The Empty Analysis Sheet and a Table Tennis Writer's Discipline of Verification

Da Nang, 1:42 a.m. on January 15, 2026, a small apartment on Nguyen Van Linh Street. Central Vietnam rain fell steadily from afternoon into the night, the kind that neither stops nor pours. I reopened the Excel file I had built two weeks earlier, named BB_VN_2026_01. Nine sheets. The Serve sheet is divided into 9 zones, all still blank. The Receive sheet holds 6 classification codes, all still blank. The Rally Length sheet has a column counting how many times the ball crossed the net, still blank. The Contact Position sheet has a coordinate grid, still blank. The Equipment sheet has three rows waiting for rubber type, blade thickness and replacement date. I sat still for about forty minutes. The screen stayed on. The text cursor blinked in cell A1 of the seventh sheet, blinked and blinked, with nothing to type into it. I closed the file, shut the machine down, went to sleep. The next morning I sent my editor one short line: no analysis piece from me this week. Not because there was no tournament to write about. Because I did not hold a single data point solid enough to support one declarative sentence. People keep saying Vietnamese table tennis lacks deep analysis. That is true, but it describes the symptom rather than the disease. The problem sits one layer lower: a shortage of raw data. Football has Opta, has StatsBomb, has hundreds of thousands of tagged events every matchday. Vietnamese table tennis has scoreboards, photographs, and a few clips shot on phones from the stands. A scoreboard tells you who won. It does not tell you why. An 11-9 set and an 11-3 set can come out of the same tactical situation, separated by three balls clipping the edge. Read only the score and the writer is forced to guess. And guessing, for me, is a form of debt. Names are not in short supply. Nguyen Anh Tu, Tran Tuan Quynh, Dinh Quang Linh, Mai Hoang My Trang — names that appear regularly in the press after every SEA Games. But attached to those names are medals and a few post-match quotes, not a dataset. Nobody publishes first-three-shot point win rates. Nobody publishes a serve placement map for any player in the domestic system. I have known that feeling for a long time. In the summer of 2026, as a first-year student, I spent three full weeks on the World Cup semifinal between France and Belgium, the match France won 1-0. I redrew 11 positional maps from minute 1 to minute 90, identified the 4-5-1 block France used in defence, and worked out that Blaise Matuidi was stationed left of centre to lock down Kevin De Bruyne: the Belgian midfielder recorded only 64 touches and did not create a single chance. Those three weeks produced 1,200 words. But every sentence in that piece had video standing behind it. In 2026, when global football stopped, I sat at home collecting data from 50 V.League 2026 matches and built a spreadsheet to measure effective pressing — ball recoveries within 5 seconds of losing possession. The result surprised me: Hanoi FC were not the best pressing team, 12% behind Ho Chi Minh City, but their average transition time was 1.8 seconds faster. From that I drew a principle: what decides a match is usually not what gets counted most often in the news feed. At Euro 2026 I wrote a 2,500-word piece on Roberto Mancini's Italy, attached heat maps I had drawn myself from UEFA data, and closed with a cautious line: Italy are stable, but that is not yet enough to declare them champions. At the 2026 World Cup I spent four days redrawing Walid Regragui's 4-1-4-1 system, cross-checking 200 defensive phases across Morocco's 5 matches, and calculated that Yassine Bounou saved 17 shots while Romain Saiss won 11 aerial duels. All those pieces shared one thing: the raw data existed before I sat down to write. In domestic table tennis, the raw data does not exist. So the first task is not to write better. The first task is to manufacture the raw material yourself. THE MANUAL TAGGING LAYER A table tennis match in a domestic tournament runs 35 to 50 minutes on average, depending on the number of sets. I set up two camera angles: one fixed behind the baseline, one tilted 45 degrees at net height. A capture rate of 60 frames per second is the floor, because at table tennis tempo, 30 frames blur the contact point and I lose the ability to read blade angle. For every serve I log four fields: placement zone across 9 boxes on the opponent's half, spin code from three options (topspin, backspin, sidespin), perceived speed on a three-point scale, and the outcome of the third shot after the serve. The fourth field matters most. The reason is concrete. From semi-professional level upward, most points are decided inside the first three shots: serve, receive, and the opening attack. If I only log set scores, I do not know whether a player won because of good serving or because the opponent made unforced errors. Those two causes lead to opposite conclusions about the same match. The cost of this layer is roughly 2.5 hours of tagging per match, not counting camera setup and audio synchronisation. That is why I cannot process three matches a week. I choose to do less and do it clean. PROXY METRICS FOR SPIN There is no spin-measurement device within the budget of a self-funded writer. I use three proxies: blade angle at the contact point, the