Silent Failure: When Football Analysis Has All the Framing But No Core

On August 13, 2026, in a small apartment on Aribau Street in Barcelona, I opened a 14-page report. The cover page had a logo, the words "Tactical and Technical Analysis," a comparative statistics table, bar charts, and even a professional glossary at the end. Not a single cell was left blank. But by page three, my spine went cold: the entire report contained not one factual information point.

I am 68 years old and have spent 52 years reading thousands of football analyses, from Belgrade to Barcelona. This was the first time I held a document polished enough that a skimming reader would believe it had value — when it had none. A report shaped like truth, but with no flesh of truth.

The summer of 2026 and the Opta ghost

In the summer of 2026, at 59, I left a traditional print newspaper to join a new online sports platform in Barcelona. My first data-driven analysis was Valencia's 3-0 win over Las Palmas on La Liga matchday two. Valencia scored three but registered only 1.4 xG. Las Palmas had an unusual PPDA of 7.2 — meaning they pressed aggressively, allowing only 7.2 passes before a defensive action. That pressing was right in spirit and wrong in structure: their back line pushed high, and Valencia only needed three straight passes to tear it apart.

Colleagues mocked me for "reading a stats sheet without watching the match." I stayed silent. But I spent three weeks building a homemade xG model to test it against the first 76 matches of the season. That was the summer I saw the Opta ghost — and from then on, my eyes stopped trusting what they saw.

That ghost never told me data is always right. It told me data can always be misread. And later, looking back at that 14-page report, I understood there is a different kind of error altogether: not misreading data, but presenting the correct form of an analysis while the content evaporated long before.

The architecture of silent failure

In the data engineering industry, this phenomenon is called "silent failure." A process runs to completion, raises no error, emits a correctly formatted output, and the end user receives a product that looks finished. No red light blinks. No warning appears. That is precisely what makes it more dangerous than a visible bug.

Modern sports analytics typically runs through multiple layers. The first layer ingests raw text — an article, a wire report, a match transcript — and decomposes it into "information points": atomic, verifiable factual claims such as "club X signed a four-year contract," "player Y is out six weeks with a hamstring injury," "team Z recorded 2.1 xG against W." The second layer takes those information points and performs tactical, financial, governance, and media analysis.

When the first layer fails — blocked by a paywall, unable to render JavaScript, or hit by an encoding error — it returns an empty array. Zero information points. The second layer still runs. And because the analysis framework is pre-built with nine dimensions, each with tables and headings, the second layer fills each cell with "Insufficient information to assess." A report is born. It has a title. It has structure. It has terminology. And it is empty.

Formal perfection is not evidence of substantive presence. That is the first lesson anyone reading football data must carve into their bones.

Information points — the ultimate unit of truth

Imagine an analysis of a derby. With no information points, we do not know the competition. We do not know the teams. We do not know the score. We do not know the lineups. We do not know the manager. All nine analytical dimensions — from tactics, finance, form, league position, governance, dressing room, risk, media, to industry transmission — require at least one anchor point. Without an anchor, every conclusion is invention.

I once watched a young editor fill the "Tactical Analysis" cell with the line "this team usually plays 4-3-3 with a high press." I asked him: which team. He went silent. He had just invented a tactical formation for a team that did not exist in the article. That error nearly went to print.

In football, a number from the wrong season, wrong match, or wrong league is an indelible stain. But a number that does not exist — presented as though it does — is worse than a stain. It is a structured lie.

On a Moscow night in the winter of 2026, I could not sleep. Not because of football, but because the numbers were whispering a prophecy. I published a piece predicting France would win the World Cup, despite their uninspiring group stage. I relied on two specific information points: France's U21 side had the highest rate of passes into the opponent's final third in Europe, and Antoine Griezmann's average shot carried an xG of 0.21 — above the average for top-tier strikers. The piece was dismissed as "dry as roof tiles." When France won, a Spanish editor told me: "You were right, but nobody reads the way you write."

That night I wrote in my notebook: truth must be told with emotion, not only with numbers. But I wrote a second line too: truth must never be allowed to be fake truth. A good article built on a false information point is a bad article in the costume of a good one.

When xG gets distorted

Over the past decade, xG — expected goals — has become the shared language of analysts. But I have seen xG misused in at least four ways.

The first: using xG to replace the scoreline. No. xG measures chance quality, not results. A team with 2.7 xG that loses 0-1 played well and met misfortune. A team with 0.6 xG that wins 2-0 played poorly and met luck. But leagues are decided by points, not by xG.

The second: comparing xG across leagues without recalibration. An xG model built on Premier League data cannot be transplanted wholesale to La Liga or V.League. Shot quality, assist quality, and defensive quality all differ. A beautiful number is like a perfect pass: it needs no explanation, only to be seen. But a beautiful number placed in the wrong context looks beautiful on paper and deceives us on the pitch.

The third: using xG while ignoring PPDA. A high PPDA means a team does not press — it waits. A low PPDA means a team presses aggressively — like Las Palmas that night against Valencia. These two extremes produce completely different xG outcomes. A high-pressing team often concedes higher xG if its back line cannot hold its spacing. A deep-defending team often concedes low xG but loses late goals through exhaustion.

The fourth — and the most dangerous — is using xG as an orphan number. Dropping it into an analysis without its birth date. No match, no date, no team. A number with no provenance will grow teeth and bite its author back.

