Turnovers in your own half are not all equal

Where a defence is facing when it loses the ball matters more than how far from goal the loss happened.

Two giveaways thirty metres from goal, and one of them is nearly harmless while the other is a chance conceded. The distance is identical. What differs is orientation, and orientation is the variable that possession-loss maps almost never show.

Lose the ball while your defenders are square to the play, facing the opponent, with the goalkeeper behind them and no runner in behind, and the situation is uncomfortable but controlled. The receiving player has to beat someone who is set. Now lose it while your defenders are turned, moving up the pitch, having just committed to an offside line, and the same turnover puts a forward into open grass running towards a keeper. Same map dot. Completely different consequence.

This is why I distrust the simple heat maps of where possession was conceded. They encode a real signal, since losses near your own box are on average worse, but the averaging hides the thing that actually matters and coaches end up telling players to be safer in areas where safety was never the issue.

The other variable is the number of opponents ahead of the ball at the moment of loss, and this one is at least measurable from tracking data. A pass intercepted with two opponents ahead of it is a counterattack. The same pass intercepted with six ahead of it is a nuisance. Teams that build patiently through the back have learned to accept the first kind of risk while eliminating the second, which is why they will play a risky short pass under pressure and then refuse an ambitious one twenty metres further forward.

That looks contradictory from the stands. It is not. The short pass, if lost, is lost in front of a set defence with everybody facing the right way. The ambitious one, if lost, is lost with three players committed forward.

So when a defender is criticised for a giveaway, the question I want asked is what the shape behind him was when he made the choice, because half the time the answer is that a midfielder had already left. The error was earlier and it was someone else's.

Where the ball was lost is the least interesting thing about how it was lost.