The goalkeeper broke every pressing statistic we had

Once keepers became the first passer, pressing turned into a ten against eleven problem that our numbers never adjusted for.

Ten against eleven. That is the arithmetic of pressing a modern build-up, and it is the reason your forwards look tired and your analyst looks confused.

The rule change that let goalkeepers receive a pass inside their own area from a goal kick is the single most consequential thing to happen to defensive coaching in years, and it arrived quietly, framed as a tweak to restarts. What it produced was an extra outfield player who cannot be tackled, stands in the safest square on the pitch, and is now coached from the age of nine to receive under pressure and pass with both feet.

Every pressing metric we use was designed for a world where that player was a formality. Count the passes in the defensive third and you now include a keeper who can hold the ball for four seconds, invite a striker forward, and clip a diagonal over him. Count defensive actions and you exclude by definition the most important thing a pressing forward does, which is stand between a goalkeeper and his preferred option while running at him.

So the arms race has moved to a place the data barely describes. Teams now select strikers partly on their curved run, on whether they can approach a keeper from an angle that removes one foot and one target simultaneously. That is a genuine specialism. It is also almost unmeasurable, because success looks like the goalkeeper choosing a worse pass than he wanted, and nobody records what he wanted.

I will make the unpopular argument. I think the change has been good for the sport and bad for the goalkeeping profession, because clubs now shop for distribution first and shot stopping second, and there are only so many hours in a training week. We have quietly decided that a keeper who saves you four points a season with his hands is worth less than one who lets you press a certain way, and I am not sure anybody checked that trade.

What I would like from the analytics people is not another ratio. It is a model of the goalkeeper's option set, before and after the press arrives, so we can finally say what a striker's run cost the opposition.

Until then we are counting the actions of ten players in a game that has eleven.