The Statistic Convicting Ancelotti Has a Hole Brazil Refuses to See

The second half drifted into its final minutes. Carlo Ancelotti stood motionless at the edge of the technical area, hands in his pockets, barely offering a gesture of instruction. The scoreboard read 1-1, a result his team had salvaged through an individual moment from Rayan Vitor rather than through any system. In Incheon, I sat in front of a screen and knew exactly what would follow: by morning, all of Brazil would be talking about a statistics table.

They talked. Goal.com, citing DataESPN data, published a piece calling this the worst start in the history of the Brazil national team, and placed Ancelotti at the centre of a wave of criticism. The story spread the way these stories always spread: a flat friendly, an eye-catching percentage, and a foreign coach standing in the middle of the storm.

I read that dataset several times and recalculated every line on paper. Not to defend Ancelotti. Because I have stood on the other side of an identical argument, and I remember what it feels like to write too fast and discover that you were right about the problem but wrong about how you framed it.

On September 25, 2026, Brazil drew 1-1 with Australia on an overseas tour. It was Ancelotti's eighteenth match in charge of the national team. The record: 10 wins, 4 draws, 4 defeats. Converted into points the way Brazilian football counts them, that is 34 out of a possible 54, an efficiency rate of 62.96%, rounded by the media to 63%.

The comparison table Goal.com reproduced contains four names. Carlos Alberto Parreira sits beside Ancelotti with 9 wins, 7 draws and 2 defeats. Vanderlei Luxemburgo leads with 15 wins and 47 points, or 87.04%. Tite follows with 14 wins and 45 points, or 83.33%. Put together, these numbers tell a very tidy story: the predecessors started far better, and Ancelotti has hit a historic floor.

The schedule ahead makes the story louder still. Brazil face Australia again on September 29, 2026, then India on October 3, 2026. Brazilian media have labelled these two of the most important matches of Ancelotti's tenure, and that framing is itself an analytical error.

I start with the simplest calculation. Ten wins is thirty points. Four draws is four points. Thirty-four points from a maximum of fifty-four across eighteen matches. The 62.96% figure DataESPN published matches exactly. The dataset is internally coherent, and that deserves credit: their raw numbers are not fabricated.

But when I calculated Parreira's line, everything changed colour. Nine wins is twenty-seven points. Seven draws is seven. That totals thirty-four. The identical points total, the identical efficiency rate, down to the decimal.

The record Ancelotti matched is not the worst record. It is the joint-worst record.

The Goal.com headline says "worst in history." The data underneath says "joint-worst." Those two sentences do not mean the same thing, and the gap between them is precisely where a media narrative gets built.

There is more. The table reproduced by Goal.com skips a detail: Parreira's record was compiled across two separate stints, in 2026 and 2026, added together to reach eighteen matches. We are comparing one continuous eighteen-match tenure with two disconnected periods almost a decade apart, in two entirely different footballing contexts. This is no longer a loose comparison. It is a wrong one.

Then comes the detail I consider the most important in this entire story: Carlos Alberto Parreira, the man who shares the worst-start record with Ancelotti, went on to lead Brazil to the 2026 World Cup title in the United States.

Goal.com names Parreira as the benchmark to indict Ancelotti but never mentions how Parreira's story ended. That silence turns a neutral statistic into a one-directional charge sheet. And it collapses the article's central argument from within: if a 63% start across eighteen matches can lead to a World Cup, the metric predicts nothing at all.

The reverse direction is more interesting still. Vanderlei Luxemburgo holds the best start in national-team history at 87.04%, won the 2026 Copa America, and never won a World Cup. Tite reached 83.33%, more than twenty percentage points above Ancelotti, and exited two consecutive World Cups at the quarter-final stage, in 2026 and 2026.

The metric being used to convict Ancelotti has never demonstrated any relationship to the thing Brazil actually cares about.

This is the kind of mistake I have made myself, and I recognise it faster each time. I was wrong at the Russia World Cup, and it was the best thing that ever happened to me. In 2026, aged twenty-three, I wrote that South Korea could not beat Iran if they kept playing the same way, was roundly abused for it, and watched the match end 0-0 with 61% possession and exactly two shots on target. I was right about the problem. But I learned something bigger from my failure at Qatar 2026, when I predicted South Korea would beat Portugal 2-0 through possession football, and they instead won 2-1 through a counter-attack in the 91st minute. It took forty minutes of rewatching footage before I understood where I had gone wrong, and that error taught me more than being right ever did.

With Ancelotti, I read the dataset with that same habit, and I found four holes.

The first is survivorship bias. The ranking only counts managers who survived eighteen matches. Those who started catastrophically and were dismissed by their eighth or twelfth game never appear in the table. So the accurate sentence would be: the worst start among those granted a full eighteen matches. That phrasing is far humbler than the headline.

The second is cross-era comparison. Parreira's 2026 side, Luxemburgo's 1990s side and Tite's 2010s side played different calendars, faced different opponent pools, and lived in different national-team cycles. In the 1980s, Brazil played a great many non-competitive matches, and those inflate win percentages. Comparing efficiency rates across eras is a weak instrument, useful only as an illustration.

