The Last-50 Gap: Reading Vietnam's Swimming Splits Before ASIAD 2026

At lane four, the scoreboard flashed 28.4 as the swimmer touched the wall to close the opening freestyle leg of the 200m individual medley. Forty seconds later, on the second freestyle leg, those same arms turned over noticeably faster and the clock returned 31.1. The gap between two freestyle segments inside a single race was 2.7 seconds.

I keep that gap in a file named fade_gap. It now holds more than a thousand rows, collected from public result sheets at domestic meets and a handful of regional championships over the past two seasons. The median for the men's 200m individual medley at national level sits between 2.5 and 2.9 seconds. Compared against split sheets for the same event, same sex and same age band at regional championships, that median falls to 1.5 to 1.8 seconds.

Most of the lost time is not in the opening segment, where the crowd sees the excitement, but in the closing one, where almost nobody is still watching the clock.

A database built from result sheets taped to walls

I began writing about swimming in 2026, at a sports desk in Saigon, when the number of competition-standard 50m pools in the country could be counted on one hand. Twenty-two years later the number has grown, but density remains low: the majority of athletes in the national selection system do more than 80 percent of their training volume in 25m pools.

That is the first input parameter, and it is not purely technical. A 1500m freestyle race in a 25m pool contains 60 turns. The same distance in a 50m pool contains 30. Each turn carries a streamline, an underwater dolphin phase and a breakout. If an athlete's training unit is 25 metres, most of the working day is spent on turning and breakout skills rather than on continuous stroke rhythm. When competition moves to a 50m pool, half the walls disappear and the athlete pays for exactly the thing they were never trained enough to do: holding stroke length while fatigued.

Every such parameter becomes a variable in my model.

The four variables I have tracked longest are split deviation, heat-to-final conversion, event load, and fade point. None says anything on its own, but layered together they establish an order of priority for fixing problems.

The window I use to calibrate the model runs from the SEA Games cycle in Thailand to ASIAD 2026 in Aichi and Nagoya, Japan, which opens on 19 September 2026. Every current training plan points at that date, and it is also the date on which errors in event allocation will be exposed most clearly.

None of my data comes from an automated collection system. It comes from sitting at both ends of the pool, clocking every 50 metres, and cross-checking against the result sheets posted after each session. The method is slow and error-prone, but it produces something aggregate results never contain: the internal structure of a race.

Variable one: split deviation

Split deviation is the difference between the final 50 metres and the opening 50 metres of the same event, measured under identical conditions. In individual medley events I use the two freestyle segments, because they occur at two points in the same race and allow comparison without changing stroke.

The amateur question is simple: why does a swimmer who is 0.6 seconds faster than an opponent at the first turn still finish 1.2 seconds behind?

My numbers answer in three layers. The first is the deviation itself. Among national-level swimmers, the men's 200m individual medley has a median around 2.7 seconds, with a spread from 1.9 to 3.6. At regional level, the median is roughly 1.6. The second layer is stroke rate, measured in strokes per minute. The third is distance per stroke.

The last two layers explain the first.

In my domestic dataset, over the closing 50 metres, stroke rate rises 6 to 8 percent against the opening 50 while distance per stroke falls 8 to 10 percent. At regional level, stroke rate rises only 3 to 5 percent and distance per stroke falls only 3 to 4 percent.

In plain language: when exhausted, Vietnamese swimmers turn their arms over faster but pull less water, while the region's leading swimmers retain almost their full stroke length. The mechanism is pure physics: propulsion in swimming depends on the mass of water pushed backwards with each stroke, not on the number of strokes. Arms that churn quickly while slipping produce the sensation of greater effort while actually producing a slower swim.

The extrapolation has practical value. Hold the opening segment constant and bring only the closing efficiency up to regional level, and the men's 200m individual medley improves by roughly 0.9 to 1.1 seconds. In an event where continental final qualification is often decided by half a second, that margin changes placings.

Variable two: heat-to-final conversion

My conversion index is the difference between the final time and the heat time. A positive value means the final was slower.

At domestic meets, the common pattern is an athlete swimming a personal best in the heats, then entering the final with reserves already spent. In my dataset, the median conversion index for men's individual events at the national championship is plus 0.8 to 1.5 seconds. Among the region's leading swimmers, the index sits between minus 0.2 and plus 0.6, meaning faster finals are normal.

