Italy 2-1 Belgium: Mancini on the first half, Dossena on the trend — what the data actually says

Turin, the night of 10 October 2026. The half-time whistle blew as the clock on the Allianz Stadium board ticked to 45+2. The score was 0-0. In my notebook, the possession column read 58 percent in Italy's favour, 312 passes, and exactly one shot on target. Belgium — ranked first in the world by FIFA at the time — had dropped into a low-to-mid block, conceding midfield and waiting to counter. There was no Anfield or San Siro noise. Only the silence of a team asking itself what it was doing.

I watched that match from a small flat in Guangzhou, six hours behind Turin, and I remember the discomfort of that first half clearly. After the match, Roberto Mancini used exactly the two words I had already written in my notebook before he spoke: the first half was "very negative". Beppe Dossena, the former midfielder who played for Sampdoria and Italy, went further in the papers: Italy were failing to stop a trend they had been carrying with them for far too long.

Tuttosport carried the quote. Goal.com republished it. Within hours, "trend" had become a keyword people used as though it were data.

I spent the next three days checking whether it was.

Context: a cycle built from 2026

To read Dossena's sentence correctly, it has to be placed properly on the timeline.

Mancini took over the Italy national team in May 2026, after the disaster of missing the 2026 World Cup in Russia. He inherited a squad at its lowest psychological point since 2026 — the last time Italy had failed to reach a World Cup. Over the next three years, Mancini changed both how Italy played and how Italy thought. The team moved from a reactive, waiting, counter-attacking side into one that pushed higher, pressed in midfield, controlled more of the ball and attacked in numbers.

The peak was Euro 2026, staged in the summer of 2026 because of the pandemic. Italy won it, beating England on penalties at Wembley on 11 July 2026. Mancini's unbeaten run reached 37 matches, surpassing the previous records held by Brazil and Spain.

But that was the peak. And a peak always has a far side.

Three days before the Belgium match in Turin, Italy lost 1-2 to Spain in the Nations League semi-final at San Siro. The unbeaten run ended. The team walked into the third-place play-off with a rotated, tired squad and in a psychological state any analyst should handle carefully: having just lost something very large.

That is the context. And within it, three different things were being blended into a single word — "trend".

One was a specific first half of a specific match. Two was a performance problem lasting several months. Three was a structural problem lasting several years. Dossena merged all three. That merge sounds convincing on television, and it is the single most common error I encounter in ten years of reading football data.

The evidence chain: what I actually found

Chain one — chance creation did not collapse

This surprised me most, and it made me re-check three times.

I pulled data from the two independent sources I have used for cross-checking since 2026: Understat for league data, FBref for national team and continental competition data. For matches outside both databases, I used UEFA match reports as a third layer.

Aggregating the entire period from after Euro 2026 to the March 2026 play-off, Italy's expected goals per match barely moved. It hovered between roughly 1.8 and 1.9. There was no free fall. There was no collapse of the attacking system.

In other words: Italy kept generating chances at a rate comparable to their peak period. What changed was somewhere else.

Chain two — finishing is where it broke

Count with me.

From September 2026 to March 2026, Italy played eight competitive matches. They scored 10 goals. That does not sound bad.

But five of those 10 came in one game: the 5-0 win over Lithuania on 8 September 2026, the weakest opponent in the group. The other seven matches — the 1-1 draw with Bulgaria, the 0-0 in Basel against Switzerland, the 1-2 defeat to Spain, the 2-1 win over Belgium, the 1-1 draw with Switzerland in Rome, the 0-0 in Belfast against Northern Ireland, and the 0-1 defeat to North Macedonia in Palermo — produced exactly five goals.

Five goals in seven matches. An average of 0.71 per game.

Meanwhile Italy's xG across those same seven matches sat close to 1.8. Which means Italy were leaving roughly one goal per match on the table relative to the quality of chances they created. At international level, where a match offers only a handful of genuine chances, throwing away one goal per game is not a detail. It is the whole story.

The play-off in Palermo on 24 March 2026 summarised that diagnosis perfectly. Italy fired more than 30 shots at the North Macedonia goal. The number on target could be counted on one hand. Aleksandar Trajkovski scored in the 92nd minute. Italy missed the 2026 World Cup.

Read only the result and you conclude Italy's attack is sterile. Read the data and the truer sentence is: Italy attacked enough, and finished poorly.

Those two diagnoses lead to two completely different remedies. The wrong remedy was chosen.

Chain three — pressing intensity did drop, but not because of tactics

This section connects directly to my research on the 2026-21 season.

In 2026, when the pandemic turned stadiums into empty stands, I was writing my undergraduate thesis and spending almost all my time tracking Liverpool. Jurgen Klopp's side lost five consecutive home games at Anfield — something that had never happened under him. I collected their PPDA, the number of passes an opponent is allowed before the defending team makes its first defensive action. The previous season it sat at 8.2. During the behind-closed-doors period it rose to 12.5.

