A Tuesday morning at a table tennis training center in Tan Binh District, a coach puts down his notebook and opens an external hard drive. On the screen appear 41,283 serves that have been manually coded over four years. He presses one key, and a heat map radiates outward: the short sidespin serve to the opponent's left succeeds 68% of the time but accounts for only 9% of total usage. That is a rare moment when the boundary between intuition and data in Vietnamese table tennis becomes tangible: a number that simultaneously indicts laziness and warns of arrogance.

I have followed Vietnamese table tennis since the years when scoreboards were written by hand in chalk and the spectator stands had only three rows. Back then, everything resided in the coach's head — the feel for friction thickness, for the bounce of the floor at Phan Dinh Phung arena, for how a 15-year-old breathes before serving. Twenty years later, most of those feelings still reside in the head, but with three additions: a phone shooting 240 frames per second, free comparison software, and a generation of young coaches who can read spin-distribution charts the way they read scoreboards.

Table tennis is an undervalued sport in Vietnam's sports market. According to the annual report of the Vietnam Table Tennis Federation published in early 2026, the country has approximately 210 officially registered table tennis clubs, concentrated mainly in Hanoi, Ho Chi Minh City, Da Nang, and Hai Phong. The number of athletes with updated national rankings is 1,470, of whom only 38 are in the national youth squad. Against an estimated 200,000 regular amateur players nationwide according to tournament organizers, the conversion rate from amateur to professional is below 0.02%. That is a frightening number through the eyes of an administrator, but an opportunity through the eyes of a tactical analyst: if the training system could improve the conversion rate by 10%, Vietnamese table tennis would gain nearly 20 potential athletes per generation.

Why talk about data in a sport where outcomes are decided in 0.3 seconds? Because within that brief window, every decision has already been shaped. A serve is not an improvised act; it is the result of a chain of choices — where to stand, how many degrees to rotate, how high to toss the ball, at what point on the racket to make contact. Each of those choices can be recorded, counted, and compared. The question is not whether data is useful, but whether Vietnam's table tennis training system is capable of absorbing data without turning it into meaningless decoration.

Over the past two years, I have had occasion to observe how training centers in Ho Chi Minh City and Hanoi handle this problem. The most common answer is: film, review, then... film again. A loop without a destination. The problem is not technology — a mid-range phone today can shoot 240 frames per second and export in a format that free analysis software can read. The problem lies in the absence of a system of questions. Data does not generate meaning on its own; meaning arises when someone poses specific questions to the data.

Take a verifiable example. At the 2026 national table tennis championship held in Da Nang, the coaching staff of a men's team conducted a three-month experiment before the tournament: coding all serve and receive sequences of six key athletes in practice matches, classified by three variables — contact position, spin type, and ball length. The results showed that 62% of their lost points came from situations where the athlete received the ball 30 cm behind the optimal position. In a sport where 30 cm is equivalent to two footsteps, that is the distance between winning and losing.

This is the point where I want to pause. When data tells a coach that every time an athlete is 30 cm late, the loss rate spikes, the intuitive reaction of most people is: force the athlete to practice standing 30 cm further back. That is a fundamental error in biomechanics. Receive position is not an independent choice; it is a consequence of how spin is read and how the nervous system responds. If you only shift position without addressing the speed of spin recognition, you will create an artificial stance and ruin the entire balance.

Data does not indicate what to fix; data only indicates what is happening. Deciding whether to fix the underlying structure or trim the surface manifestation is the work of an experienced coach, and this is precisely where most table tennis training centers in Vietnam go wrong.

Another example, from a different angle. In table tennis, win-rate data has little value without data on the structure of point sequences. An athlete may win 70% of long rallies but lose 80% of decisive rallies at 9-9. This is often explained by the concept of "competitive psychology," but tactical analysis points to a much more specific cause: at 9-9, athletes typically shorten their swing amplitude to reduce risk, which reduces ball speed by roughly 12-15%, giving the opponent the opportunity to attack first. For this reason, some coaches have introduced the "9-9 drill" — where every rally begins from a 9-9 score and athletes are forbidden from narrowing their swing amplitude. This is a drill designed to counter the body's natural defensive reaction, and it can only emerge when the coach has examined data on the correlation between swing amplitude and win rate at decisive points.

There is an interesting detail that few notice: Vietnam's top athletes are often those with the least personal data. The reason is practical — training centers have limited budgets and prioritize equipping young prospects who still have many years of development. The consequence is that elite athletes often play on intuition forged from tens of thousands of hours, but lack the ability to objectively cross-check themselves. This creates a paradox: the best players are the hardest to improve, not because they have reached their limit, but because they lack the tools to see their own limits.

I once witnessed a national-level athlete review footage of himself for three consecutive hours, and the only thing he said afterward was: "I didn't know my serve was that bad." That is not a lack of confidence; it is a truth obscured by feeling. In table tennis, the feeling of a good serve is often established by the sound of the ball hitting the table and the sensation of the racket in the hand, not by its actual outcome. A serve that sounds very "satisfying" can be entirely harmless to the opponent.

