Rotate the unit, not the individual

Changing four players scattered across a team damages more than changing four who already play beside each other.

Managers talk about rotation as a headcount. Six changes, four changes, freshen it up. The number is almost meaningless on its own, because the cost of a change depends entirely on who the replacement has to coordinate with.

Think about what a defensive line actually is. Four players holding a shared height, stepping together, communicating a trap or a drop in about a second. That is not four skills, it is one skill distributed across four people, and it is built by repetition together rather than by individual quality. Swap one of them and you have degraded the whole unit. Swap the left back and the left centre back on the same night and you have effectively replaced the left side of your team with two people who have never done it in a competitive fixture.

Now compare a straight swap of both wingers. Wingers operate at the edge of the structure, with fewer coordinated actions, mostly one against one, and the cost of unfamiliarity is far lower. Two changes there is not the same event as two changes in the back four, even though the team sheet shows the same number.

So the sensible policy is to rotate along the seams. Keep the spine intact and change the outside. Or, if you have real depth, change an entire connected block at once, so at least the replacements have trained as a set, rather than sprinkling fresh players through a familiar structure where every one of them creates a new untested relationship.

There is a version of this argument I do not accept, which is that you should barely rotate at all because the partnerships are worth more than the rest. Fixture congestion is not a preference. Play a first choice centre back through three competitions and you are not preserving a partnership, you are borrowing against it, and the repayment arrives as a soft tissue injury in March that costs you the player for eight weeks rather than for one game.

The awkward part is that rotation policy is judged retrospectively and dishonestly. If the rotated side wins, the manager was bold. If it loses, he disrespected the competition. Nobody ever credits the change that prevented the injury, because the injury that did not happen has no highlight.

Count relationships, not players. It is a different number and it is the one that predicts the mess.