The third is the composition of the eighteen matches. Goal.com does not disclose how many were qualifiers, how many were World Cup fixtures, and how many were friendlies. If Ancelotti's eighteen are friendly-heavy, 63% is worse than it looks. If they are qualifier- and World Cup-heavy, the figure is much better than the way it has been read. This single missing data point is enough to reverse the entire conclusion.

The fourth, and in my view the most serious, is that the article contains not one line of tactical analysis. No xG, no passing data, no pressing metrics, no formation, no description of how Brazil build possession. The whole piece operates on outcome data: wins, draws, losses. That is accounting, not football analysis.

Data whispers while the whole stadium is screaming. I learned to listen. And the only thing worth listening to in this dataset is not the four draws. It is the four defeats. Brazil losing four times in eighteen matches is unusual for a national team that once taught the world how to defend. Those four defeats are the structural signal. The four draws are noise.

Based on my experience tracking matches, I record everything recordable: touches, pass-completion rates, distance covered, successful pressures. I have done that since 2026, after being abused for saying South Korea could not beat Iran while keeping the old approach. That habit taught me one thing: a table containing only wins, draws and losses tells me nothing about how a team plays. It only tells me the outcome of how they played.

So what is actually driving the pressure on Ancelotti? Not the draw in Australia. That friendly carried almost zero sporting information value while carrying enormous narrative value, a dangerous asymmetry. What feeds the criticism is the memory of the 2026 World Cup knockout stage, where Brazil were eliminated in the round of sixteen. For a national team, public patience collapses at the moment of tournament failure, and everything afterwards is re-read as confirmation. The Australia draw is being processed as evidence of continuation, not as an isolated event.

This is a mechanism I have watched too many times in my career to still be surprised by. A friendly cannot be "vital" in any sporting sense. It is vital only in the political and media sense, for the man in the coaching seat. Brazilian media elevating the Australia and India fixtures into turning points is narrative inflation, and that inflation raises the probability of an artificial crisis.

One more factor goes largely unmentioned. Ancelotti is the first foreign coach to lead Brazil in decades. When a federation breaks with tradition, every failure is interpreted at the level of the model, not merely the individual. People do not say "Ancelotti lost." They say "it turns out hiring a foreign coach doesn't work." A domestic coach is protected by networks, local media relationships and national identity. Ancelotti has none of that armour. The pressure on him is amplified by the very decision to hire him.

Then there is the commercial dimension nobody wants to name. The Australia and India tour is designed to generate revenue. It opens two markets, creates two broadcastable fixtures, and delivers commercial and sponsorship income to the Brazilian federation. Financially, it makes sense. Sportingly, it is questionable: long-haul travel, time-zone shifts, disrupted training blocks, unfamiliar pitches. The same tour that makes money for the federation strips the coach of serious preparation. That is a conflict of interest sitting inside the structure of the game.

An empty stadium is the most honest mirror football has ever had. But I remind myself it is only one of many mirrors. Noise, money and the fanaticism of the stands are also facts, and sometimes they matter more than any dataset.

Where could I be wrong?

There is one undisclosed variable that determines almost the entire story: the terms of Ancelotti's contract. Its length, his salary, the compensation payable on early termination. If CBF must pay an enormous sum to dismiss him, then "under fire" is largely media theatre. If the contract contains a low-cost release clause, the threat is real. Without that data, any estimate of Ancelotti's sacking risk is invention dressed in statistics.

I also have to concede the possibility that the four defeats are structural rather than variance. Four losses in eighteen matches is anomalous for Brazil in any era, and it is the only data point in the article that makes me pause. Beyond that, the tactical quality of Brazil under Ancelotti is something I cannot assess, because there is nothing to hold on to. It is possible something is wrong inside the dressing room that the press has not touched. But what stands out is that across everything Goal.com aggregated, there is not a single player quote, not a word from federation leadership, not one internal voice. If the dressing room were turning, this kind of aggregation would normally say so. That absence is a signal, even a weak one.

And I warn myself about my old habit: forcing shock to protect the hot-take guy label. The people who hate me read every line I write more carefully than the people who love me, and I know that creates a very specific temptation. So I ask myself: if nobody argues about this piece anymore, is it still worth reading? The answer lies in the fact that the argument does not depend on who is right about Ancelotti. It depends on whether a results table is qualified to convict a coach at all.

My prediction for the September 29 and October 3, 2026 window is verifiable.

If Brazil beat Australia convincingly, the crisis story dies within two weeks, because it lacks the data to live longer. If Brazil draw flatly or lose either match, and especially if they lose to India, the pressure moves from the media level to the federation level, and the real question then appears: what is the cost of dismissal? And if Brazil beat India heavily, I will be the first to say that win proves nothing, because a friendly in a developing market is the most efficient generator of false positives in football.

Brazil do not lack money, talent or players. What they lack right now is a yardstick that measures the right thing. After the 2026 round of sixteen, the road to 2030 began, and the first friendly of that road has been turned into a trial. What worries me is not who that trial convicts. It is how many people believe a results table is qualified to sit in the judge's chair.