There is a question I always pose to the coaches I interview: "If you had to throw away all the data and keep only one number to train by for a year, what would that number be?" The best answer I ever received was not win rate, nor the number of successful serves, but "the percentage of rallies in which the athlete actively changes the tempo before the opponent does." That is a difficult number to measure, requiring manual coding, but it reflects the true tactical essence of table tennis: tempo control.

In table tennis, matches typically proceed at a tempo of 60 to 90 rallies per game. In each rally, one can hear the sound of the ball hitting the table — a structured sequence of sounds. Experienced coaches can read the rhythm of a match through these sounds, like a musician reading the beat of a piece. When the rhythm breaks — a serve slower than usual, a return faster than usual — that is a sign of a tactical shift. And a tactical shift, at the highest level, is the only thing that can turn a match around.

From a data perspective, this is a major blind spot. Automated analysis systems today can count strokes, ball speed, and placement, but they are very poor at measuring "tempo" — that is, the interval between strokes. In tennis, systems like Hawk-Eye can measure this interval, but in table tennis, cost and space constraints make such systems nearly nonexistent in Vietnam. Therefore, manual coding remains the primary tool, and this is precisely what creates the biggest barrier: time.

Let us calculate. A national-level table tennis match can last from 35 to 50 minutes. If coded manually at a detailed level (position, spin type, length, tempo, outcome), an experienced person needs about 3 to 4 hours per match. With a team of 8 key athletes, each playing 3 matches per month, the workload is 96 to 128 hours per month — beyond the capacity of most centers. This is precisely where training centers give up, and also where technology needs to intervene.

But which technology? Machine-learning-based motion recognition systems today can automatically detect strokes with fairly high accuracy under stable lighting conditions, and can classify strokes into basic groups with over 85% accuracy. This means that much of the manual coding work can be automated. However, there is a paradox: these systems perform best when the input training data is large enough, and that data must come from the very environment in which they will be used. A system trained on international matches will perform poorly in a Vietnamese arena with different lighting and different flooring.

This is one of the points I consider most important in the entire story: Vietnamese table tennis does not lack data, but lacks a structure to turn data into questions. And that structure cannot be bought with foreign currency; it must be built from within, by people who understand the rhythm of their own arena.

Over the past two months, I have followed a group of young coaches in Hanoi undertaking a small but meaningful project. They selected three young athletes, filmed all training sessions over four weeks, and used open-source software to automatically detect strokes. The result was a raw dataset of about 12,000 strokes, but it took nearly two weeks to calibrate the software to the arena's lighting conditions. That is a hidden cost most centers do not anticipate.

But the more valuable lesson lies elsewhere. When they compared the automated data with the manual coding of an experienced coach, they discovered that the software counted strokes accurately but missed the most important strokes — those executed with the intention of changing tempo. Those are strokes that are technically indistinguishable but temporally distinct. And it is timing that creates victory.

A table tennis match can be read at three layers. The first is the technical layer — whether this stroke is correct or not. The second is the tactical layer — whether this stroke serves the rally's objective. The third is the tempo layer — whether this stroke can change the match. Current data systems mainly address the first layer, occasionally reach the second, and almost never touch the third. That is why data, no matter how abundant, can never replace the eye of a good coach.

Now consider the other side: the athlete. In table tennis, each athlete has their own "language" — a set of signals they use to communicate with the opponent through the ball. Some speak through the silence of a short serve; some speak through the noise of a loop. Data analysis can indicate whether this language is effective, but it cannot indicate whether this language should change. An athlete who has built a personal language over tens of thousands of hours of training will take a very long time to learn a new language, and during that period, performance will decline. This is a price most training systems dare not pay.

I have witnessed a specific case: a female athlete with a very strong topspin serve, but a success rate only at an average level because opponents had grown accustomed to it. The data said she should change. But when she tried to change, she lost her feel for the racket, and over six months her ranking dropped from top 5 to top 12. She then returned to her old serve, but with a small adjustment in standing position — a change no data system could suggest, yet far more effective than a complete overhaul. That is a lesson in the patience of data.

What about the psychological dimension? In combat sports, psychology is often considered the decisive factor, and table tennis is no exception. But psychology in table tennis has a peculiar characteristic: it is governed by speed. A table tennis rally can last from 2 to 5 seconds, far shorter than a tennis rally (5-10 seconds) or a football sequence (15-30 seconds). This means psychological decisions must be made in an extremely short window, almost at the reflex level. And reflexes cannot be trained by saying "calm down"; they must be trained by creating repeated situations.

That is why the "9-9 drill" I mentioned above is effective. It creates a pressure situation repeated over and over, and through repetition, the athlete gradually builds a new reflex — one that does not shrink under pressure. This is an application of the principle sports psychologists call "controlled exposure," and it requires no data technology whatsoever. It only requires a coach who knows how to ask the right question.

