SWOLF is better, and the interesting thing is that it can stay the same while speed changes, as long as the swimmer trades off rate and length correctly. This is why I say the results board is being misread. When a swimmer finishes slower, most fans immediately conclude form has dropped. But if SWOLF is stable or lower, technique has not deteriorated at all — only pacing has changed. In a training cycle aimed at a bigger stage, deliberately swimming slower in the early phase is a tactical choice, not a sign of decline. Let me return to the specific problem. In an internal tracking sheet I keep for a number of long-distance swimmers, I divide each 1500m race into four blocks: the launch block (first 200m), the base-building block (middle 600m), the transition block (penultimate 500m), and the sprint block (final 200m). For each block, I record average time per 100m, average SR, average SL, and average SWOLF. The first thing worth noting: in the launch block, average 100m time among the top Southeast Asian swimmers is almost identical — the spread between the fastest and the fourth fastest is only about 0.4 seconds. The real difference does not lie in this block. It lies in the transition block, the penultimate 500m, where most swimmers begin to pay for poor pacing in the first half. In this transition block, I measured a significant gap: the leading group held average SR around 34 cycles per minute with SL around 2.1 metres, while the chasing group dropped to around 31 cycles per minute without a corresponding rise in SL — meaning speed fell directly. In other words, the weaker swimmer does not lose because they swim slower in the sprint, but because they lose stroke rhythm about 300m earlier than their rivals. This is a pattern I have seen repeat many times. One citable specific example: at the most recent regional meet, the swimmer who took the highest position in the 1500m freestyle had a third 100m split (metres 201 to 300) about 0.6 seconds faster than the second 100m split (metres 101 to 200). This detail sounds trivial, but it shows that swimmer did not swim at an even rhythm — they deliberately accelerated after stabilising technique. Meanwhile, most of the other rivals swam the third split slower than the second, what we call losing rhythm. Now we come to the hardest part: reading metrics and then interpreting them. Over the past twenty years, I have seen one error repeat in Vietnamese sports analysis, and it is particularly severe in swimming. It is confusing correlation with causation. Specifically: analysts see that swimmers with higher SL tend to have better 1500m times, then immediately conclude that "to swim fast you must extend your stroke". But in reality, high SL may simply be a consequence of that swimmer being taller or having stronger arm propulsion — not the cause. Training SL mechanically while ignoring SR can destroy rhythm and raise SWOLF, that is, make everything worse. Before saying "X leads to Y", I always force myself to point to the concrete physical mechanism linking the two variables. In swimming, that mechanism is the trade-off between propulsion and drag. When a swimmer extends their stroke, the time the hand spends in the water increases, raising the body's drag area — but if propulsion rises correspondingly, speed holds or improves. If propulsion does not rise, the swimmer is simply reaching further without added momentum — the result is a drop in speed. This is a physical mechanism, not an abstract law. For this reason, I am very cautious of advice in the form of "imitate the world star". The metrics of a world-class swimmer are built on their own physique, arm span, and muscle power. Copying their numbers while ignoring body parameters is a methodological error. I have seen too many young Vietnamese swimmers trained on mechanically copied Western programmes, only to reveal problems in breathing rhythm and pacing when they actually compete. That is the methodology. Now the empirical part. I track split data for the country's leading long-distance swimmers across the last three seasons. A sample large enough to extract stability. What I want to stress is this: most improvements in total time do not come from swimming faster in the sprint — they come from reducing the rhythm drop between the 700m and 1200m splits. More specifically, I once analysed data from 12 swimmers in a four-month training cycle. Group A (6 swimmers) focused on improving the transition block, Group B (6 swimmers) focused on boosting the sprint. After four months, both groups improved their 1500m times, but in very different ways. Group A improved by an average of 4.2 seconds, mainly by reducing the rhythm drop in the middle splits. Group B improved by an average of 2.8 seconds, mostly in the final 200m. Notably: Group B had result volatility about one and a half times higher than Group A when competing in different pools. Stability — something the coaching world rarely measures — has higher competitive value. Here I want to be clear: swimming is unlike football in its metric structure. Football has a huge luck factor, and probability models must account for it. Swimming is more compact: everything is measurable, and most results are inevitable if we have enough body and fitness parameters. But precisely because almost everything can be measured, people easily believe they understand everything. That is the trap. Every shock has its own probability. We call it a shock when we have not yet checked the numbers. In swimming, a shock is usually the result of a low-rated swimmer unexpectedly beating a high-rated one. But if we have prior split data, most of these cases are predictable at a certain probability. For example: a swimmer with high split stability over three consecutive prior meets is usually rated lower than a swimmer with a higher personal best but higher volatility. In most cases, people remember the peak result and forget the volatility. This is one of the things I have mentioned many times in previous articles: a personal best is a data point, not a trait. A swimmer's true trait is the probability distribution around their times, including their poor performances. If we read a swimmer only through their best time, we are reading a point at the tail of the distribution and mistaking it for the mean. This is a basic mathematical error, yet it appears on almost every sports news page. Now I want to pose the problem from the reverse angle. Many assume data will help predict precisely. I do not believe that. Data helps narrow the uncertainty band, not eliminate it. In swimming, the uncertainty does not come from physics — it comes from physiology and competitive psychology, two fields my models have never fully covered. This is the point I often have to remind myself of. There is one parameter I cannot quantify but cannot ignore: the stands. In pools in Vietnam, crowd noise varies greatly between venues. An indoor pool with large capacity and a packed crowd can generate noise more than 20 decibels higher than an open-air pool with few