The DPS Threshold and the Blind Spot of Vietnamese Swimming: Lessons from the 200m IM Splits at SEA Games 31

On the night of the women's 200m individual medley final at SEA Games 31, I did not look at the scoreboard. I watched the split graphic replayed on the secondary screen. When the swimmer turned at the 150-metre mark, a number made me put down my pen: stroke rate rose 8%, but distance per stroke (DPS) fell 11%. That is the classic signature of an impending tempo collapse. The final 50 metres confirmed it. Swimming differs from football in one respect. There is no rebound here, no shot off the post, no referee. Water responds with brutal honesty. But precisely because of that, the numbers along the lane become more trustworthy than the medal table. Data never lies, but it knows how to hide. I began building a Vietnamese swimming dataset in 2026, when I was still working as a transfer-market administrator in Nha Trang. At first it was just a habit of logging splits from domestic meets. By SEA Games 31, I had a data chain covering more than 60 swimmers, enough to reconstruct the tempo structure of each lane. No lane is meaningless, not even the lanes without medals. SEA Games 31 took place in Hanoi in May 2026, at the My Dinh pool. It was the edition where Vietnamese swimming set its biggest target in years, with the successor generation to Nguyen Thi Anh Vien and new faces carrying expectations. The organisers published a dense schedule: many swimmers had to race heats in the morning and finals in the evening on the same day, and some contested three events in two days. That schedule created a variable the medal table cannot express. A gold medal in a swimmer's specialty can be undervalued if, on that day, the athlete had already swum two events beforehand. Conversely, a personal best in a secondary event, with a tired body, says far more about the physical base. I chose the women's 200m individual medley as my sample because it tells the story most clearly. Four strokes—butterfly, backstroke, breaststroke, freestyle—force the swimmer to allocate energy deliberately. There is no room for pure explosion. It is a resource-management problem, and every deviation shows up in the splits. I divided the lane into four 50-metre blocks and calculated three metrics for each: split time, stroke rate (strokes per minute), and DPS (metres). The product of stroke rate and DPS yields speed. When a swimmer raises stroke rate while DPS drops, it means they are fanning the water rather than swimming it—the arms turn fast but do not push any volume of water. In the first 100 metres (butterfly and backstroke), the leading group averaged a DPS of 2.05 metres and a stroke rate of 42. By the breaststroke block—the slowest stroke—DPS fell to 1.4 metres and stroke rate dropped to 32. That is normal, because breaststroke has a long glide phase. The problem lies in the final freestyle block. In the last 50 metres, the medallists held DPS at 1.85 metres and raised stroke rate to only 48—a 14% increase over the first block. The group behind pushed stroke rate to 54 but let DPS fall to 1.55 metres. They swam faster in arm tempo, but each stroke was shorter, and their total time was 0.9 seconds slower. This is what I call the DPS threshold: when DPS falls below 1.6 metres in the final block, the swimmer has almost certainly exceeded their metabolic limit. My data shows this threshold held stable across three consecutive SEA Games, even as the swimmers' absolute speeds rose. Another metric worth noting is turn time. In a 50-metre pool, a good turn takes roughly 0.7–0.8 seconds from wall contact to the head leaving the wall. In the SEA Games 31 leading group, that figure was 0.72 seconds. In the middle group, 0.85 seconds. A 0.13-second gap sounds small. But a 200-metre lane has three turns, and multiplied out it becomes nearly 0.4 seconds—exactly the margin between gold and bronze in this event. While building the dataset, I once sent requests for official splits to the organisers of three domestic meets. Two went unanswered, and one returned a photo of a scoreboard at such low resolution that the digits in the outer lanes were unreadable. That is why I had to time from video myself—a time-consuming job, but one that forced me to rewatch every stroke. Rewatching repeatedly is what helps me notice things a single live viewing would miss. What is interesting is that these metrics do not contradict the crowd's feeling at all. When the stands roar at a sprint finish, the data shows that is the moment stroke rate spikes. But the decisive moment is not there. It is in the breaststroke block, where the swimmer holds DPS while rivals begin to lose the structure of their stroke. Compared with swimming nations in the region, Vietnam has one clear strength and one clear weakness. The strength is acceleration in short events—a product of natural physical base. The weakness is pacing distribution in long events, where DPS must be held steady across the whole lane. The region's strong swimming nations usually win not because they swim faster in one block, but because they do not collapse in any block. The medal table tells one story, but the splits tell the opposite one. A swimmer who wins gold in a secondary event may be showing signs of overload; one who leaves the final empty-handed may have a more stable improvement curve month by month. If we look only at medals, we inflate one and forget the other. The problem lies in the fact that Vietnamese swimming data is almost never made public. Official splits usually exist only at major international meets, while domestic meets depend on the on-site scoreboard. That creates a blind spot: we know who won, but not how they won. And without knowing the how, we cannot predict who will improve and who will stall. I once faced harsh pushback for saying that a rising swimmer should not be over-hyped. The coaching staff said I was judging from outside the pool. But data does not care who says such a thing. Three months later, that very swimmer's form dropped in the freestyle block, exactly as the DPS curve predicted. There is another trap. A high DPS is not necessarily good. In the butterfly block, some swimmers reach a DPS of 2.2 metres thanks to a long underwater glide—effective in the short term, but oxygen-costly and reducing their ability to sustain over later blocks. Correlation is not causation. A beautiful metric at one moment can be the cause of a collapse at another. Luck is something I do not have. I have probability and enough data. And in swimming, luck barely exists—there is only tempo managed well or lost. If I had to pick one signal to track over the next two years, it would not be medals, but the DPS threshold in the final freestyle block. When a new generation of swimmers holds DPS above 1.8 metres in the last 50 metres, that is a sign the physical base has changed. And while that number still hovers around 1.6, every grand expectation is just applause in front of a curve that is not yet thick enough. The pool hides nothing. We simply have not chosen to read it.