There is another aspect I want to address: the role of referees and competition rules. Table tennis is a sport with relatively stable rules over decades, but recent changes — particularly the adoption of the 40+ mm ball and changes to service rules — have significantly altered the structure of matches. A larger ball means less spin, slower speed, and longer rallies. This means tactics built on heavy spin and high speed become less effective, while tactics built on control and patience become more effective. This is a change Vietnamese coaches have recognized, but adjusting the training system to match it remains slow.

The reason is practical: training centers in Vietnam typically have few practice balls, and 40+ mm balls are significantly more expensive than the old ones. A center with 200 practice balls would need about 15 million dong to replace them all, and with limited budgets, this is often postponed. But the consequence is that young athletes trained with old balls will develop techniques suited to old balls, and when they transition to competing with new balls, they will struggle. This is a form of structural disadvantage that data can detect but cannot resolve.

A training system cannot overcome its material limits by better data analysis. But it can use data to identify precisely where those limits lie, and from there make more meaningful investment decisions.

In that context, there is a question I think Vietnamese table tennis administrators should ask: if you could invest in only one of three areas — facilities, analysis technology, or coach training — which would you choose?

My answer, based on what I have observed, is coach training. The reason is simple: a good coach can work effectively with modest facilities and limited technology, but a poor coach cannot leverage the best facilities or the most advanced technology. And in table tennis, where every decision occurs in a fraction of a second, coach quality is most evident in the ability to read what is not written in any data table.

This does not mean denying the value of data. On the contrary, data can help coaches ask better questions. But it can only do so if the coach already has a solid enough knowledge base to understand the answers. A coach who does not understand why a sidespin serve is effective will not be able to understand data saying it is 68% effective — and therefore will not be able to decide whether to strengthen or change it.

Over the past two years, I have seen some positive signs. Some training centers in Ho Chi Minh City have begun organizing seminars on data analysis in table tennis, attracting dozens of young coaches. Some sports universities have introduced data analysis into their curricula. And some former athletes have become pioneers in applying technology to training.

But a large gap remains. Most training centers still rely on the personal experience of the head coach, and the transmission of this experience to the next generation still occurs mainly through direct observation, not systematization. This means each generation of coaches must relearn what the previous generation knew. That is an enormous waste of time and capability.

There is a way to break this loop: build an open database of Vietnamese table tennis matches. Not commercial data, but public data, usable by any coach. Such a database would allow young coaches to access hundreds of coded matches, and from there draw lessons that previously could only be gained through years of direct observation.

But such a database requires significant initial investment — not only in technology but in human resources for coding. And it requires cooperation among training centers, which is currently very limited due to competitive reasons. This is a problem sports administrators need to solve, but it is also an opportunity for those who want to lead.

In the meantime, individual coaches can start with small steps. They can select one match per month and code it in detail. They can share data with colleagues. They can organize discussions about what the data shows. These small steps, if carried out persistently, can create a cumulative effect.

I recall a story an veteran coach told me. He said that in the 1990s, when Vietnamese table tennis faced severe facility constraints, coaches often had to make their own tables and craft their own rackets. But precisely under those conditions, they trained athletes with very solid fundamental techniques. When conditions improved, these athletes were able to go further. The lesson is: material conditions can limit the speed of development, but cannot limit the quality of the foundation.

This holds true for data in a similar way. A center without the budget for expensive analysis systems can still build a solid data foundation using free tools and a disciplined manual coding process. What matters is not the technology, but the method.

That method can be summarized in three principles. First, start from a specific question, not from data. Second, choose a small number of variables and track them consistently. Third, always cross-check data against direct observation to verify plausibility.

These three principles sound simple, but in practice they are often violated. Many centers start by buying a camera system and software, then think about what to analyze. The result is that they have a lot of data but no answers to any specific question.

Conversely, some of the best coaches I know have no technological tools at all. They have only a notebook and a pen. But they record systematically, and they ask questions consistently. Over many years, their notebook becomes a personal database more valuable than any software.

This is what I want to emphasize: data is not a technological product; it is a habit of thought. And this habit can be developed at any level, with any resource.

When I look at the overall picture of Vietnamese table tennis over the past five years, I see a slow but clear shift. Young coaches are beginning to use data more naturally. Young athletes are beginning to review their own footage more systematically. Training centers are beginning to invest in technology more intelligently.

But this shift is still very slow compared to the pace of world table tennis development. Leading nations such as China, Japan, Germany, and South Korea built comprehensive data analysis systems years ago. They have dedicated data analyst teams, automated camera systems, and massive databases of international matches.

This gap cannot be closed overnight. But it can be narrowed if Vietnam focuses on what it can do best: training coaches with analytical thinking, people who can work effectively with limited resources.

And that is why I believe the data revolution in Vietnamese table tennis will not begin in a high-tech laboratory, but from a notebook and a right question.

In the coming months, I will continue to follow how different training centers handle this issue. I will look for examples of coaches who have found ways to combine data with intuition effectively. And I will try to answer the question I posed at the beginning of this article: does Vietnamese table tennis have enough capacity to absorb data without turning it into meaningless decoration?

The answer, as with all answers in sport, will come from the arena. Not from the data table.

In a sport learning to count, the best are those who know when to stop counting and start listening.