I am 68, but the data is younger than I have ever seen — each season it grows another layer of teeth. And the newest layer is machine-learning models that automatically emit analytical reports. They can decompose thousands of data points an hour. But they can also confidently present an empty report just because the pipeline lacks a gate.

The empty gate and the three-source rule

When the stadiums fell silent in 2026, I suddenly understood: football never died — it only stripped off its costume and revealed its skeleton. That summer, the pandemic halted football, then restarted it in empty stadiums. I was granted real-time data access to a Segunda-level club in Catalonia. Home win rate fell from 46% to 38%. But passes into the final third rose 11%. A paradox only data could see: without crowd pressure, players circulated the ball more confidently, but lost the invisible spiritual weapon the stands had always provided.

I wrote a long essay about "lost space" and "digitally recorded psychological pressure." But before publishing, I cross-checked three independent data sources. Three days. Not because I am slow. Because I learned that a number not verified three times is a number not yet born.

From that point I proposed the concept of the "empty gate" — a mandatory check in any analytical pipeline. If input data contains no information points, the process must halt and return a clear status: "Source content insufficient." It must not emit a beautiful report. It must not let the reader believe a filled-in table is a table with substance.

The transfer market is a monastery where numbers chant; I merely transcribe what they pray. But if that monastery is built on an empty foundation, every chant is an echo of the void. During the transfer window, thousands of rumors pour into platforms daily. Only a fraction have identifiable provenance. A player is linked to three clubs in the same week. A deal is "done in 48 hours" for two months. If an analytical report accepts these rumors as information points, it will generate a false report. If it refuses them for lack of provenance, it will generate an empty but honest report. I always choose the second.

The trap of formal perfection

Looking back at that 14-page report, I realized its structure closely resembled the structure I still use in my deep analyses. It had an introduction, context, core analysis, contrarian angle, conclusion. It had a risk table, an assessment matrix, a comparative statistics table, even a glossary. But every cell read "Insufficient information to assess." A hundred times "Insufficient information." A hundred times.

What is frightening is not that it was empty. What is frightening is that it looked full. If someone does not read carefully, they will believe it. They will cite it. They will feed it into another pipeline. And the emptiness will multiply.

I have seen this in football. A club receives a scouting report from a data company. The report looks polished. The club signs the player. The player fails. On review, the report was built on match data from a different league, under a different measurement system. Nobody at the club checked, because the report was too beautiful to doubt.

A young player is valued at 20 million euros because of one season with 0.42 xG per 90. A beautiful number. But that number was measured in a league where the average defense allowed PPDA 13 — meaning opponents did not press. When the player moved to a league with an average PPDA of 8, he lost the ball twice as often. The report was not wrong arithmetically. It was wrong structurally. And structural error is always harder to detect than arithmetic error.

In modern football, the biggest risk is not a lack of data. The biggest risk is data presented in the right framework with the wrong core, and nobody patient enough to check the core.

The injury story is another example. Medical confidentiality blinds fans and media. Clubs only disclose injuries that benefit share price, contract negotiations, or brand image. A player with a "minor injury" may have a six-week muscle tear. A player "ready to play" may be performing at 70% capacity. When an analysis relies on club-released injury information without a second source, it is building on sand. And that sand can wash away overnight.

I once believed in feeling. After Opta, I believed in probability. After COVID, I believed in structure. When the stadiums emptied in 2026, I understood that football did not die — it only shed its costume. And looking at that empty report, I understood one more thing: some analyses also shed their costume, but when the costume falls, there is no skeleton inside.

What I want readers to carry away

Esports taught me one thing: human reflexes never beat the speed of an algorithm. But that sentence has a second half few remember: an algorithm never knows when it is wrong. Only humans do. And only humans can build a gate to stop the error from spreading.

For fans, this means something very concrete. When you read an analysis with tables, look for the information points. Do not look for the conclusion. Conclusions can be staged. Information points cannot. If a piece says "team X presses high" without telling you the PPDA in which match, ask a question. If a piece says "player Y has an impressive scoring rate" without telling you xG, minutes, and league, slow down. If a piece looks perfect but contains not one number with a birth date, close it.

For people in my profession, this means an iron rule: before writing anything, identify the central information point. If there is no information point, there is no article. I once intended to write about a player based only on a rumor. I stopped. Two weeks later, the rumor collapsed. I do not regret not writing. I only regret nearly writing.

Numbers do not lie. Only the people who read numbers lie. But there is a subtler lie than misreading a number: presenting a number as though it had been read, when in fact no number exists at all.

Opta saw this three months ago. Not the Opta of any specific company, but the Opta inside my head — the ghost I saw in the summer of 2026. It always sees first. And it always reminds me that a number presented in the correct format does not mean the number exists.

At 68, I still choose xG over intuition. But I also choose honest emptiness over fake completeness. A report that says "I do not know" is a trustworthy report. A report full of words that says nothing is a frightening report.

In the coming transfer window, as thousands of articles pour in, I will read each one a single way: find the information point first, find the conclusion second. If there is no information point, I will close it. Not because I am strict. Because I am 68, and I no longer have time to read beautiful reports that are empty.

When the stadiums fell silent in 2026, I understood football was only shedding its costume to reveal the skeleton. But I understood something more: some writers are also shedding their costumes. The problem is that not everyone who sheds a costume has a skeleton beneath.