Why is a fast heat and a slow final a measurable phenomenon?

Three reasons, none of which concern competitive nerve.

The first is opponent density. In domestic heats, a lane may hold three or four swimmers of similar standard, producing a natural chase effect. In the final, after cuts, the leader often breaks clear within the first 50 metres and has nobody to chase mid-race. Times do not arise from solitary effort; they arise from a specific opponent in the next lane.

The second is schedule compression. Some domestic meets run heats and finals on the same session, six to eight hours apart. For events with high metabolic demand such as the 400m individual medley or the 1500m freestyle, six hours is not enough to restore local energy stores in the arm and shoulder muscles.

The third is the condition of the stands, and this is the variable I handle most carefully. At SEA Games 31 in Hanoi in 2026, with a full arena and noise levels few squad members had ever experienced, the home team's conversion index flipped clearly positive. At championships held abroad, with the stands no longer theirs, the index deteriorated immediately.

When the stands fall silent, home-pool advantage collapses into a number close to zero.

I raise this so the model knows which parameters can be switched on and off.

Time of day matters too. Heats at international meets are usually swum in the morning, while most domestic training happens in the afternoon or evening. In my file, domestic swimmers tend to be 0.6 to 1.2 seconds slower in morning heats than their personal bests, against 0.2 to 0.5 seconds at regional level. The mechanism involves body-temperature rhythm and neural activation. Training at the competition hour is not a ritual; it is a physiological variable, and it costs nothing but consistency to fix.

Variable three: event load and turnaround

Event load is the number of competitive swims an athlete takes in one meet, relays included.

The record in my file belongs to a SEA Games more than a decade ago, when a Vietnamese swimmer entered eight individual events and won all eight. Nobody in the region has repeated that load since, and I use it as the reference for an extreme loading configuration.

My data shows a stable relationship: athletes entering six or more individual events record split deviations roughly 0.4 seconds larger than the rest at their last swim of the day. Accumulated across three days, that becomes about 1.2 seconds, the equivalent of the distance between a medal and fourth place in many events.

The problem is not the absolute figure but the mismatch between domestic and regional rhythms. A typical domestic meet demands six to eight swims across three days. A regional championship demands four to six across five or six days, with morning heats and evening finals. A body calibrated to a dense rhythm enters a sparse one with neural arousal at the wrong moment, and vice versa.

This is the kind of error result sheets never display. It only appears when the competition calendar is placed next to the training calendar.

Variable four: fade point

Fade point is the distance at which a swimmer's 50-metre split exceeds their own average by one second. It is the point at which the race starts slipping away.

In my 200-metre dataset, the average fade point for domestic swimmers falls between 130 and 140 metres. At regional level it retreats to 160 or 170 metres. In the 1500m freestyle, the domestic fade point sits around 1,150 to 1,200 metres, while regional swimmers pass 1,300 metres before declining.

The tactical meaning of the fade point lies in its position immediately after a turn. That is not coincidence. After each turn, a swimmer must rebuild stroke length out of the streamline, and that moment of rebuilding is where skill is tested. With reserves available, rebuilding is automatic. Once depleted, it becomes a conscious decision, and conscious decisions are always slower than trained reflexes.

A 200-metre race is lost around metre 140, not metre 200.

The 1500m freestyle case

The continental silver medal in the men's 1500m freestyle is the highest coordinate Vietnamese swimming has ever reached at distance. I treat it as the benchmark against which everything since is measured.

The splits I keep show a striking structure: through the middle section, roughly metres 400 to 900, this swimmer's 100-metre splits were almost flat, varying by less than 1.5 seconds. That signals a solid aerobic base. But over the final 100 metres, splits dropped 3 to 4 seconds below that flat pace.

The real question is where that drop comes from, given that the previous 900 metres showed no weakness in conditioning.

The answer lies in the interaction between sit-and-kick tactics and stroke rate. In a 1500m race, sitting on the leader's feet is a rational energy-saving choice: the athlete lets an opponent set the rhythm and benefits from not having to generate pace. But it also means the race rhythm belongs to someone else. When the leader surges around metre 1,200, the follower must lift stroke rate by 4 to 6 percent within a very short window, while distance per stroke has already begun falling. The result is a collapse in efficiency, not a collapse in will.

