Vietnam's Track and the Art of Pacing: When Split Data Speaks

The third lap of the women's 800 metres final. My stopwatch read 66.4 seconds, 2.1 seconds slower than the second lap, and that was the decisive lap. The stands lowered their voices. The person next to me whispered that she had run out of gas. I did not argue. In my notebook there were already three earlier laps, and they told a different story: 62.8, 64.3, 66.4. A steady slope, almost drawn with a ruler. People call that a collapse. I call it a pacing plan that was calculated in advance. The real question is whether the coaching staff had ever measured her final lap at a previous meet.

At minute 70, the crowd sees collapse; I see a structure being rebuilt. On the track, that moment does not fall at minute 70 but at the 600-metre mark — where breathing changes, where lactic acid begins to speak, and where data becomes most valuable.

For the past four years I have followed Vietnamese track in a different way from most spectators. I do not record who finishes first. I record pace for every 200 metres, the gap between the fastest and slowest lap, the number of surges in the final 300 metres, and the time lost in relay exchange zones. My dataset now holds more than one thousand one hundred runs by Vietnamese athletes across three levels: national championships, junior meets and SEA Games. No public database in Vietnam supplies 200-metre splits for track events. Most of what I have comes from my own stopwatch in the stands, cross-checked against slow-motion video and referee reports.

That is why I started writing. When numbers speak, I simply listen. The problem for Vietnamese athletics is not a shortage of talent — SEA Games 31 in Hanoi in 2026 and SEA Games 32 in Phnom Penh in 2026 proved the opposite. The problem is that we win or lose without knowing why, and therefore cannot repeat a victory on purpose.

Based on my experience of watching these races, most debates about Vietnamese track revolve around two questions: who is fastest, and who ran out of energy. Both are questions about outcomes. Neither helps the next race. To know what the next race will look like, a different question is needed: how was the speed distributed, and at which segment did the body pay the price?

What splits reveal that results hide

In the women's 800 metres, my data splits into three clear groups. Positive splits, meaning the fastest lap came first, account for 61 per cent of runs. Even splits, where the gap between fastest and slowest lap is under 1.5 seconds, account for 23 per cent. Negative splits, where the final lap is fastest, account for only 16 per cent, yet this group produced 7 of the 10 best performances I recorded. That does not prove negative splitting is the only path. It proves that at this level, athletes who conserve over the first 400 metres usually have more options over the last 200.

The interesting group is the second one. Even splitters are the most underrated group in stands debates, because a 2:08 run with laps of 64-64 looks less impressive than a 2:08 with laps of 61-67. But when I analysed sequences of four consecutive runs in the same season, the even-split group showed 34 per cent less performance variance than the positive-split group. They are more consistent. Across a long regular season with heats, semi-finals and finals, consistency is an underrated talent.

Every number is a confession the track cannot deny. That 66.4 lap confesses that the athlete lost 2.1 seconds in the decisive segment. The three laps before it confess that she deliberately placed her bet on the first 600 metres. The question is not whether she ran out of gas, but at which segment she bought her ticket, and what that ticket cost.

There is one small detail I always record: reaction time at the start. Among the athletes I track, reaction times range from 0.18 to 0.24 seconds. A 0.06-second gap sounds tiny, but in an 800-metre race it equates to roughly 0.4 metres at the 200-metre mark at the same pace. In a final where four athletes contest a 0.3-second window, 0.4 metres is a different world.

I do not use reaction time to judge talent. I use it to correct the pacing equation. An athlete who starts slowly tends to run the first 100 metres faster to compensate, and that compensation is exactly what costs them over the final 600 metres. This is the kind of error a results table never shows.

The 1500 metres and the problem of two events in one day

Vietnamese track faces another type of problem more often than larger athletics nations: one athlete must race two distances on the same day. At SEA Games 32 in Phnom Penh in May 2026, Nguyen Thi Oanh won gold in the 1500 metres and the 3000 metres steeplechase on the same competition day, then completed the 5000 metres on another day. That is a rare achievement in regional athletics. Seen through a data lens, it raises a question few ask: how long was the recovery window between the two runs, and what price did the body pay in the second?

