I’ll be honest, it’s been a long time since I’ve played basketball. I still remember who the point guard is, though. He’s that guy that dribbles better than you, passes better than you, and is probably just better than you all around. He’s the playmaker; if he’s really good at his job, you just get him the ball and he takes care of the rest.
Arizona State University researchers Jennifer Fewell and Dieter Armbruster recently published a network analysis of ball movement during the 2010 NBA playoffs in which they claim to have found an ideal model for predicting results of games. This model was built by tracking ball movement during games and graphing this movement as a network; the nodes consist of the players, various means of getting possession, and various results of possession, while the directed links indicate ball movement. They come to the conclusion that the standard point guard-centric offense is not nearly as successful as what has come to be known as the “triangle offense”, in which the ball moves freely from player to player without necessarily going through the point guard first.
In the point guard-centric offense, the point guard seems to occupy a role loosely akin to that of a gatekeeper; the strongest paths from the “inbound pass”, “rebound”, and “steal” nodes to the “success” node pass through the “point guard” node. In fact, if we only look at the strongest (red) links, he is a gatekeeper in the strictest sense. The effect of this is that, if the defense is able to prevent the ball from moving to or from the point guard effectively, there is little redundancy in the network to allow the ball to move around him. This is where the triangle offense succeeds; there are multiple paths of strong links connecting the “inbound ball”, “rebound”, and “steal” nodes to “success”, making it much harder to defend against.