An environmental variable belongs here too. With most training volume in 25m pools, a 1500m swimmer trains with 60 turns but races with 30. The skill practised most is the one that pays least in competition. No amount of extra training fixes that; only different training does.

Given the current data structure, I consider the 800m freestyle the better conversion event for Vietnam's distance swimmers in this cycle. It is long enough for the aerobic base to matter, yet short enough that the fade point moves beyond the finish wall instead of sitting mid-pool.

The 200m butterfly and 200m individual medley cases

Butterfly exposes the fade point most clearly, because the recovery phase occurs above the water and drains the shoulders.

In my domestic dataset, women's 200m butterfly split deviation sits between 3.0 and 3.6 seconds, well above the 2.5 to 2.9 range in the men's individual medley. The mechanism is the fatigue threshold of the shoulder muscles. When it is crossed, stroke rate rises while recovery amplitude shortens, and distance per stroke collapses faster than in any other stroke.

The common coaching error is adding volume to raise the aerobic ceiling. That helps general conditioning but misses the deciding muscle group. What is needed is strength-endurance work specific to the shoulders, combined with tethered swimming and rate-controlled sets, so athletes learn to hold rate rather than push it when tired.

In the women's 100m breaststroke the problem sits elsewhere. Split deviation is smaller, but lost time concentrates in the underwater pullout after the turn. In my data, domestic swimmers typically surface 7 to 9 metres off the wall, while regional swimmers extend to 11 or 13 metres. Extra metres do not automatically create an advantage, but surfacing early to breathe costs 0.3 to 0.5 seconds for a decision lasting a fraction of a second.

In the men's 200m individual medley, the closing freestyle leg decides the race, and it depends most on something rarely trained in isolation: switching from breaststroke rhythm to freestyle rhythm without losing stroke length. Changing stroke while exhausted is a skill of its own, not a natural consequence of swimming a lot.

Sprint events: where split deviation nearly disappears

Over 50 and 100 metres, split deviation loses most of its diagnostic value. In my file, the men's 100m freestyle has a median deviation of only 0.9 to 1.3 seconds, and most of that comes from the start and the underwater phase rather than stroke rate.

At sprint distance, three parameters decide the outcome: reaction time off the blocks, underwater distance before breakout, and the number of breaths taken. My data shows domestic swimmers reacting 0.05 to 0.10 seconds slower off the signal, breaking out 2 to 3 metres earlier, and breathing roughly twice as often in a 100m freestyle.

One breath at racing speed costs about 0.1 to 0.15 seconds through changes in stroke rhythm and propulsive force during recovery. Two extra breaths mean 0.2 to 0.3 seconds. Adding the start and the underwater phase, domestic swimmers surrender roughly 0.5 seconds in a 100m race before the race has properly begun. At sprint distance, half a second is the entire gap between a semifinal place and elimination in the heats.

These parameters also improve far faster than aerobic capacity. An 8 to 12 week block focused on starts and underwater work can yield 0.2 to 0.3 seconds, while distance swimmers need years to build a base. That asymmetry explains why sprint investment often shows visible results sooner, even though its medal value at continental level is lower.

Juniors and the puberty threshold

Among the variables my model handles poorly, puberty is the hardest.

For female swimmers aged 12 to 15, structural change produces a period in which performance can stall or decline while training volume is unchanged. In my dataset, girls in this age band show split deviations 0.3 to 0.5 seconds larger than their own previous season, even though stroke rate and distance per stroke have not deteriorated much.

The mechanism involves the ratio between limb length and body mass, and how propulsion transfers from shoulder to hand. When that ratio shifts, old skills no longer return old efficiency, and the swimmer must relearn stroke rhythm on a different body. For boys, the growth phase usually comes with rapid strength gains and deviations tend to narrow, which sometimes creates the illusion of a breakthrough when it is simply a larger body.

The coaching consequence is concrete: for girls in this age band, judging progress by absolute times is a mistake. The metric to track is distance per stroke, remeasured every 6 to 8 weeks, because the body threshold changes faster than the standard training cycle. Swimming is a sport where peak performance age for women typically arrives two to three years earlier than for men. A female swimmer may peak at 19 and then must find a new way to swim to hold her position, while male teammates of the same cohort are still rising.

