Nine Layers of Data and One Ethical Line in Sports Analysis

1:40 a.m. in Busan, the city quiet enough that I could hear the laptop's cooling fan. On the screen was the transfer-window tracker I had kept open for three weeks: each row a name, each column a piece of evidence. Row seventeen was blank. The source column empty. The transfer-fee column empty. The contract-length column empty. I looked at it for about two minutes, then typed one line into the notes field: “not enough information to assess.” Closed the laptop, turned off the light, went to sleep. The abacus never sleeps, but football does. People who write about football have to sleep too, unless they want to publish a line of speculation dressed up as data the next morning. From Busan to Munich: one night changed how I read a match. In 2026 I was fourteen, sitting in front of the TV watching South Korea play Germany in the World Cup group stage. I wrote on paper: Germany held 72% possession but managed only three shots on target; South Korea had five quick counterattacks generating 0.4 xG. I concluded that if the opponent lost focus late, South Korea could win 1-0. The match ended 2-0, with goals from Kim Young-gwon and Son Heung-min. My blog post was shared three hundred times. The 2026 World Cup taught me that a 1% probability is still data. But the bigger lesson lay elsewhere. I was right, and precisely because of that I nearly learned the wrong thing. Being right once does not prove the method is right; it only proves the denominator had not betrayed me yet. It took the 2026 lockdown for me to understand that. When the leagues paused for COVID-19, there were no matches left to write about. I stayed home for three months, collecting data from 380 Premier League matches of the 2026-20 season. I calculated Liverpool's PPDA at 8.2, the highest in the league, with only 22.1 xG conceded. During the lockdown I learned to hear data with my ears rather than my eyes — hearing the pressing in each duel, then checking it against the numbers. I wrote a two-thousand-word piece on the correlation between pressing intensity and defensive performance. A large forum reposted it. In the piece I made clear there were many confounding factors and that the correlation's strength was only about 70%. By Euro 2026 the method had its test. I used qualifying data to assess the teams. Italy had an average PPDA of 7.9, the lowest among the big sides, and an 82% pass-completion rate in the opponent's final third. I wrote that Italy would reach the semifinal or the final, even as Korean media shrugged. When Italy lifted the trophy, the old post resurfaced, and an editor reached out to offer a collaboration. I declined because I was still in school, but agreed to write for an amateur column. The Euros do not end with the final; they end when I finish my summary table. Since then, every analysis I write carries a methodology section: how many matches, how many metrics, where the data limits lie. The transfer window is when that section matters most, because the noise here is louder than at any other point in the year. Release-clause structures and the wage bill are the real story of any transfer. The fee is only the published figure; release clauses, agent fees and performance-linked payments are the submerged part of the iceberg. Readers are already drowning in rumours, and my job is to hand them a filter rather than one more rumour. The structure I use has nine layers. That sounds like a lot, but the logic is tight: each layer answers one question, and every conclusion must trace back to a specific information point. No information point, no conclusion. I call it the traceability principle. Layer one is patch and meta. In esports, this is where everything begins: an update shifts the numbers, and the entire tactical system pivots. The update itself says nothing. You need win rates, pick-ban rates, and the gap between the tournament server and the practice server before it counts as data. In football, this layer corresponds to semi-automated offside or how stoppage time is calculated — small changes that tilt how the whole match is read. Layer two is tournament format. Single-elimination and round-robin league play produce two entirely different upset probabilities. Schedule density, travel distance, the timing of a patch change mid-tournament — all are variables of preparation risk. Layer three is teams and players. Here I split two columns: paper strength and positional fit. A player's value is only an equation missing an unknown — the transfer fee measures expectation, not adaptation. In June 2026 I looked at Kim Min-jae's profile from Fenerbahçe: a 71% aerial-duel win rate, 2.3 tackles per match on average, a sprint speed of 32.5 km/h. I compared him with Napoli's existing centre-backs and saw the numbers fit Spalletti's high defensive line. On July 18 I published “Napoli, the right signature for the defence.” The deal went through. In that piece I kept the numbers strictly separate from the inference and stated the date of the prediction. That approach lets readers check me. Layer four is the regional landscape. A region's strength depends on the discipline: a region's standing in football differs sharply from its standing in a tactical game. Ignoring local context and applying one region's yardstick to another is the mistake I made most when I started, because my tactical reflexes were trained in a different football culture. Layer five is club finance. Sponsorship revenue, distributions from the league or publisher, the wage bill, incoming capital — these four cells decide whether a deal is an investment or a gamble. A club that overspends on one contract carries with it the pressure to sell someone next window. Layer six is rules and governance. I check competitive integrity, transfer and registration rules, contract compliance, and issues involving underage players. One question shapes this layer: if the worst happens, who bears the punishment? Layer seven is the risk profile, and this is the layer I put on the table first, not last. Four signals make me stop immediately: unpaid wages, suspected match-fixing, a patch aimed straight at a team's dominant style, and an injury to a key player. These four do not need deep analysis; they need to be stated before anything else is analysed. Layer eight is the media narrative. Every team carries a label: a new dynasty, an all-domestic roster, a generation's last dance. The prettier the label, the more easily the gap between expectation and reality becomes a cliff. I measure that gap with one question: how many matches of data does this expectation stand on? This layer is also where I raise refereeing. That referees treat big clubs and small clubs differently is not a conspiracy theory; it is crowd and media pressure, measurable through the number of VAR consultations, the review time, and how cards are distributed in decisive minutes. Layer nine is industry transmission. A change upstream — a publisher, an update, a tournament licence — flows down to the midstream of clubs, tournaments, and streaming platforms, then to the downstream of sponsorship, derivative markets, and mainstream adoption. This flow has a lag, and the lag is exactly where readers are most easily led. Nine layers, but only one line. That line is: when the input data is empty, the correct conclusion must be “not enough information to assess.” Not “possibly,” not “likely,” but a refusal to answer. I once received an analysis whose source-information section was completely blank. No title. No source. No information points. All that remained was one domain label: esports. The analysis still had all nine layers, all its tables, all its risk checkboxes. But every cell read “insufficient information.” At a glance it looked like a failure. Looked at closely, it was the most honest result the framework could produce. Every table is a cut, and every cut is a story — and when there is nothing to cut, the right story is the story of the gap. If someone filled that gap with a game title, a team, a player, a transfer-fee figure, the analysis would look more perfect and be entirely wrong. That is the trap I fear most in this trade. There is a paradox I have to state plainly. Correlation is not causation, and a beautiful analytical framework is not a correct conclusion. The more complete the framework, the more easily it is mistaken for finished work. Nine layers with full tables can convince a reader the job is done, when in fact not a single information point has been verified. A further danger comes from the chain. An empty analysis, if read as a complete one, generates imagined conclusions. Those imagined conclusions get cited, then cited again. By the third round, no one remembers it started from an empty cell. In the transfer window this loop runs many times faster than usual, because every rumour has someone who wants it to be true. The only defence is to classify news by source and by impact, then state the confidence level right beside the conclusion. I do not say “certain.” I say “70% strength.” I do not say “a source close to the deal.” I say “unverified, not entered into the table.” It sounds slow. But the speed of a correct piece is always slower than the speed of a wrong one, and I choose slow. This transfer window will bring more blank rows. There will be names tied to clubs where no one has verified the fee column, the length column, the source column. The question I ask myself about each such row is simple: if this row is blank, do I dare leave it blank? If the answer is yes, then my table is still worth trusting.