When Data Is Empty: The Sports Analytics Industry and the Nightmare of Foundationless Reports

At 2 a.m. on March 15, in Busan, I reopened a thirty-seven-page report an independent analytics firm had sent me. The cover page stated "Summer 2026 Transfer Potential Assessment Report." A complete table of contents. Form curves, comparative metric tables, projected goal contributions. Everything was there — until I reached the appendix with the raw data.

Not a single figure.

No, I'm mistaken. A few empty cells were tagged "data processing," and the rest was just dashes. But on the first page, the status line still read "Complete — Version 1.0." The person who emailed it to me, a young analyst I had once mentored, did not realise that every conclusion in the report rested on nothing.

And this is what kept me awake. Not the missing data itself. Rather, the fact that a formally complete document could be produced out of thin air — and could have entered a club's system, triggered a bad transfer, and gone undetected.

That is the story I want to write today. Not about a specific match, but about something more dangerous: confidence generated by empty data.

I began writing about sports data in 2026, as a first-year student in Busan. My first piece to draw attention — many in the industry still remember it — analysed Asan Mugunghwa, then top of K League 2, and concluded they would collapse late in the season. My basis was a simple metric: their xG per match was only 1.02, while Busan IPark, ranked below, stood at 1.48. A team leading the table without generating enough quality chances to sustain its position.

The notable thing is not that I was right — Asan finished fourth and lost in the playoffs. The notable thing is that I never used the league table to reach my judgement. I used data. At the time, a student blog reaching 2,000 views was an enormous figure. But the lesson was not about views. It was that I had grasped something many in the industry still have not: data only has value when it actually exists. Don't trust the league table, ask xG. The table tells the past, data tells the future.

Nearly a decade later, sports analytics in Vietnam and Asia has changed enormously. K League clubs hire full-time data analysts. Esports teams in the LCK, VCS, and LPL have analytics departments including both tactical coaches and data engineers. Major tournaments like the V.League or VCS broadcast with real-time statistical overlays. But alongside that growth, a new disease has appeared: we have built an ecosystem where the shell of analysis — clean charts, advanced terminology, tidy slides — can be produced far faster than the layer of data genuinely beneath it.

In other words, we live in an era where the form of analysis has overtaken its content.

Let me tell a second story. In 2026, at the World Cup in Russia, I wrote a controversial piece about South Korea's 2-0 win over Germany in Kazan. Germany's PPDA in that match was 5.8 — extremely low, meaning they pressed very hard. Many analysts, including prominent voices in Korea and Asia, used this figure to conclude that coach Shin Tae-yong was lucky, that South Korea only won because Germany collapsed entirely.

I dug deeper. I split the data into 15-minute intervals and found something different. Germany's highest running distance came between minutes 60 and 75, and their pressing system broke down after Kim Young-gwon was substituted on. In other words, PPDA 5.8 was a correct figure placed in the wrong context. It did not describe the whole match. It described a specific window. I wrote a rebuttal and published it on an Asian football forum. Three weeks later, FIFA released a report confirming exactly what I had said.

I tell this story not to boast. I tell it because it taught me a lesson that, now, looking at a 37-page report with no data, I realise is more valuable than ever. A metric can be technically correct and semantically wrong. And a report can be formally correct and substantively wrong — far more dangerously so. I was once attacked for daring to question PPDA. FIFA confirmed it. But what mattered was not that I was right. What mattered was that I did not rush to conclude from a single metric.

From 2026, when the pandemic forced European and Korean leagues to play in empty stadiums, I began collecting data from 214 matches in the Bundesliga and K League 1 between May and August. People called it a natural experiment. I called it a chance to measure luck. The result: home win rate in the Bundesliga dropped from 43.2% to 37.8%, and average goals per match rose from 2.79 to 3.12. 214 empty-stadium matches taught me: home advantage is data, not just atmosphere.

When I published that small study on Medium, an editor at Football Analysis reached out to collaborate. It was my first time working with a professional editor, who forced me to standardise everything: comparison tables required, data source footnotes clear, language neutral. And most importantly, I had to learn to distinguish "weak data" from "non-existent data."