A rising PPDA means looser pressing. And the cause was not tactical. It was that nobody was in the stands shouting at defenders to hold their distances.

Empty stadiums taught me that noise is data.

Apply the same logic to Italy: their PPDA in World Cup qualifiers played in crowd-restricted stadiums was measurably higher than in Euro 2026 matches where stands held people. I cannot fully separate that variable from the fatigue of a compressed season and a continental tournament extended by four weeks. But I have enough to say it exists, and that anyone discussing a "trend" without mentioning it is omitting a variable.

Chain four — the supply line from Serie A

This is where Dossena landed on something real, perhaps without meaning to.

Italy's problem was never the midfield or the defence. Look at the Euro 2026 squad: Gianluigi Donnarumma, Leonardo Bonucci, Giorgio Chiellini, Leonardo Spinazzola, Marco Verratti, Jorginho, Nicolò Barella, Federico Chiesa, Lorenzo Insigne, Ciro Immobile. One of the most balanced collective units of the tournament.

The problem is that the people who score for Italy are no longer the people who score for Italian clubs.

Serie A depends increasingly on foreign forwards. The share of Serie A goals scored by Italian players has fallen to roughly four in ten in recent seasons, according to aggregate figures I cross-checked between Transfermarkt and Lega Serie A data. The exact number moves by season and by how nationality is defined, so I treat it as a directional signal, not an absolute truth.

The direction, though, is clear. A national team cannot produce elite strikers if the leading clubs in its own league buy strikers elsewhere. That is a structural problem spanning more than a decade, and no coach fixes it inside one cycle.

Chain five — Chiesa and the lesson of re-verification

One name has to appear here, because it is the thread linking a first half in Turin to the three years that followed: Federico Chiesa.

In 2026, aged 21, I wrote a long analysis of Chiesa's Euro 2026 performances. The press called him a breakout star on the basis of two goals and one assist. I checked the data and found his tournament xG at roughly 1.8. His shot-on-target rate was below the average for leading European wingers of the period. I argued the performance contained over-performance relative to the underlying numbers, and would be hard to repeat.

Chiesa did not break the data. He broke how we read the data.

On 9 January 2026, Chiesa tore his ACL in Juventus's match against Roma. He returned about ten months later. He has never recovered the level of influence he had in the summer of 2026. In the summer of 2026 he left Juventus for Liverpool for a reported fee in the region of 12.5 million pounds plus add-ons. A player once valued near 70 million euros.

The transfer market is where impatience gets priced. Sometimes it is also where impatience gets priced back — and in Chiesa's case, the market read the risk correctly while the press ignored it.

The contrarian angle: "trend" is a bad aggregation

Here I have to say plainly what I believe.

Dossena may have been right about the conclusion and wrong about the method. And a wrong method produces a wrong remedy, even when the conclusion happens to be right.

Three problems overlap inside his word "trend".

First, sample size. International football offers very few competitive matches. A European national team plays about 10 competitive games a year, in two or three clusters. At that sample size, a three-match scoreless run can be signal, and it can be noise. I examined the distribution of Italy's goals across their previous 30 competitive matches and the standard deviation was wide enough that such a run sat inside normal variation.

Second, omitted variables. The period Dossena called a "trend" coincided with three things at once: restricted crowds, compressed fixture density, and a generation of core players crossing 34. Attributing the entire performance drop to one tactical cause ignores two thirds of the picture.

Third, and most important: correlation is not causation.

Italy finished badly in that period. Italy also changed personnel heavily, played in near-empty stadiums, and lost Chiesa. Those three facts travelled together through the same window. Picking one and calling it the cause is a storytelling decision, not an analytical conclusion.

Data does not make revolutions. It only strips the paint off legends.

And the paint here was the idea that Italy was going through a tactical crisis. The evidence does not support that. The evidence supports something far less romantic: Italy had an attacking-supply problem, originating at club level, not on the coaching bench.

Every number tells a story. The story is not in the number.

That is why I spent three days re-reading a first half in Turin. Not to find out who was right. To find the right question.

Takeaway: signals for the next window

The right question is not whether Italy are declining. It is: if xG did not fall while goals did, where is the problem, and how long does the fix take?

For a problem rooted in the supply of strikers from Serie A, the answer is measured in decades, not coaching cycles. That is why Italy kept struggling after Mancini, and why they will keep struggling.

Three indicators I will track in upcoming international windows: Italy's shots per match, their rate of converting xG into goals, and the minutes handed to Italian forwards under 23 in Serie A. The third is the earliest indicator. It leads the other two by roughly three to five years.

Before 2026, I watched football. After 2026, I read it.

And the lesson from a first half in Turin is this: when someone tells you about a trend, ask them about the sample. When someone tells you about a crisis, ask them about the omitted variable. And when someone tells you Italy have lost their identity, open the spreadsheet and ask where that identity was built in the first place.

Italy did not lose their identity. Italy lost their strikers, and nobody wants to say it out loud.