spectators. Noise affects the start reaction timing, and in some young swimmers, it raises heart rate just before the start, which in turn affects pacing in the first 200m. The specific magnitude of this effect is beyond my measurement capacity, but I acknowledge it exists as a confidence interval, not a precise parameter. When the stands go silent, the home advantage dissolves into a number close to zero. This was demonstrated during the pandemic, when competitions took place with empty stands. Back then, the rate of swimmers setting personal bests in venues considered "home" fell to almost the same rate as in neutral venues. In other words, what we call home advantage in swimming does not lie in the pool — it lies in the stands. But I do not want to go too far into the psychological aspect. Back to the main analytical framework. With three seasons of data, I built a general descriptive table for the top Southeast Asian long-distance swimmers: SWOLF in the first half of the race correlates positively and fairly strongly with the final result, while SWOLF in the second half correlates negatively — meaning those who keep SWOLF low in the second half usually finish faster. This detail has clear coaching implications: focus on maintaining technique under fatigue, not on accelerating while fresh. I want to tell a story from my own direct monitoring experience. In 2026, when football and most sports paused due to the pandemic, I was invited to review GPS data for several athletes at a training centre in Saigon. The work was unrelated to swimming, but it showed something applicable to swimming: errors in sport usually appear before the result — often weeks before. Small changes in training load, small changes in resting heart rate, small adjustments in training volume — all are signals. If we only read competition results, we read too late. The same is true in swimming. A swimmer who declines at a big meet almost always shows early signs two to three weeks earlier: slightly elevated resting heart rate, lower sleep quality, longer recovery time after hard sessions. These are metrics most coaches do not record. And because they do not record, they cannot trace. A contract is not a signature, it is a hypothesis signed with a name. I always think this when I see a swimmer recruited on high terms. That contract contains a hypothesis that the swimmer will keep improving. That hypothesis may be right or wrong, and the only way to know is to track data over time. No contract guarantees performance. Only time-series data tells us whether a contract was right or wrong. Back to the central story of this article. Over the four recent races of a leading Vietnamese long-distance swimmer, I noted the change in pacing described at the start. But what I want to stress is this: this change did not appear by itself. It is the result of a deliberate training cycle, and it can only be seen when we record split data consistently over many months. This is the point I want to stress to the swimming coaching world in Vietnam. The results board is the output of a crude measurement system — it only records the finish time. But the finish time itself is a composite variable, influenced by dozens of small factors. A swimmer finishing 1.5 seconds slower may actually have improved, if other factors have changed. To know that, we need detailed split data, not just finish data. I admit this is laborious. Pressing a stopwatch at every 50m requires someone present and recording consistently. But this is the minimum investment for a sport that can measure almost everything. There is no reason we should continue to read swimming results only through finish times. Here I want to offer a comparison that may be controversial: swimming and esports have more in common analytically than swimming and football. Football and esports do not differ in essence, only in reflex rhythm — but swimming and esports share one important feature: both are sports that can be almost fully measured through individual metrics. A swimmer and an esports player can both be tracked by reaction index, rhythm index, and stability across multiple competitions. This is something a footballer cannot have, because most of their value lies in interaction with teammates and opponents. This leads to a consequence I consider important: the career span of swimmers and esports players has a similar structure — significantly shorter than footballers — but the post-retirement support systems are very different. A Vietnamese swimmer retires at twenty-six or twenty-eight, and most of them have no career transition system. They can become coaches, but the number of coaching positions with stable salaries is very small. This is a systemic problem, and it cannot be solved with data. But it can be seen through data — if someone bothers to record. I want to return to one of my professional principles: I never praise a tactic or a coaching choice without pointing out the conditions for it to work. In the case of long-distance pacing, the tactic of swimming slower in the first half to preserve technique in the second half only works under conditions of sufficient fitness margin and sufficient technical stability. For a novice swimmer, this tactic can backfire, because they lack the fitness base to exploit the pacing advantage. This is why I do not give general recommendations. Every swimmer is a system of their own, with their own parameters. Any recommendation applied to everyone is likely to be wrong for each individual. In swimming, the differences between swimmers lie in height, arm span, muscle ratio, lung capacity, and recovery ability — none of which appear on the results board. Now I want to pose a question I believe will become a main topic in coming years: will detailed data become a mandatory standard in Vietnamese swimming competitions? Currently, almost no domestic swimming meet publishes split data. This makes analysis dependent on direct observation, which cannot be as accurate as instrumented measurement. If competitions begin providing split data, analysis quality will rise enormously. This is one of the things I find worth anticipating. Not a new record, but a new data infrastructure. A technical era fades when nobody reads its numbers anymore. If Vietnam wants to raise swimming quality at regional and continental level, improving data infrastructure may have a larger impact than importing training programmes from abroad. In most articles about Vietnamese swimming, writers focus on the results of major meets. This makes sense, because results are what fans care about most. But if we only read results, we miss most of the story. The real story lies in the splits, in the training sessions, in the small adjustments nobody sees. This is where I want to end this article: with a question. If Vietnamese swimming's data infrastructure improves in the next five years, will we be able to see improvements that currently cannot be measured? My answer is yes — provided we invest in collecting data before investing in analysing it.