In sports analytics, these two problems are often conflated. Weak data means few matches, small samples, or unrepresentative samples. You can analyse, but you must state the limits. Non-existent data means you have nothing at all. No figures, no sources, no observations. In that case, any conclusion you draw is fiction, whatever name you give it.

And this is the industry's problem: we lack a system strong enough to detect fiction.

Let me explain more clearly. When a traditional football analyst claims "team A presses better than team B," he can be wrong, but wrong in a verifiable way. You can rewatch the match, count pressing actions, compare ball-recovery positions. Even without numbers, he still has observational grounds. But when a data analyst claims "team A has a higher xG/90 than team B" without actual xG data, he is not wrong in a verifiable way. He is wrong in a disguisable way. Because the form of the statement — technical terminology, assumed values, assumed charts — creates a sense of precision that the content does not possess.

This is why nobody spotted the problem in that 37-page report. It looked too professional to be wrong. And that is the most dangerous thing.

Now, to esports. I have worked at the intersection of football and esports since 2026, when I was both a competitor and a tournament organiser before moving into media. And I can tell you that in esports, the "empty data" problem is more severe than in football, for three reasons.

First, esports data ecosystems depend on publishers. Riot Games, Valve, Tencent, Krafton — each publisher discloses data differently. Some key metrics — map movement counts, pick-ban rates at semi-pro level, internal economy indicators — are never public. Without access, analysts can be pushed into situations where they "must" reach conclusions without grounds. And when pressured, humans tend to speculate rather than admit.

Second, meta change in esports is far faster. A League of Legends patch can invert the power order within two weeks. A CS2 update can turn a useless weapon into a meta one. This means old data can become obsolete fast enough that using it can lead to wrong conclusions. But it also means many analysts, lacking new data, still try to use old data. The result is a kind of "grounded fiction" — not entirely invented, not entirely correct.

Third, and this is what I consider most dangerous, esports culture prizes speed and decisiveness. Coaches, analysts, and casters are valued for fast judgements. "I don't know" is a discouraged sentence in a pre-match analysis session. This creates pressure to say something, even when that something has no foundation.

I recall a time when I worked as a transfer market administrator for a K League 1 club in June 2026. I proposed signing Lee Kang-in from Mallorca for 8 million euros. My data showed he was top 10 in La Liga for chances created per 90, at 2.8, higher than Isco. But the board rejected it, arguing he "didn't show defensive capability."

Six months later, Lee Kang-in shone and helped Mallorca survive. My club finished eighth.

I collected all the emails, data reports, and meeting minutes to write a 15-page internal analysis, sent to the board, acknowledging the process failure without blaming any individual. What I learned was not "data is always right." It was: "when data is absent, people substitute bias, and bias always looks like truth."

A transfer fee is a value one person is willing to pay. True value is a data value that needs no negotiation. But if data does not exist, even true value becomes a narrative. And narrative can always be more attractive than truth.

I want to pause here to discuss an aspect I consider important but rarely debated: the organisational one.

In most clubs and esports organisations, data analysis is not done by an individual. It is done by a team, and results are reviewed by others. This creates a phenomenon I call the "accountability gap." When a report passes through multiple layers, no one is ultimately responsible for its accuracy. The report's author says "I had data." The reviewer says "it looked reasonable." The decision-maker says "I relied on the report." And no one checks whether the underlying data actually exists.

I have witnessed this more than once. In Korean football, transfers have been made on reports that later turned out to be compilations of news and impressions. In esports, roster decisions have been made on analyses later revealed to be written by people who never watched the matches live. What is frightening is not the individual mistakes. What is frightening is that those mistakes are never detected, because the system has no mechanism to detect them.

I once saw a Vietnamese esports team prepare for an international tournament based on opposition analysis from a report generated by an automated tool. The report contained full statistics on win rates, pick rates, and average match durations. But when I checked the data source, I found the tool only pulled data from matches streamed on one specific platform, and the opponent's most important matches — regional qualifiers — were not in the dataset. The Vietnamese team prepared based on an incomplete picture, and nobody knew.

This is the most dangerous form of empty data: it is not entirely hollow. It contains part of the data. Only, that part does not represent the real picture. And precisely because it is not entirely hollow, it easily slips past reviewers.