Relays and take-off timing

In a four-leg relay, the time that is not swum but spent waiting is the easiest thing to overlook.

The gap between the incoming swimmer's hand touching the wall and the outgoing swimmer's feet leaving the block, in my domestic dataset, usually falls between 0.28 and 0.35 seconds. Among the region's leading nations, that gap narrows to 0.15 to 0.22 seconds. Multiplied across four legs, the difference amounts to 0.4 to 0.8 seconds.

The cause is not reflexes. It is that the domestic training environment rarely simulates a take-off with disqualification risk attached. Because leaving early costs the entire result, swimmers default to a safe cushion. That cushion only breaks with hundreds of repetitions using immediate feedback timing equipment, and with coaches accepting that disqualifications will occur in practice.

No aggregate result sheet ever shows this. Seeing it requires someone with a stopwatch at both ends of the pool.

The overseas training model as a control group

My dataset contains a small group of Vietnamese swimmers trained abroad. It is too small for statistical conclusions, but sufficient as a qualitative control.

The difference is not absolute speed. On raw speed the gap between them and domestically trained swimmers is modest. The difference is conversion: swimmers raised in the international competition environment routinely post negative conversion indices, meaning faster finals than heats. For them the final is the objective and the heat is a formality.

One notable case is a butterfly and individual medley swimmer who competed in the United States. In the meets I tracked, his ability to hold stroke length over the closing 50 metres ranks among the best of any Vietnamese swimmer, keeping his split deviation below the domestic median regardless of his finishing time.

The finding that training location shapes conversion more than raw speed is valuable. It suggests the bottleneck in Vietnamese swimming sits in competition, not in training.

At the top tier of Asian swimming, Japan and China retain dominance in most events, especially individual medley and butterfly. South Korea is solid in middle-distance freestyle and the men's medley. The chasing group of Hong Kong, Singapore and Thailand has a lopsided distribution, usually concentrated in a few sprint events and women's breaststroke.

For Vietnam, the competitive opening lies in freestyle from 800 metres up, women's butterfly and women's breaststroke. These are events where qualifying times fluctuate more between championships, meaning a wider door than at sprint distance, where the density of qualifying athletes is heaviest. That narrowness is not bad news. It simply means errors in resource allocation will be more visible, because no event remains to shield another.

The contrarian read: the last-50 gap is the price of a choice

Reading fade_gap and concluding that Vietnamese swimmers lack conditioning would be a basic inferential error.

There is a less comfortable explanation: those with the largest split deviations are precisely those who dared to swim the opening fast enough to be in the race. A swimmer who chooses even pacing never approaches the national record, because they never leave the comfort zone in the first 100 metres. A large split deviation is the price of ambition, not evidence of deficiency. I call this the selection effect, and it renders naive group comparisons meaningless unless opening speed is controlled for.

More carefully still, correlation must be separated from causation. Monthly 50m-pool sessions correlate fairly tightly with split deviation in my file, but the correlation only means something once a mechanism links the two variables. That mechanism is turn economy and the ability to rebuild stroke length, not pool length itself. In other words, the issue is not the 25m pool but the ratio of training volume allocated to turning skills versus sustained rhythm. Without identifying that mechanism, I should say only that two data series are related, not that one produces the other.

Finally, there is variance I cannot explain. At SEA Games 31 in Hanoi, arena noise far exceeded anything the home squad had experienced. Under such conditions, part of the outcome cannot be attributed to any variable in my model. When judging races where crowd conditions exceed historical thresholds, I widen the confidence interval and say so.

And I do not call any result a shock before checking the numbers. Every shock has its own probability. We call it a shock when we have not yet checked the table.

The signal for the next cycle

I will track one figure next season: the closing 50-metre split deviation in the men's 200m individual medley at the national championship.

If that figure narrows below 2.0 seconds while the opening 50 metres holds, it is the leading signal for ASIAD 2026 and for the Olympic cycle beyond. If the opening segment slows at the same time the deviation narrows, that is a different kind of progress and requires a different reading.

A single swim happens once. Its trajectory lasts for years. Most people look at a medal table to understand a championship; I look at split sheets to understand the years.