Rotation only works if the substitutes play the same game

Squad rotation fails less from fatigue than from tactical mismatch between the first eleven and the men replacing them.

Six changes, a flat performance, and the manager gets asked whether he disrespected the competition. The honest answer is usually that he did not rotate players, he rotated systems, and nobody noticed until the game started.

Here is what I mean. A team pressing with a specific trigger, say the pass into a full back, relies on all ten outfield players agreeing on when to go. That agreement is built in training, over months, with a particular group. Change six of them and the trigger fires late, or fires for three players and not the other seven, and the press becomes worse than no press at all, because now you have a broken line and forty metres behind it.

The players who came in are not less fit. Frequently they are fresher and faster. What they lack is the shared timing, and shared timing is not transferable in the way that fitness is.

So the clubs that rotate successfully are not the ones with the deepest squads. They are the ones with the most legible systems, where the job of each position is defined precisely enough that a replacement can execute it without three months of rehearsal. That is a coaching choice with a real cost, because a highly legible system is also easier for an opponent to prepare against. You trade some tactical surprise for the ability to change personnel.

I would rather have that trade than the alternative, and I think most managers underrate it. The alternative is a beautifully idiosyncratic first eleven and a bench that cannot be used, which means the first eleven plays sixty matches and arrives at the decisive month in pieces.

There is a partial rotation that works better than either extreme, and good sides use it constantly. Keep the spine, change the edges. The centre backs and the holding midfielder carry the structure; the wide players and forwards can be swapped with much less disruption because their jobs are more individual. Change the spine and the whole thing wobbles.

Judge a rotation by whether the shape survived it, not by how many names changed. Those are different measurements and only one of them tells you anything about the manager.