In football, I have seen this when analysing smaller V.League teams. A team might average 62% possession, but most of its passes are sideways in midfield, creating no threat whatsoever. If you look only at possession, you think they are playing well. If you look at key passes, box entries, or xG, you see a completely different picture. Possession can be the most deceptive metric in modern football, because it measures ball ownership rather than ball usage quality.

In esports, a similar phenomenon occurs with KDA. A player can have a very high KDA, but KDA does not say whether he participated in decisive teamfights, whether he generated map pressure, whether he secured vision for teammates. KDA is an easy metric, and because it is easy, it is often used in place of harder-to-measure but more meaningful ones. This is a subtle form of empty data: the data exists, but it does not measure what everyone thinks it measures.

And when a metric is misused long enough, it becomes part of the culture. People start to believe in it. Decisions start to rest on it. And when someone questions it, they are treated as an ignoramus. I went through that with PPDA. I know what it feels like to be attacked for asking the right question. And I know that in the long run, data always wins — but only when someone is patient enough to read it correctly.

I think this is the moment to state something I believe deeply, and it runs counter to the prevailing view in the industry today.

In modern sports analytics, there is an implicit belief that more data is good, less data is bad, and no data is a disaster. From that, we build complex pipelines to collect ever more data. We buy access to paid data platforms. We hire data engineers. We build dashboards. And we believe that with enough data, problems will solve themselves.

But I think that is a mistake in approach.

The biggest problem in sports analytics today is not a lack of data. It is a lack of ability to recognise when data does not exist. And that is a process problem, not a technology problem.

Think about this. If an analyst is trained to always produce a conclusion, he will always produce a conclusion. If a process is designed never to stop, it will never stop. If a system is built to always return a report, it will always return a report. And in all those cases, empty data will not lead to a halt. It will lead to fabrication.

The solution is not more data. The solution is building "fail-closed" capability — the ability to halt safely when input is invalid. In software engineering, this is a fundamental principle. A good system should not try to continue when essential data is missing. It should stop, report an error, and wait for correct data. But in sports analytics, we largely lack this principle.

PPDA 5.8 sounds scary, but a team running out of gas in the 75th minute is what's truly scary. I was once attacked for daring to question PPDA. FIFA confirmed it. But the real lesson was not "I was right, they were wrong." The lesson is: if an analyst cannot say "this data is not enough to conclude," he will always reach a wrong conclusion in some way. Maybe wrong by using data in the wrong place. Maybe wrong by using stale data. Maybe wrong by using non-existent data. And in all those cases, the report's form will make readers believe everything is fine.

There is one more aspect I want to address, concerning how empty data is exploited in markets. In sports betting markets, information asymmetry is an advantage. Those with better data can make better decisions. But when empty data exists, it creates a different kind of asymmetry: people who produce the appearance of data can influence crowds without having real data. A fabricated report that circulates can move odds. A groundless claim from a credible figure can shift money flows. This is a dark corner of the industry that I believe deserves more serious attention, because it affects not only those who place money, but also how information is produced and consumed across the entire ecosystem.

I do not write this to encourage anyone to enter those markets. I write to point out that when empty data is packaged in professional form, it can produce consequences far beyond the scope of one analysis. It can affect club decisions, players' careers, and fans' trust. The responsibility of an analyst is not only to reach correct judgements. It is also to ensure those judgements rest on a real foundation.

So what should we do?

I don't have a complete solution. But I have a simple principle I apply to every piece I write from 2026 to now: if I have no data, I will write that I have no data. If I have weak data, I will write that my data is weak. If I have strong data, I will write that my data is strong, but I will still state its limits clearly.

That is not an attractive principle. It does not produce viral pieces. It does not gain me followers. But it helps me sleep at night, because I know what I write is not fiction packaged as a report.

And in an industry where the form of analysis has overtaken its content, that is perhaps the most important thing an analyst can do.

The 37-page report that night in Busan went nowhere. I called the young analyst the next morning, and we spent three hours rebuilding from scratch — this time from real data, not from the form of data.

But I know not every report gets checked that way. And that is what keeps me writing. I started from a student blog with 2,000 views. Data doesn't care who you are, only whether you read it correctly. And in an industry where every report can become the basis for a transfer decision, reading data correctly is no longer a skill — it is a responsibility.