An Empty Sheet Is Not Neutral: A Verification Lesson from Croatia 2026

In the final forty minutes of the 2026 World Cup semi-final between Croatia and England at Luzhniki Stadium, I did not write another line. My two-thousand-word analysis had been finished before kick-off, and it argued that Croatia would collapse physically: eight starters over thirty, three consecutive matches that had gone to extra time. Croatia won 2-1 with 58 percent possession, Ivan Perišić equalising in the 68th minute and Mario Mandžukić sealing the score in the 109th, while Luka Modrić set the tempo in midfield. For a month afterwards I sat with the full match tapes, logging 214 set-piece situations, and found what I should have seen earlier: Croatia were not exhausted, they chose to play slowly. Slowing the tempo in extra time was a deliberate decision, repeated across all three knockout rounds, and it sat in exactly the data I had not bothered to collect.

This morning, at a desk in Chengdu, I received a spreadsheet. The column headers were complete: possession, expected goals, distance covered, successful duels. The body of the sheet was entirely empty. The sender, a young editor, asked whether I could build a tactical analysis from it.

Thirty-eight years in the trade have taught me that empty sheets are not rare, but each era produces them differently. In 2026 I wrote for a print newspaper in Chengdu. A young blogger reported that Sichuan Longfor had gone bankrupt, attaching a photograph of an empty training ground. I had attended twenty-three live training sessions and knew the club had just signed a sponsorship worth 2.8 million yuan. My rebuttal ran three days later and drew 1,200 reads. The false story had already travelled everywhere.

In 2026, when the pandemic stopped every competition, I dug out the tape archive of the China national team from 2026 to 2026. With no new matches to watch, I measured the old ones again and found a pattern: when first-half possession rose more than 12 percent above that team's own average, the second-half defeat rate was 71 percent, seven out of ten. A series called The Numbers That Weep drew 60,000 reads, and Chengdu Rongcheng approached me to work as an informal consultant.

Those three chapters left me with one habit: when an empty dataset arrives, the first thing I do is look for the cause of the gap, not for a conclusion.

An empty sheet has three kinds of cause, and each demands different handling. The first: empty because nobody watched. This is standard in smaller leagues, where the analysis budget only stretches to a few showpiece fixtures each round. The second: empty because the observer refused to observe. The conclusion was already settled, so only what fitted it got written down. The third: empty because the event left no measurable trace — a decision to slow the tempo, for instance, produces no shot to count.

The Croatia case belongs to the second kind, and I was the one who created the gap. I had the tapes, the notebook, three weeks with the squad. But I recorded against a single hypothesis, fitness, so every line circled heart rate and distance. A team deliberately slowing down to conserve energy only appears when you bother to count sideways passes in extra time, count the seconds held before each restart, count how often defenders rolled the ball back to the goalkeeper. I did not count. I believed the arithmetic before I believed the pitch; that was my most expensive mistake.

In V.League, data providers have begun publishing detailed per-round indices, but the infrastructure has a characteristic: it concentrates on the end of the action. The shot, the final pass, the duel. What happens before — the rhythm of the preparation, the delay in the decision, the silence between two phases — is barely encoded. That is why model-driven analyses so often look suspiciously alike.

The return of the back three in recent seasons is a clear example. It gets called tactical evolution. Read the defensive data along a timeline and a different story appears: most teams switched to a back three after a run of being carved open through the middle. Adding a centre-back does not create more chances. It reduces goals conceded, and reduces the number of incidents a head coach has to explain in the press room. The back three is a reputational insurance policy. To test it, the only way is to count the expected goals a team generates from organised counter-attacks across two consecutive seasons. If the back three were genuine progress, that number should rise. In most cases I have tracked, it falls.

The refereeing story follows the same arc. When VAR arrived in V.League, I supported it. After a while, I began to doubt myself. Millimetre offside lines rest on frames sampled at an interval, and the gap between two frames, at a sprinting striker's speed, covers a distance that is anything but small. The assistant referee is no longer the final decision-maker; he has become the editor of a cut: choose a frame, award a goal, disallow a goal. Frame error is only part of it. The rest is what cannot be measured: attackers stopping their runs because they trust the line more than the instinct.

Transfers close the picture. The race between the big clubs in V.League and across Southeast Asia is largely a branding race: buying to prove you have money, selling to balance the books. The deals that actually change a league's outcomes usually sit at low-budget clubs, where scouts have to sit through footage of lower divisions. Transfers are where value gets replaced by price, and nobody says it out loud. When a small club signs a player no newspaper covers, that is usually a more interesting signal than the marquee deal of the week.

Vietnam won the 2026 ASEAN Cup, and throughout that tournament I noticed something familiar: after every match, people argued about the scoreline far more than about how the scoreline was made. The 71 percent rule I found in the 2026-2026 archive needs a warning attached. I have recalculated it at least four times and each time found the same problem: rising first-half possession does not cause second-half defeat. It merely accompanies it. A team holding the ball in the first half is usually trailing and pushing its line higher, or leading and deliberately slowing down. Both directions end in losing control after the interval. That rule is a question-generating tool, not a prophecy. I still use it, but always with one extra column: the conditions under which I can be wrong.

Outsiders look at an empty sheet and assume it is neutral. They believe that with no data, nobody is biased. The reality inverts: a gap is the strongest state of bias, because it will be filled by whatever is at hand — narrative, emotion, headline. An empty sheet does not stay silent. It invites the loudest voice to fill it.

The same result, and the insider and the outsider are watching two different matches. The outsider sees the gap and writes. The insider sees the gap and asks who missed it, at which stage, and why at this particular moment.

False stories do not die because people stop believing them; they die because people stop verifying them. In Chengdu in 2026, the false report about Sichuan Longfor lived for three days and reached tens of thousands. My verification reached 1,200 reads. I used to treat that as a failure of the profession. Now I read it differently: those 1,200 are the people who will verify for themselves next time, and they verify faster than the rest.

There is one thing I still cannot explain, and I am leaving it where it is. People look for the goal; I look for the forgotten pass. Between two rolling balls there is a silence that holds the whole truth.

Tonight I will answer the young editor with a different spreadsheet. The first column keeps the original headers. The second column I add myself: conditions under which this can be wrong. The first three rows will state who observed, with what instrument, and for how long. If those three rows are empty, I will write no analysis at all, and say plainly why. The next signal I am watching is not in the scoreline of the coming round. It is in who steps into the gap first, and whether they step in with data or with a story. Old files gather dust, but the long shot from that year still leaves its curve in the data — and today's gap will leave its own trace, once somebody is willing to bend down and look.