When the Swimming Data Pipeline Breaks: Rebuilding from the Charred Data

At 22:47 on August 13, I opened my swimming spreadsheet in a small apartment in Thu Dau Mot, Binh Duong. Nine tabs. Nine analytical dimensions I had built over six years, dating back to my time in the data room at Becamex Binh Duong. All nine were blank.

I had not abandoned the work. All nine tabs were blank because the input layer — the source-deconstruction layer my workflow calls Stage-1 — returned an empty payload: no title, no source, no core viewpoints, and most importantly, a completely empty information-points field. A document with no entities, no competition data, no events, no signals about source quality.

What matters is how the system responded. The nine analytical dimensions each ran, and each returned the same single sentence: insufficient information, cannot assess. Technical. Performance and data. Competition system. World landscape. Rules and anti-doping governance. Athlete career. Risk profile. Public narrative. Industry ripple. Nine out of nine, no exceptions, not a single line of speculation.

I stared at that empty table longer than a technical error deserves. Then I realised: it was the most honest document I had produced in years. And it told the truth better than any glossy analysis I had ever written about Vietnamese swimming.

What I do, and why swimming is the hardest room

I am Bui Phong, 41, a sports data analyst specialising in swimming. My career started in 2026 at Thanh Nien newspaper, as a reporter covering the blue lane. Back then I wrote with my eyes and a hand-held stopwatch. Later I wrote with a spreadsheet.

In 2026, while working as a senior specialist for Becamex Binh Duong, I analysed all 26 rounds of the V-League season. The club's average PPDA was 8.4 — the lowest in the league — meaning opponents completed only 8.4 passes before being pressed. Their xGA was 0.68 per match, with 14 clean sheets. The article, with 17 charts, drew more than 250,000 reads. From then on I set myself a professional rule: every number in a piece must pass three cross-checked sources, and every chart must state its statistical timestamp.

In 2026, I built an xG model from 180,000 shots across five European leagues for a broadcaster in Russia, and called 14 of 16 knockout matches correctly. In 2026, when the Bundesliga returned with 312 matches behind closed doors, I measured home advantage falling from 54% to 47%, with home-team PPDA rising 0.9. In 2026, I carried the same framework into the Euros and the Tokyo Olympics, predicting Italy to win from the quarter-finals and correctly identifying 8 of Brazil U23's 10 most important players.

But swimming is my professional homeland. And swimming is the cruellest sport for anyone who works with data, for two reasons.

First, swimming is decided in hundredths of a second. An athlete can miss a final by 0.03 seconds — a span no human can perceive, but electronic timing captures it. Data error in this sport is not an academic matter; it is a medal matter.

Second, and this is Vietnam's specific problem: the public data layer for domestic swimming is so thin it barely exists. Football has the V-League with 26 rounds, statistical providers, match reports. Swimming has a results sheet. A results sheet answers one question: who touched first. It does not answer why.

The nine dimensions in my system are built to answer "why": technique (starts, underwater phases, turns, stroke efficiency), performance and data coordinates, competition systems and selection mechanisms, the world power map, rules and anti-doping governance, athlete careers, risk profiles, public narrative, and industry ripple.

Attached to it is a principle many in the trade treat as weakness: if a dimension lacks sufficient information, I write plainly "insufficient information, cannot assess" instead of guessing. A framework proves its worth only when it dares to say "I do not know". A framework that can always answer is a framework that is making things up.

The technical layer: the silence lives in the metres nobody records

Start with the easiest thing to verify. In a 200m individual medley, there are four turns, one start, and four underwater phases. Together, the non-swimming portions account for a large share of total time. At continental level, the gap between two athletes usually sits not in straight-line speed but in turn quality and the length of the underwater phase.

Watching Nguyen Thi Anh Vien in her peak years, I was always struck by her dolphin kick off the wall. It was one of the few genuinely continental weapons Vietnamese swimming possessed. But I cannot prove it with numbers. There is no public split dataset for domestic meets. I know she was strong in that segment because I sat at the pool edge counting footbeats. I cannot cite it as a figure.

Nguyen Huy Hoang is the mirror case. In the 1,500m freestyle — the event where he won silver at the 2026 Asian Games in Jakarta — what decides the outcome is not peak speed but the flatness of the pacing curve. Over distance, a badly distributed race can cost an athlete one to two seconds across the final 300 metres, once lactate has crossed threshold. I have hand-timed his 100m segments in internal test sets, but that data does not belong to me to publish.

That is the paradox of the technical layer. An analyst of Vietnamese swimming usually knows more than he is allowed to write. The technical column in my spreadsheet is always the fullest by intuition and the emptiest by evidence.

The performance layer: coordinates without a map

To judge a swim, you need three coordinates: the world record, the all-time list, and the current-season ranking. Without those three, a number is just a number.

