The Back Three Is Back, but the Transfer Money Is Flowing to the Flanks

On 24 September 2026, Arsenal beat Chelsea 3-0 at the Emirates. Antonio Conte, who had just left his post as Italy head coach after Euro 2026, sat in front of the cameras and talked about "balance between the lines". Ten days later he scrapped the 4-2-3-1 and rebuilt his defence with three centre-backs. Chelsea won 13 consecutive Premier League matches and finished the 2026-17 season with 93 points and the title.

I remember that detail not because of Conte. I remember it because across four decades spent between the stands and the data room, I have watched the back three return at least five times, and every single time it was sold to the public with the word "revolution". Every time I went back to the tape. Every time the real motive sat somewhere completely different from where the media pointed.

The back three is not new. Serie A lived on it in the 1990s. Arrigo Sacchi broke it at Milan with a pressing back four, and a whole generation of European coaching ran after him. At Euro 2026, Conte brought the 3-5-2 to France and turned it into Italy's identity. By Euro 2026, nearly half the teams reaching the knockout rounds started with three centre-backs. In 2026, Thomas Tuchel switched Chelsea to a 3-4-2-1 mid-season and won the Champions League. The same year, Conte took Inter Milan to the Scudetto with a 3-5-2.

Each time it comes back, it gets a new label. In 2026 it was "flexible reaction". In 2026 it was "total football in a pandemic". In 2026 it was "control without the ball". Right now, in the middle of the transfer window, it is called "the modern defensive structure" — and it is being sold alongside the most expensive centre-back contracts in history.

I do not believe the naming. I do believe the data. And the data tells a different story.

In 2026-24, Xabi Alonso's Bayer Leverkusen went unbeaten in the Bundesliga: 28 wins, 6 draws, 90 points, and the first league title in the club's history. Their unbeaten run across all competitions lasted 51 matches, ending on 22 May 2026 with a 0-3 defeat to Atalanta in the Europa League final, a match in which Ademola Lookman scored a hat-trick.

Everyone talks about Alonso's back three. Very few people talk about how much the two players who decided that system cost.

Alejandro Grimaldo joined Leverkusen on a free transfer in the summer of 2026, after his contract with Benfica expired. Jeremie Frimpong arrived from Celtic in January 2026 for a reported fee of around 11 million euros. Those are Leverkusen's two flanks — the players who touched the ball most in the opposition half, the two who decided whether the back three became a back four or a back five in possession.

The core point sits here: a back three does not create an advantage, it only creates a gap — and the value of the whole system depends on who fills that gap.

With the ball, three centre-backs create a spare man in build-up. Opponents usually press with two forwards, so one centre-back is always free. If that free centre-back can play through the lines, the team bypasses the first pressing line without needing a midfielder to drop deep. If he cannot, the team can only pass among the three of them and then go long.

Without the ball, three centre-backs create a thicker defensive line. The price is losing a midfielder in the middle. If the opponent plays long into central areas, a back-three team often loses the second layer.

In other words, a back three is a form of insurance contract. It shifts risk away from the coach and onto the structure. When a back four gets cut open, the first question in the media is "why didn't he change it?". When a back three gets cut open, the question is "where did the players go wrong?". That is an investment in reputation, not in points.

And the transfer market has misread the signal.

Clubs are paying for centre-backs. Joško Gvardiol joined Manchester City in August 2026 for a reported fee of around 77 million pounds, and he spends most of his time playing as a left-back. That is money spent on a hybrid player, but it is still booked under "centre-back". Meanwhile Grimaldo — the man who generates most of the system's value — arrived for nothing.

The money flows to the wrong place. It pours into the position the valuation tables find easiest to measure: young centre-backs, good physical profile, safe passing. A wide player who can produce goals from a defensive position is still priced as a defender. And a free transfer is never free either: the fee simply moves from the transfer column to the signing-on fee and wage column. Leverkusen paid Benfica nothing, but they paid Grimaldo — and they paid him very well.

That is why I think most of the teams currently buying a back three will be disappointed. They are buying the frame. They are not buying the engine.

Based on my experience watching matches, the earliest tell is not the starting shape published before kick-off. It comes in the fifteenth minute, when the team has the ball in its own half and you count how many times the wing-backs touch the ball on the highest line of the touchline. If that number is low, the back three is just a back five written differently.

I have also been wrong often enough to know I need to argue against myself.

The Leverkusen sample is a single club. One unbeaten season is beautiful data, but it is still one sample. If I use it to draw conclusions about the whole of Europe, I am doing exactly what I just criticised: taking an outlier and calling it a law.

Expected-goals-against numbers for back-three teams are also usually distorted by game state. Teams that are leading tend to switch to a back three to protect the result. Which means the best defensive teams contribute most of the minutes to the "back three" statistical sample. I am not sure I have separated that variable, and anyone who claims certainty should show their method.

Then there is Morocco. Morocco reached the semi-finals of the 2026 World Cup while playing without a back three for most of the tournament. They got there with a disciplined back four, with Bono saving two penalties in the shootout against Spain, and with Youssef En-Nesyri's header in the 42nd minute against Portugal. Morocco had no miracle. They had homework, and they did it very carefully.

If the lesson of Morocco is "any shape works as long as you train it properly", then my whole argument about choosing a formation becomes much smaller. I may be wrong to focus on shape instead of on the quality of training.

One more example keeps me careful. In 2026 I wrote that Germany would not get out of the group, based on average age data and a drop in distance covered. On 27 June 2026, Germany lost 0-2 to South Korea in Kazan, with goals from Kim Young-gwon in the 90+2nd minute and Son Heung-min in the 90+6th, and finished bottom of the group. But on that broadcast I mispronounced Toni Kroos's name three times. A correct argument does not save a wrong detail. I have carried that lesson into every piece since: when you write about systems, check the names again.

Here is a judgement for readers to test. Over the next two seasons, the number of touches in the opposition third by wing-backs will separate successful back-three teams from failing ones more sharply than ever. Any club that buys a back three without buying a wide player capable of producing goals from a defensive position will slide backwards.

People call me a contrarian. I treat that as a job description. In 2026 they called me mad for proposing that matches be cut to 80 minutes. This year they call me a researcher, and I am still looking for a way to prove myself wrong.