Nine Layers of Tennis Data and the Lesson of an Empty Spreadsheet

On July 14, 2026, Roger Federer served in the 16th game of the fifth set on Wimbledon's Centre Court. The score was 8-7, the points 40-15. Two championship points rested in his hands, two serves away. When the scoreboard froze, Novak Djokovic had won 7-6(5), 1-6, 7-6(4), 4-6, 13-12(3) after 4 hours and 57 minutes, the longest Wimbledon final in the tournament's history.

The official match statistics told a different story. Federer won 218 points; Djokovic won 204. Federer struck 94 winners; Djokovic struck 54. The 37-year-old Swiss served better, approached the net more often and controlled most of the long rallies. He still lost.

I revisit that match for a reason other than any argument about luck. It is the cleanest example of a problem anyone working with tennis data meets every week: a spreadsheet describes a match, but a spreadsheet does not hand out trophies. Between those two things lies a gap, and a writer must know where he stands inside it.

The Input Gate

Drawing on my experience tracking matches over more than twenty years, from evenings watching Masters events on cable in the United States to nights spent following Challenger tournaments to gather numbers for articles in Vietnam, I distilled a single operating rule: check the input before you analyse.

My working framework has nine layers, moving from technique and tactics, through data and form, tournament structure and scheduling, tour landscape and player standing, rules and governance, team and management, risk, media narrative and expectation, all the way to the transmission flow of the entire industry.

Ahead of all nine layers sits one gate: does the input exist at all? If a report cannot carry a single concrete fact, no player, no tournament, no date, no score, then the nine layers behind it become meaningless. The only correct output at that moment is an input diagnostic, not a prediction.

Data is never in a hurry. The person in a hurry is the one who gets it wrong.

Layer One: Technique Cannot Speak for Nerve

At the technical layer I sort every player along three axes: serve structure, return structure and surface adaptability. These axes are independent. A player can serve at a top-five world level while returning at a top-forty level, and that very gap decides how many tiebreaks he wins.

From the 2026 season, the ATP introduced a 25-second serve clock at main-tour events. It is the most underrated technical change of the decade. It does not make a serve stronger or weaker, but it shortens the time a player is allowed to process pressure between points. Players with long, rigid preparation routines are hit harder than instinctive ones. That is the kind of variable a statistical table never records.

The off-court coaching rule behaves the same way. When the ATP trialled and then formalised coaches communicating with players during changeovers, the boundary between "the player solves it himself" and "the player is being directed" dissolved. For a data analyst, that means every model built on a player's capacity for self-correction needs rewriting. No dataset will announce this to you. You have to read the rulebook.

Layer Two: When the Score Is Binary

Tennis has the harshest data structure of any mainstream combat sport. Football has xG, a continuous measure of chance quality. Tennis has no equivalent. Every point carries only two values: won or lost. A best-of-five match lasting nearly five hours contains roughly 200 to 220 points. That is an extremely small sample for any statistical inference.

The best metrics available, first-serve points won, return points won, break-point conversion and the Dominance Ratio developed by Jeff Sackmann at Tennis Abstract, are all ways of compressing those 200 points into a few readable figures. They are useful. They are also easy to misread.

Break-point conversion is the classic case. A player converting 1 of 8 break points gets labelled mentally weak by the press. But the denominator is eight. Across eight attempts, the difference between 1 of 8 and 4 of 8 sits entirely inside statistical noise. No conclusion about mentality can be drawn from eight observations. That is why I always place an error horizon beside every metric, even when it makes the piece less attractive.

The ranking-points structure, by contrast, is far clearer. A Grand Slam awards 2,000 points to the champion and 1,200 to the runner-up; a Masters 1000 awards 1,000; ATP 500 and ATP 250 events award exactly what their names say. Because points are defended on a 52-week cycle, every player carries pre-determined risk windows. A place in the top eight for the ATP Finals can be decided not by October form, but by a player dropping 600 points back in March.

Layer Three: Tournament Tiers and the Scheduling Trap

The professional tier system is sharply defined: Grand Slam, ATP Finals, Masters 1000, ATP 500, ATP 250, then Challenger and ITF. Each tier has its own entry requirements, prize money and calendar slot.

For Vietnamese fans, the distance between these tiers matters more than its surface appearance. A Vietnamese player winning an ITF title or reaching the deep rounds of an Asian Challenger collects points worth only a fraction of what entering the main draw of an ATP 250 delivers. Domestic media tends to collapse all of it into "an international title". A spreadsheet does not collapse. A spreadsheet adds.