There is one historical variable any swimming analyst must handle: the polyurethane suit era of 2026–2026. At Rome in 2026, a wave of world records fell in a single meet, largely thanks to buoyancy properties in the fabric. Only in 2026 did the international federation tighten the competition-swimsuit rules. Which means every cross-era comparison must label which era a record belongs to.

In Vietnam we rarely do that. Nguyen Huy Hoang's 2026 Asian Games medal in Jakarta was celebrated by the media, and deservedly so. But very few articles placed it in coordinates: in the men's 1,500m freestyle, how many seconds separated a world final berth from an Asian medal in that same period. Without coordinates, fans default to the assumption that an Asian medal means world class. It almost never does.

I once treated models as scripture. Now they are only a compass — but without one, we get lost.

The competition layer: which cycle decides what

A swim result means different things depending on where it sits in the cycle. The SEA Games is a regional stage, where many delegations field young squads to accumulate experience. The Asian Games is a continental arena where finals berths are largely controlled by Japan, China and South Korea, with occasional exceptions from Singapore. The Olympics is where A-cuts and B-cuts turn qualification itself into a strict selection problem.

For Vietnamese swimming, this cycle creates a familiar trap. Athletes peak at the SEA Games; the next step is the Asian Games, where competitive density jumps sharply. That is a cliff, and the distance between a SEA Games peak and continental standard is usually underestimated.

SEA Games 31 was held in Hanoi in 2026, with competition at the My Dinh Aquatic Sports Palace. SEA Games 33 is scheduled for Thailand in December 2026. The 2026 Asian Games will be held in Aichi–Nagoya, Japan, from September 19 to October 4, 2026. Those three markers should be read as one straight line, not three isolated events.

Selection mechanisms are another overlooked variable. When an athlete reaches a final without hitting the A-cut, an Olympic berth often comes through other routes. How a berth is created determines how it should be read. A world-level entry and a world-level medal are two entirely different assets.

The world map: where Vietnam stands

In the men's 100m freestyle, the current landmark is the 46.40-second world record set by Pan Zhanle at the Paris 2026 Olympics.

In the middle and distance groups, the picture has changed completely since the previous era closed. Bobby Finke, Daniel Wiffen, Ahmed Hafnaoui, Gregorio Paltrinieri, Elijah Winnington and Lukas Maertens have reshaped the 400m, 800m and 1,500m freestyle events. In breaststroke and medley, Adam Peaty, Qin Haiyang and Leon Marchand form a leading group with a clear technical gap over the rest.

In the women's events, Katie Ledecky, Ariarne Titmus, Summer McIntosh, Mollie O'Callaghan, Kaylee McKeown and Zhang Yufei have shared almost every major final berth for years.

As for Southeast Asia: Singapore has the most complete swimming system in the region, in both selection and sports medicine. Thailand and Indonesia have depth in certain events. Vietnam has two historic spearheads — women's medley, associated with Nguyen Thi Anh Vien, and men's distance, associated with Nguyen Huy Hoang — but thin overall depth.

The talent supply chain is the most troubling part. Vietnamese swimming runs on a small number of training centres, and each centre depends on a handful of key coaches. When a coach leaves, an entire cohort vanishes from the results map. No system records that disappearance.

The rules layer: a grey zone nobody teaches athletes

Swimming has one of the most detailed technical rulebooks in sport. The 15-metre rule limits the underwater phase after the start and after each turn in most events. In breaststroke, the restrictions are tighter still. Every turn type has its own technical description, and a faulty turn can bring disqualification.

Higher up sits anti-doping governance. This is where swimming has carried the heaviest trust burden of the past decade. The 2026 case involving a group of Chinese athletes who tested positive for a banned substance, and the way the world anti-doping body handled it, generated years of argument about procedural transparency.

For Vietnamese athletes, the problem is rarely intent. It is process education: whereabouts filings, out-of-competition testing, handling therapeutic-use situations. A procedural slip can destroy four years of training, and that is rarely taught systematically at grassroots level.

Numbers do not lie, but people always find a way to lie about numbers.

The career layer: the puberty wall

In women's swimming, the career curve is broken by a biological variable: puberty. Many Vietnamese female swimmers post their best results between 13 and 15, when their strength-to-mass ratio is most favourable. As the body changes, speed drops, and very few are supported through that phase.

Nguyen Thi Anh Vien is the exception, and that exception did not happen by accident. She sustained a peak career across multiple SEA Games cycles, alongside a long-term coach. A long career in this sport is almost always the product of a system, not an individual.

Among men, the curve arrives later. Nguyen Huy Hoang was born in 2026 and won an Asian Games medal at 18. He is the template Vietnamese swimming needs to replicate, and there is still no process for replicating it.

This section of the spreadsheet is also empty, and it is empty because we do not track athletes after they leave the pool. To understand why a generation of talent disappears, you need data about the disappearance. We do not have it.