height of the ball's trajectory after leaving the racket, and the type of return error the opponent makes on the second shot. The logic is fairly simple. If the ball leaves with a steeply angled blade, a low trajectory and the opponent sends it skyward, the likely reading is heavy backspin. If the ball travels flat, the trajectory is level and the opponent pushes it long, the likely reading is topspin. Where the method fails is in the opponent's freedom to choose. A skilled player can deliberately loop the ball high to change the rhythm. In that case the proxy measures the receiver's perception, not the spin-maker's action. I state this limitation explicitly in every piece that uses the metric. Without that note, readers would assume I measured spin. I do not. FOOTWORK MAP AND RECOVERY TIME This is the part I carried over from football. For each shot I mark contact position by coordinates, the player's distance from the table edge, and lateral displacement compared with the previous shot. From that I derive two metrics: distance travelled per point, and the time from the end of a rally to the moment the player is back in a ready stance for the next receive. The second metric is a direct copy of the V.League 2026 lesson. Back then I found that the most control-oriented team was not the best at recovering the ball; the difference-maker was transition speed, 1.8 seconds faster. In table tennis, the interval between two points is the most neglected interval of all. Spectators watch the rally. I watch the silence between rallies. I have not published this metric for any Vietnamese player, because I do not yet have a sufficient sample. That is a deliberately empty column. Every hand-drawn diagram is a story the numbers cannot tell, but a diagram only earns trust when it rests on enough logged rallies. THE EQUIPMENT COLUMN The seventh sheet in the file is the one I use least and the one that irritates me most. It has three rows: rubber type, blade thickness, replacement date. Those three fields say nothing on their own, detached from the player's feel for the ball. But combined with rally data, they explain abnormal miss patterns. A player who has just changed rubber usually loses control on the second shot first, and only then on the third. Without this column, I would misattribute the cause of a form dip. The smallest detail on the court has its reason, including the details that are not on the court. THE MINIMUM SAMPLE THRESHOLD I set a threshold of 12 matches per player before making any statement about a trend. Twelve matches times 2.5 hours equals 30 hours of tagging for one person. A player who goes deep in domestic tournaments can reach that threshold within a season. Which means a personal model needs a season, not a week. That threshold has a price. It keeps me silent for weeks while other outlets keep publishing after every matchday. I accept it. I look at the diagram first and the reputation second. THREE RISK FLAGS I PLANT MYSELF Before every piece, I check three flags. First, does the technical claim have enough data behind it. Second, is the player in an adaptation period after an equipment change, since switching rubber or blade usually takes weeks to settle and all data in that window is noisy. Third, is the playing style being discussed countered by a specific opponent type. For the file BB_VN_2026_01, none of the three flags could be planted anywhere, because there is no data. That is why the file sits still. The counterintuitive part is that an empty analysis sheet is less dangerous than a half-filled one. An empty sheet forces the writer into silence. A half-filled one invites confidence, and confidence built on missing data is the hardest error to correct, because it wears the clothing of evidence. The second blind spot lies in the choice of what to measure. Analysts tend to measure what can be measured instead of what decides. Inside Vietnamese gymnasiums there are variables that directly affect results and that nobody writes into any column: humidity, floor grip, lighting, and the daytime job of the player. An athlete who works eight hours in an office and trains at eight in the evening has a completely different form curve in the fifth set from a full-time professional. That column does not exist in any statistical table of Vietnamese table tennis. But it exists on the court, every night. My conclusion may irritate readers: at the current stage, the most valuable column in my spreadsheet is the column I refuse to fill. Writing a wrong figure into an empty cell is worse than leaving it blank. The only way a Vietnamese table tennis analysis system becomes trustworthy is by publicly accepting the part it does not know, and stating how large that part is. When the pitch falls silent, I learn to listen to the data speak. But when the data falls silent too, the right move is not to speak on its behalf. If by June 2026 I have gathered 12 tagged matches for at least three domestic players, I will publish the first rally-rhythm model together with the entire raw dataset so that anyone can recheck it. If I do not have enough, I will publish that empty sheet itself, along with a list of what is missing and why I refused to fill it with speculation. A mature table tennis scene is not measured by how many articles get published. It is measured by how many blank cells get explained.