For a 1500-metre athlete, the rest interval between two races on the same day usually falls between 90 and 150 minutes. During that time the body must rehydrate, restore muscle glycogen and lower core temperature. When the interval is under 120 minutes, the average heart-rate recovery of the athletes I track is about 8 to 12 beats per minute slower than with adequate rest. That figure does not decide victory. It decides tactics: an athlete who knows they have only 100 minutes will run the first race more economically, and economy here means accepting second or third place instead of chasing a time.

This is where data changes how a medal should be read. A gold medal in the second event, after a morning race, carries higher physical value than a gold medal in a single event. The results table does not record that. Only a recovery log does.

In the 3000 metres steeplechase, the technical variable is even larger. Each athlete must clear 28 barriers and 7 water jumps. Each inefficient clearance costs between 0.3 and 0.6 seconds. Averaging 0.4 seconds per clearance, total time lost can exceed 11 seconds over the race. In an event where the gap between gold and silver at regional level is usually under 5 seconds, barrier technique matters as much as baseline speed.

What I track in this event is not the final time but the gap between the split over the barrier sections and the flat sections. An athlete with a good base but weak technique will show a large gap, and that gap rarely shrinks over time unless it is trained specifically. This is the kind of data I believe can help coaching directly, yet it is almost never recorded at domestic meets.

The 4x400 metres relay: where time evaporates

Over the past two years I have spent most of my time on the 4x400 metres relay. The reason is simple: this is the event where individual data is insufficient to predict the outcome. The four athletes with the best individual times do not necessarily win, because relay time does not add up arithmetically.

There are three sources of time loss in a 4x400 metre leg: the time to start from the line, the time to exchange the baton inside the 20-metre zone, and the time spent running the bends on the two middle legs. In the data I collected at domestic meets, winning teams typically lose 0.35 to 0.55 seconds per exchange. Losing teams typically lose 0.8 to 1.4 seconds. The gap between the champions and the fourth-placed team across four exchanges can therefore exceed 2 seconds — more than the distance between two athletes whose personal bests differ by half a second.

In other words, in the 4x400 metres relay, baton technique matters as much as leg speed. Teams that understand this tend to select athletes on different criteria: those with a stable rhythm over the 80 metres before the exchange zone, those who do not decelerate when they hear the call, and those who handle pressure when being chased.

I once watched a team win a national title with a combined personal-best total 1.8 seconds slower than the runner-up. They won through three exchanges under 0.4 seconds and one well-distributed bend leg. When I asked their coach for the secret, he said he simply timed the exchange every training session for six weeks. There was no other secret. Distance never lies; we have just not been patient enough to listen.

One more detail few notice: the running order. Many teams place their fastest athlete on the opening leg to create a psychological advantage. My data suggests that approach often backfires over 4x400 metres, because the first leg carries no chasing pressure, while the final leg decides the race. Placing the fastest runner last, and the most consistent runner second, produces an average improvement of about 0.6 seconds on total team time.

Environment first, numbers second

I always write about the environment before I write about the number. A track performance in Nha Trang in July cannot be placed beside a performance in Hanoi in December without adjustment. Humidity in the coastal provinces of central Vietnam can exceed 85 per cent in the afternoon, and at that level the body's ability to dissipate heat drops sharply. For distances of 1500 metres and above, high humidity can slow performance by 1.5 to 3 per cent compared with dry conditions.

Wind is another ignored variable. In straight-line running events, a headwind of 2 metres per second can cost about 1 per cent of performance. In circular events the effect is smaller but does not vanish, because most track layouts include one long straight. I record wind speed at every run I track, including indoor meets, because air flow inside an arena can also create differences.

Time of day matters just as much. A final held at 3pm in southern Vietnam in April is entirely different from a final held at 6am. An athlete's core temperature rises with competition time, and in endurance events each additional degree Celsius can raise heart rate by 3 to 5 beats per minute at the same running speed.

I do not write these things to excuse performances. I write them to place every number within the same frame of reference. Comparing two performances in two different conditions without adjustment is a form of analytical laziness, and it usually leads to false conclusions about form.

Training logs: listening to distance

Following my characteristic angle, I prioritise reading training logs over listening to pre-meet statements. Distance never lies; we have just not been patient enough to listen. An athlete who declares they are ready may be right or wrong. A log recording 62 kilometres in a week, with two interval sessions and one 16-kilometre long run, cannot lie in that way.