The risk layer: the biggest risk sits in the process

When I ran the risk matrix across the nine dimensions, the output was a row of cells that could not be assessed: competitive risk, career risk, doping risk, rules risk, psychological risk, systemic risk. None had enough data to score.

The only entry I could record was a risk that belongs to no athlete: analytical-process risk. The data pipeline broke at the input layer, and all nine layers behind it collapsed in sequence. That was the most expensive lesson of August 13.

When the stadium empties, every model collapses. I rebuild from the charred data.

The public layer: hype cycles and the expectation gap

Vietnamese fans lived through a very high expectation cycle around Nguyen Thi Anh Vien, especially between 2026 and 2026. The way media framed it led many to believe a SEA Games medal was a direct launchpad to Olympic glory.

This is the point I always want to state clearly: expectation is not bad, but expectation without coordinates manufactures its own disappointment. When a regional champion is measured against a world yardstick, both outcomes are unfair — either the athlete is inflated too high, or they are treated as a failure.

Reputation is only a name. What remains is always how you read the race.

One encouraging sign is that coverage in recent editions has become less absolutist. But public data has not improved in parallel.

The industry layer: pools, markets and numbers we cannot forget

Swimming has a feature few sports share: its value extends beyond medals. It is a life skill. Vietnam is among the countries with a high rate of child deaths by drowning, with a figure frequently cited at around two thousand children per year. Every number here carries a weight entirely different from a hundredth of a second in a race.

On infrastructure, major cities such as Hanoi and Ho Chi Minh City have relatively developed pool systems, but rural and peri-urban areas remain severely short, in both quantity and water quality.

On markets, swimming sponsorship is unevenly distributed: concentrated on a few elite faces while the youth development system lacks long-term resources. The domestic swimming equipment market is largely imported, with margins sitting at retail. The content market — where I work — is barely tapped because there is not enough data to tell stories with.

This point matters: without data, there is no content. Without content, there is no audience. Without an audience, there is no sponsor. That chain begins with an empty spreadsheet.

Where I doubt myself

At this point I have to confront myself, because that is the mandatory part of the job.

Correlation is not causation. The fact that an athlete wins a SEA Games medal and never reaches continental level does not prove the SEA Games damaged their career. The fact that a famous coach is attached to a successful swimmer does not prove his method can be replicated. The fact that Vietnamese swimming has little data does not prove data will generate medals by itself.

My instinct — and I say this as someone in the personality group that always wants to reorganise everything — is to build a new model the moment the old system collapses. That is precisely the trap. Building a new model on a leaking data pipeline only creates a new error layer, more sophisticated and harder to detect.

That empty table on August 13 taught me the opposite of my instinct: the first task is to patch the pipeline, not to add another storey.

One thing also needs stating about the limits of models. xG is not wrong; football is simply irrational. After 2026, I learned to count the irrationality too. Swimming has its own irrationality: an athlete can swim faster in the heats and slower in the final for psychological reasons no metric captures. That irrational portion must be counted as a variable inside the model, not pushed outside so the charts look cleaner.

And one more thing: most analyses I read about Vietnamese swimming conclude too quickly. A meet happens, an article appears the same day, a conclusion lands immediately. I do not trust those conclusions, even when they match my own view. A single anomalous data point is not enough to conclude; sample size is part of the conclusion. If I have only one meet, I call it a hypothesis, not a trend.

In this trade, the hardest task is publishing a map of what you have not yet read. An honest data analyst is not someone who can explain everything. It is someone who draws the borders of their own understanding clearly, and takes responsibility for what lies beyond those borders.

One more question readers often ask me: why not write praise pieces, which are easier to read. Because praise requires no skill. If I merely summarised results sheets and added adjectives, I would not be practising data analysis; I would be practising results-sheet translation. The value I can add lies in stating the distance between a medal and a world final berth — and that distance is almost always larger than audiences imagine after a final.

A final reminder to myself: never let the writer's name grow larger than the number. When readers trust the analyst more than the data, the analyst has failed.

Signals for the next cycle

Three markers to watch, and I am locking them in here as an early prediction so I can grade myself later.

First, SEA Games 33 in Thailand in December 2026 will show whether Vietnamese swimming's next generation exists. If medals decline in the women's events, that is a signal about the development system, not about one athlete performing poorly.

Second, the 2026 Asian Games in Aichi–Nagoya will test the continental-gap hypothesis. A few final berths would be a good outcome for this cycle; a medal would be an anomaly requiring analysis, not an anomaly requiring celebration.

Third, and this is the indicator I care about most: whether a public split dataset for domestic swimming meets appears within the next two years. If it does, every analysis of Vietnamese swimming moves to a new layer — harder for lazy writers, easier for serious readers.

As for that nine-tab empty table, I am keeping it exactly as it is. Not deleted, not filled in. Because it is the only map that shows precisely where the data pipeline broke, and in this trade, knowing where you do not know is a more valuable asset than any elegant conclusion.