I once wrote about this gap and received no small amount of irritated feedback. But without tiering, readers will never understand why a player can win repeatedly at small events and still sit outside the world's top 300 for years.

Layer Four: Tour Landscape and Standing

The tour landscape always splits into four groups: title contenders, the top-10 seed tier, the top-30 backbone and the top-100 fringe. The borders are not fixed. They shift with each generation.

Generational comparison is the most uncomfortable job in tennis analysis, because it demands normalising data across decades of differing conditions, surfaces and calendar density. A claim that this generation is stronger than that one almost always exceeds what the data permits. What I can do is compare the share of major titles held by each age cohort at a given moment, then describe the trend rather than judge the value.

Layer Five: Rules Are Data

Rules and governance are the most neglected layer, yet every technical committee decision flows into the statistical tables a few months later.

Medical timeouts, the serve clock, off-court coaching and the regulations governing ranking-point structures all change competitive behaviour. Alongside them sits a stricter compliance layer: anti-doping and match integrity. Tennis has lived through match-fixing cases at Challenger and ITF level, where prize money is low but financial pressure is high. That is why governing bodies invest in betting monitoring even at events nobody broadcasts.

For a data writer, this is a layer to track continuously, because one rule change can invalidate an entire prior model.

Layer Six: The Team Decides More Than the Forehand

A professional player is a small enterprise: head coach, fitness coach, physiotherapist, nutritionist, commercial agent and sometimes an entire dedicated analytics unit.

The fit between player and coach is the hardest variable to measure in this layer. It appears in no statistical table, yet it explains form shifts the numbers miss. When a player changes coach and results improve, the media credits the newcomer. I always check whether something else coincided: an easier draw, a healed injury, or a change in serve structure.

Layer Seven: Risk

I sort risk into six groups: injury and fitness, points defence and ranking, career, rules, commercial and media, and systemic risk.

Systemic risk is the most underrated group. The 2026 pandemic proved it: when the calendar froze, the entire protected-ranking system became meaningless, and young players lost the advancement opportunities a normal schedule would have handed them. The crowd can leave the stadium, but physical data never takes a day off.

Layer Eight: Narrative and the Expectation Gap

Tennis media runs on heat cycles. A player winning three straight matches becomes a title contender. Losing two becomes a crisis.

The only check is comparing expectation with reality numerically. If a player is described as surging while first-serve points won stays flat and return points won inches up only slightly, that is a media story, not a spreadsheet story. I still remember the feeling of predicting Germany's collapse at the 2026 World Cup based on a pressing coefficient falling from 8.1 PPDA to 12.6 and average distance covered dropping 6.2 kilometres per match. When Germany lost 0-2 to South Korea and went out, colleagues who had called me a statistics fanatic changed their tone. But the lesson was not in getting the call right. It was that I only dared speak when ten matches stood as evidence, not three.

Layer Nine: The Industry's Transmission Flow

Finally comes the transmission layer of the whole industry, running from youth development, equipment and facilities, through players, tournaments, broadcast rights and sponsorship, into derivative markets.

For the Vietnamese market, this layer deserves more attention than it gets. The number of ITF and Challenger events staged domestically directly affects the points-accumulation opportunities of players such as Lý Hoàng Nam and Nguyễn Thùy Linh. A low-tier domestic event does not create a star, but it creates data: matches played, points earned, meetings against opponents of comparable level. Without that data, every analysis of Vietnamese tennis is merely an educated guess.

The Counter-Intuitive Point

Layers eight and nine together reveal something: this industry rewards the person who delivers a verdict, not the person who withholds one.

A headline saying Player X has rediscovered his form always travels faster than the sentence not enough data to conclude. But in tennis, the state of insufficient data is the default, not the exception. Each match holds only a few hundred points. Each surface generates a different behavioural set. Each minor injury wipes out a month of data.

Correlation is not causation. A player whose first-serve points won rises and who wins more in the same month may simply have met weaker opponents. If I write that the serve has come back, I am selling a conclusion the data has not yet bought.

People remember results. I remember the conditions that produced them.

What to Track Next

Over the next twelve months I will track three signals at tennis's data layer. Richer movement data, including court position, ball trajectory and reaction time, could move tennis closer to an xG-style metric if it is made public. How low-tier Asian tournaments, Vietnam included, are folded into the ranking system will shape the next generation of players. And every time the rules on off-court coaching or the serve clock shift, an older layer of data turns worthless, forcing analysts back to the input gate.

Every serve is a hypothesis. My job is not to believe it, but to verify it with the score.