Rotation should replace partnerships, not people

Changing five players damages every partnership on the pitch, which is why heavily rotated teams look strange rather than tired.

A manager makes five changes for a midweek tie and the team plays like strangers. The usual explanation is rustiness. I think the real explanation is combinatorial, and it is a bit more interesting than that.

Football is played in pairs and small groups. The centre back and the holding midfielder have an understanding about who steps into the space in front. The full back and the winger have a private agreement about who goes outside. Those relationships are learned, and they are learned specifically, between those two men, in those roles. Swap one member of the pair and the relationship resets to something generic.

Now count. If you change five players scattered across the back four, midfield and attack, you have not disturbed five relationships. You have disturbed every adjacency each of those players sat inside, which on a pitch of eleven is a large fraction of the total. The team is not five-elevenths new. Functionally it is most of the way to new.

Which suggests an alternative that hardly anybody commits to. Rotate by unit. Rest the whole full back and winger pair on one side and leave the other flank untouched. Change the front three together and keep the spine alone. You lose the same number of legs, you buy the same amount of freshness, and you preserve the surviving partnerships whole instead of halving all of them at once.

The counterargument is that you cannot always afford to lose an entire unit, and that the strongest available eleven does not respect these tidy blocks. True. Also the reason most rotation looks the way it does. Managers pick the best eleven available minus whoever needs a rest, and the result is a scatter, not a structure. Then everyone blames the players for looking disjointed.

There is a recruitment angle too. Some players are unusually easy to slot beside anybody. They talk, they point, they take the simple option, they position themselves so the man next to them does not have to guess. That trait is worth real money in a squad that plays sixty matches, and it is almost invisible in the numbers we publish. It shows up only as the absence of chaos.

I would rather have a rotation policy that costs me a slightly weaker eleven and keeps four partnerships intact than one that fields the theoretically best players and asks them to improvise the shape.

Freshness is not the only thing that decays.