Reply to thread

Message: [QUOTE="nelsonmuntz, post: 4804893, member: 833"] That is exactly right. The biggest problem with the current version of NBA analytics is that it does not take into account that every event in a basketball game has a direct impact on the next event in the basketball game. In football or baseball, everything stops after every play and the players set up again. There are some variables, such as down and distance in football or players on base in baseball that change the probability of the next event, but each play is a discrete event with its own calculatable probability. Play is continual in basketball, and even when it stops, such as for a free throw or a ball out of bounds, play starts up again pretty quickly. Very few models take into account things like offensive rebounds, which are the highest probability shot an offensive team can have, but are also events dependent on the play right before it. Depending on the dataset, a short range miss will result in an offensive put-back opportunity about 40% of the time. So a 55% inside shooter missing his shot had a 40% chance of getting the shot and getting a 60% scoring opportunity. The datasets are varied, but a missed 3 pointer only results in an offensive rebound about 25% of the time, and that is rarely a putback opportunity. So a 33% 3 point shooter and a 50% interior 2 point shooter are not actually equivalent, because if you look at the probability of the next event, the interior player has a much higher probability of creating a scoring opportunity even if he misses than a 3 point shooter. I would also make the case that drives to the basket are lower probability plays than a pass to an interior player, both in the likelihood of the player scoring from the same spot on the floor, and the quality of shot that the action of getting the ball to the interior in a drive vs. a pass creates when that player kicks the ball out. I think that as basketball incorporates continuous probability analysis into its models, there will be changes in what is considered a good shot and the type of players that will be considered analytically attractive. I think the game will come around to players like Sanogo again. [/QUOTE]

Verification: First name of men's bb coach