World Cup 2026 and the Small-Sample Problem: When Data Isn't Enough to Judge

Minute 88, a missed penalty. Instantly, social media fills with explanations: weak mentality, drained fitness, the coach picked the wrong taker. I sit in Marseille, reopen the footage alongside my data sheet, and the first thing I write in my notebook is not a conclusion but a question: do I have enough data to say anything at all? The answer, in most knockout matches, is no. A missed penalty is an event, and an event has never been a trend. That is why I open a fresh spreadsheet — seven columns, not a single number — and leave it empty until I have gathered enough sample. World Cup 2026 is repeating an old lesson: a major tournament is where data is scarcest, yet where people rush to conclusions fastest. A club season gives us 38 matches. A World Cup gives us seven at most. National teams spend less time training together than any club, so every match carries an enormous amount of noise. In statistics, the smaller the sample, the larger the variance, and a player can score three goals in two games then vanish for three more without his true ability changing at all. My job, for more than forty years, has been to separate signal from noise. I began my career in 2026, the year The Independent was founded, and I learned that writing discipline starts with observation, not inspiration. In my notebook at every major tournament, each national team has its own page, divided into four columns: process metrics including xG, xGA and PPDA; match context including opponent, pitch and weather; human factors including injuries and cards; and a blank space kept in reserve. That blank space is the most important one. It reminds me that every conclusion is provisional, and that I am always short of data until the tournament ends. In the summer of 2026, I learned to trust something no one had yet named: xG. When Opta first released xG tables for Ligue 1, I did not rush to believe. I hand-recorded 1,204 shots from 20 teams in the first half of the 2026-18 season and compared them with actual goals. The correlation coefficient reached 0.84 — enough to build my own striker valuation dataset, and enough to understand that a metric is only trustworthy when you know its sample size. Colleagues said my reaction was slow. I need verification before use, and I never cite a new metric without stating how many events it was calculated from, across how many matches, and in what context. That is the principle I carry into every major tournament. At World Cup 2026, it let me do something many considered reckless. I tracked all 64 matches and counted PPDA — the passes an opponent is allowed before each defensive action. In the semi-final Croatia – England, Croatia allowed England only 8.2 passes before each defensive action, while England allowed Croatia 12.5. I wrote a note predicting Croatia would win through extra-time pressing. They won 2-1, with Luka Modrić dictating midfield. I did not shout in celebration; I reopened the spreadsheet to hunt for outliers. Croatia champions of a low-PPDA tournament? Then PPDA is merely a letter, not yet the truth. From then on, I shifted from describing matches emotionally to presenting PPDA tables, running distances and successful pressing counts. The phrase "the team pressed better" only appears when the numbers genuinely support it. In 2026, football returned to empty stands, and I was assigned to cover the Bundesliga. I sat in Marseille analysing 81 matches in empty stadiums in the 2026-20 season. Home teams won only 26%, against 43% before the pandemic. I wrote the report "Empty stands kill home advantage", and a Ligue 2 club, Le Havre, used it to negotiate down the price of a young striker who had shone at home. Empty stands are the finest laboratory for a data lover, because they remove a variable we normally cannot isolate. But the deeper lesson lies elsewhere: every table of mine since then separates home and away metrics, and I remind readers not to trust pre-lockdown form when judging a player. World Cup 2026 took the principle to another level. When pundits praised Achraf Hakimi for 142 sprints and 2.3 chances created per match, I dug into the data and found the channel behind him was open 34% of the time. Morocco stayed safe because their centre-backs ran above 31 km/h. I wrote a note warning that this trend of the high attacking full-back only holds if the defence is fast enough to compensate. Against France, the opponent poured attacks down Morocco's right flank, where Kylian Mbappé kept exploiting that space. I no longer praise a new tactic without weighing its compensating variables. In every analysis, I list the necessary and sufficient conditions for a system to work, rather than applauding a trend. In the group stage, teams usually face weaker opponents, so attacking metrics get inflated. A side can reach 2.4 xG per match against lower-ranked teams, then drop to 0.9 against a well-organised defence. If I read only the 2.4 without splitting the sample by opponent quality, I will predict wrongly in the knockout round. That is why I always divide my data into two parts: form against teams of the same level and form against stronger teams. The second part is always smaller, and always more trustworthy. This is also why I am always sceptical of deals signed right after a major tournament. A young player who shines in three group-stage matches can be valued twice as high, while dressing-room chemistry — something that appears in no xG table — is undervalued. The industry's transfer data models tend to overrate young potential and underrate integration. I have watched deals built on a seven-match tournament collapse within a 38-match season. Small samples do not only fool viewers; they fool those who sign the cheques. I see the same thing in esports. Professionalisation is turning players into the product of an assembly line, where individual flair is smoothed away in digital training. But esports tournaments also provide only small samples, and people are just as quick to crown a champion after a handful of games. A single mouse click on an esports screen carries the shape of a pass: it only means something inside a system. This is where I want to be honest with myself. There is a temptation greater than trusting a wrong metric: filling a data gap with a story. When I open a spreadsheet and it is empty, the storytelling instinct wants me to write a conclusion immediately, because an empty analysis makes us uneasy. But correlation is not causation, and emptiness is not permission. I am 66, old enough to know a number never tells a story unless we ask. Before concluding anything about a national team at a major tournament, I force myself to write at least three hypotheses explaining the same phenomenon: one tactical, one physical, one random. If all three are plausible, I have no right to pick the one that sounds most appealing. For example, when a team concedes three goals in a knockout match, the tactical hypothesis might be a high defensive line, the physical one might be lingering extra-time fatigue, and the random one might be three long-range shots into the corner. Only by reviewing each goal do I know which belongs to which cause. There are matches won on the pitch but lost on the data sheet — I choose the data sheet. But there are also empty data sheets, and then the only honest choice is silence, a note reading "insufficient information", and waiting for more sample. What I want to say about the current stage of World Cup 2026 is this: do not let one match, one shot, or one missed penalty shape your entire story about a team. A player is a variable, the market is a function, but most of my life has been a constant — and that constant is patience. The signal I will track in the next round is not the scoreline but the structure. I will watch whether a team keeps creating the same number of chances against stronger opponents, whether their defence is fast enough to cover an advanced full-back, and whether their home metrics are clearly separated. Data will thicken round by round, and the right story will emerge on its own — not from a moment, but from a sample large enough to be trusted. A cancelled match is not lost points but a lost page of the diary; so is a rushed conclusion.