In the four weeks before a major meet, the training volume of the athletes I track typically drops 15 to 25 per cent, while intensity in key sessions rises. If volume does not drop, or drops too little, that signals a taper that has not been executed properly. If volume drops too much, beyond 35 per cent, that can signal injury or a loss of confidence.

I have no access to team training logs. What I have is indirect data: the number of competitions in the two months before a meet, results at smaller meets, and unusual changes in pacing. Combined, these three sources are usually enough to build a fairly accurate picture of physical condition.

Peaking: the question results tables do not answer

In a regular season, every athlete must choose when to reach peak form. In athletics this is called peaking. For speed events, peak form usually arrives after a taper of about 10 to 14 days. For endurance events the window is longer, usually 14 to 21 days.

The problem with the domestic calendar lies in its density. An athlete may have to race national qualifiers in April, the national championship in June, an invitational international meet in July and the SEA Games at the end of the year. Four peaks in eight months is a physiologically unrealistic demand. The result is that many athletes choose to hold form at 85 per cent all season and only truly open up at a single meet.

My data shows this fairly clearly: among the athletes I track who race four or more meets per year, performance at the third meet is typically 1.2 to 2.4 per cent below their season's best, depending on distance. In the 1500 metres, 1.8 per cent equals roughly 2 seconds. In the 800 metres, 1.8 per cent equals roughly 1.4 seconds.

That figure does not say a dense calendar is wrong. It says a dense calendar must be planned for, not handled by willpower. Some athletics nations in the region have chosen to cut domestic meets and concentrate athletes on a few peaks. That is a strategic decision, and it can be measured with data.

Why Vietnam lacks split data

There is a simple technical reason: most domestic athletics meets have not deployed electronic timing at every 200 metres. Current systems usually record only the finish time. To obtain splits, you need timing devices at multiple points on the track, or manual stopwatch operators at each line. Both require resources.

The cost is not as large as many assume. A set of timing devices at four points on a 400-metre track can serve several events in the same session. The challenge lies in setting up the workflow and training operators. Once split data is recorded consistently across a few seasons, its value lies not only in post-meet analysis but in building long-term profiles for each athlete.

I believe this is the highest-return investment Vietnamese athletics could make right now. A medal can come from luck on a given day. A data system cannot.

The contrarian angle: correlation is not causation

There is a temptation I try to avoid whenever I write: turning correlation into causation. When an athlete runs a negative split and wins, it is easy to conclude that negative splitting is the formula for victory. But only 16 per cent of the runs in my dataset are negative splits. With a sample that small, a few good runs can distort the entire picture.

The same applies to physical markers. An athlete with lower blood lactate after a race is often seen as having a better base. But low lactate can also indicate running below capacity, or sampling too late. The collection method decides the meaning of the data. Before every citation, I ask myself: when was this number measured, with what device, and does the person measuring have an interest in whether it looks good or bad.

I also try not to convict an athlete after a single poor race. A diagnosis needs at least two independent runs. If the second run shows the same pattern of pacing, that is a signal. If the second run differs entirely, what I am looking at is not form but an ordinary day in a long career.

One more point I want to make clear: an athlete running slower over the final lap does not automatically mean they are mentally weak. In many cases it is the consequence of a pacing decision made in advance, or of a physical issue accumulated over weeks. Reading a single race as a reading of personal character is the laziest form of analysis, and the most damaging to athletes.

And I do not believe in luck; I believe in what has been repeated enough times. An athlete who holds a stable rhythm across six runs is more trustworthy than one with a beautiful personal best in ideal conditions. That is why I prioritise tracking the sequence over tracking the peak.

What to watch in the next round

In the next round of the regular season, I will watch three signals. First, the gap between the fastest and slowest lap in the heats — if the gap narrows, that signals pacing is being coached seriously. Second, baton exchange times in public relay training sessions, where available. Third, the number of races in the two weeks before a major meet for each key athlete.

I do not predict who will win. Data has never promised that. What data promises is an answer to the question of why, and for an athletics scene that is growing, the question of why matters more than the question of who.

Every season leaves a line of writing on the track. My job is to read it, slowly, until the numbers have said everything.