How Pep Guardiola Learned To Trust His Players

He didn't tinker this time. He didn't need to.

27th April 2011. The last time Pep Guardiola won a Champions League semi-final first leg. It took him a full decade plus one day to do it again.

His team that day was as follows: Victor Valdes, Dani Alves, Gerard Pique, Javier Mascherano, Carles Puyol, Sergio Busquets, Xavi, Seydou Keita, Pedro, Lionel Messi, David Villa. Guardiola didn’t have Andres Iniesta, so Keita needed to fill in. He also had no fit left-backs, so in came Puyol at an unfamiliar position. But other than those enforced changes, he just played his best team and won. He stuck with the same side in the second leg and they got the job done.

The following year, he played a 4-3-3 shape with Iniesta on the left that became a kind of 4-2-3-1/4-2-4/4-2-amorphous blob in possession. Presumably, this was to get Cesc Fabregas into the team. They lost. In the second leg, he switched to a sort of 3-3-3-1 shape with Iniesta as the only player wide on the left and Isaac Cuenca (yes, he exists) on the right. They drew and went out.

In his first year at Bayern, they came up against Real Madrid in the semis. Pep sprang a surprise in the first leg by playing Bastian Schweinsteiger as a number ten to retain possession better and close down Madrid’s deeper midfielders before they could launch fast counters. Bayern lost the game. In the return leg, he did the most uncharacteristic thing of all and played the shape his players wanted, dropping what they’d done all year for a gung ho 4-2-3-1. It failed spectacularly. The next year against Barcelona, he went for an extreme man-marking job before switching to a bizarre diamond shape, attempting to have four passing midfielders to control possession and stop the counters. It failed. He stuck with it for the second leg, and it still couldn’t get him through. The third year he went out to Atletico Madrid mostly playing his own way, but Atleti did an Atleti and there’s not too much to do about that.

At City, the problem has so far become more acute. He didn’t even make the semis previously. In his first year against Monaco, he moved Fernandinho over to left-back, but this had to happen since his team just weren’t very good that year. Afterwards, when he finally had the quality he needed, he still couldn’t let it settle. The following year against Liverpool, he played Ilkay Gundogan from the right to add in extra control in midfield and better prevent fast counters from happening. It did not work at all. In the second leg, he actually gave up on control and attempted to beat Liverpool at their own transition-based game, which sort of worked ok but it didn’t matter in the end. The next year against Tottenham, he played a weird 4-4-2 shape because... fuck, I don’t even know why, you’d need to ask him. The second leg he played closer to his default system, but with Gundogan at the base of midfield, likely to help control possession and stop counters again. It also didn’t get him through.

Last year against Lyon was the most cautiously reactive he’d ever been. He was so scared of the counters – against Lyon! – that he switched to a three centre back system with more solidity without the ball. As you surely know, it didn’t work. As Kevin De Bruyne said: different year, same stuff. It inspired me to write a lot of words on why Guardiola kept failing at this stage. At the time, I felt as though he was trapped in an endless cycle of repeating these mistakes:

“So we have two problems: there is a weakness in his system when playing these sorts of games, and his attempts to correct it do not solve the problem and might actually confuse his players. And this is maybe the sad part: I don’t think he’s going to fix it. That doesn’t mean he can’t win another Champions League title if things go right one year. But I do not think he’s going to fundamentally reshape his problems in these games.

Ultimately no system in football is complete. Everything has its flaws. As time goes on and certain elements dominate, teams will more and more look to exploit those flaws. And such, unless he really, truly does the work to rethink some elements of his system, will Guardiola keep facing these problems in the Champions League knockout stages. There’s an opening in the Death Star thermal exhaust port. Teams are going to keep firing at it.”

It looked like Guardiola was on a clear path to decline as recently as October. His style seemed desperate to add in more controlled periods of possession along with the high pressing style and it just wasn’t working. Every time he put in a better passer over a more physical option, it made them more open on the transition.

And then, suddenly, he just fixed it. City won both legs against PSG on the night without any huge surprises to the starting eleven. And he fixed it by becoming exactly what he wasn’t supposed to be: efficient, ruthless, boring. His team this season is probably the least enjoyable side he’s ever produced, at least in my subjective opinion. It’s also the most effective he’s had since Barcelona.

His previous approach to adding more control was to swap out more classically defensive players for nice passers to help retain possession. As Fernandinho went from starting every game to a rotation option, he was replaced by the gentler souls of Rodri and Ilkay Gundogan. Fabian Delph gave way to Oleksandr Zinchenko. But the attack kept its mix of technique and speed in behind. Last season, he would typically start Raheem Sterling and one of either Sergio Aguero or Gabriel Jesus for that threat in behind and goalscoring presence.

This season, he often hasn’t started anyone of that profile at all. It’s been about players who want to receive the ball to feet and keep possession. The emergence of Phil Foden has helped here, but he’s hardly the root cause. Guardiola has deliberately moved away from hitting the ball fast to players in space. Last season, about 8% of the passes City completed were “progressive”1. This year it’s 7%. If those sound like small margins, that’s because they are. Guardiola has kept the core of what he’s been doing, but just moved the sliders towards controlled possession rather than fast chance creation and high pressing.

As Ryan O’Hanlon has pointed out, City have become the slowest team in the Premier League. Their sequences of possession are longer than anyone else, and those sequences are slower at moving the ball forward than every other team. In any other context, you’d assume those stats were highly critical, but this has been extraordinarily effective. In the pandemic era, Guardiola has found a way to kill off the counter attack: make everyone play at walking pace.

At the same time, he’s turned the high pressing down. In terms of passes per defensive action, essentially how many passes you let the opponent make before trying to win the ball back, City are now only the seventh most aggressive in the Premier League, after leading it as recently as 2018/19. This has been a gradual scaling back, generally doing it a little less each year, and this season feels like the apex. What we have is a team that can defend quite differently. In the second leg against PSG, you had Ruben Dias making block after block as City dug deep to defend with their hearts on their sleeves. You just wouldn’t have seen that with previous aggressive high pressing Guardiola teams.

All of this has been why City have defended so well this season. They’ve cut out those terrible high-quality chances conceded when they would just get ripped open on the break. And with Guardiola having mastered the defensive control to prevent those transitions, it wasn’t a surprise at all to see him trust his team and avoid making wholesale changes.

This City side, the second great City side Guardiola has produced, is a very different beast from what we’ve seen before. While Jose Mourinho loves to claim possession is “for the philosophers”, his famous nemesis is using it for nefarious means. City are the most streetwise, solid, hard to break down team in football right now. And Guardiola could finally win the Champions League again because he can turn his methods into hard-fought “ugly” wins.


Stats are from FBRef using data from StatsBomb, except PPDA which is from Understat.