Why possession percentage doesn't predict win probability
Possession correlates weakly with league points because ball retention does not equal shot quality. A team can control 65% of the ball through sideways passing and defensive possession yet create zero clear chances. Conversely, a team defending deep on 35% possession can generate three shots on target through direct play and set pieces.
This disconnect emerges because possession is a volume metric while points depend on efficiency. An underdog team pressing high and winning the ball in advanced positions might average only 40% possession but score more goals per 90 minutes than a team that holds 55% but plays predictably in front of goal.
Contextual possession: the misunderstood metric
Some teams deliberately cede possession to draw opponents forward, creating space for transitions. Luis Enrique's Roma, for example, often finished matches with 40-45% possession but outscored opponents through counter-pressing. Meanwhile, possession-dominant teams can become predictable, allowing defenders to set shape and limit space.
Expected goals (xG) and shot-creating actions explain points far better than raw possession percentage. A team needs to combine possession with penetration: either moving the ball into dangerous areas or using possession to isolate defenders one-on-one.