Updated November 17 , 2018 by Seattle Hoo
The Successful Possession Index arose out of a debate over the measurement of a player's contribution to the team. Scoring, rebounds, assists, steals, etc., these things are obvious and often not as meaningful as on first view. What about the things a player does that don't end up in the box score? Everybody agrees that these plays are important, but quantifying them is difficult. The concept of plus-minus attempts to get at it indirectly by measuring the point differential a team experiences with a certain player on the floor. I wanted some way to more directly find a player's contribution to success. Second, I wanted some way to measure the "glue guy" contributions without falling back on a flawed metric like plus-minus.
First came the Glue Index, to measure the plays that do not land in the box score. At the same time, I was collecting highlights and jokingly came up with the Highlights Index, another limited measure of a specific kind of play. But it was too subjective and I found that what I did or did not consider a "highlight" varied based on my feelings about that game while breaking down film (not very many highlights in a loss, I found). So I broadened it and tried to make it more objective.
Thus, the Successful Possessions Index. The Successful Possessions Index is a count of the number of possessions in which a player directly contributes to the successful completion of the possession. It includes all of the "Glue Plays" in addition to box score plays.
The Successful Possessions Index is not intended to be an all-encompassing final word on player evaluation. It is intended to be part of an overall evaluation. I think it dovetails nicely with efficiency ratings. It could answer the question of "how does this player contribute to that +- rating? Is he really doing things that produce it, or is he just out on the floor with four other guys?" A player with a high plus-minus but a low SPI might just be occupying space.
Now to define the elements of the SPI:
A "Possession" is one team's control of the ball. There are Offensive Possessions and Defensive Possessions. When the other team has the ball, that is a Defensive Possession (this is completely UVA-centric. The rest of college basketball exists to be part of our drama). When Virginia has the ball, that is an Offensive Possession. A Possession begins when one team gains control of the ball and ends when control passes to the other team. An offensive rebound merely continues the current possession. A held ball does not end a possession unless the arrow awards the ball to the defending team.
The definition of a Successful Possession is as clear and simple as that of a Possession: A Defensive Possession is successful if the other team does not score at all, and an Offensive Possession is successful if Virginia scores one or more points. The why or the how do not matter. If a defensive possession is successful through no agency of the Hoos (unforced turnover, for example), that possession is still successful, but no "Tags" will be scored to any player, thus no player's SPI will be incremented.
This is where the fuzziness comes in. No scoring system can be completely devoid of judgment calls. What does it mean for a player to have "directly contributed to the success of the possession?" We start from the end and move backwards. A score, a rebound, a steal, a forced turnover, or a drawn foul on a defensive possession all ipso facto produce a successful possession. I then go back and identify the "play" that produced the successful result. A "play" is a continuous series of discrete actions that create an advantage and then turn that advantage into success. Where the setup ends and the play begins is, admittedly, quite subjective, akin to the decision of what makes an "assist".
A scoring play might be as simple as a player picking up the ball and putting it in the basket, or it could involve a rapid series of passes and cuts that result in a player getting a layup. If those actions take place in a short amount of time and lead directly one to the other, then I count it as one play. So a scoring play could be as involved as London runs off a Wilkins screen, curls, catches a pass from Devon, drives into the lane and dishes to Mamadi cutting down the baseline for a dunk. The screen created the advantage for London, and his drive created the opening for Mamadi. On that play, Isaiah would have a Screen, Devon a Pass, London a Cut, Drive, and Pass, and Mamadi a Cut and Dunk. All four players' SPIs are incremented.
Or on defense, Isaiah might Hedge a dribbler off course, Jack Rotates to Isaiah's man, Challenges the shot, which misses. Jack and Devon Boxout, and London Isaiah Recovers to grab the Rebound. Isaiah is scored a Hedge, Recover and Rebound; Jack has a Rotation, Challenge and Boxout; and Devon is credited for a Boxout. All three players receive an SPI point.
The following are the plays that are charted. I call them Tags. Each term is a link to its definition in the Hoopedia. Each of these Tags can be searched for in the Hoos.Space Possessions Database.