Possession is 9/10 of the Law

Note: Written in November, 2009.

As in football, I believe rating a team’s offense and defense should be based on a per-possession* basis rather than a per-game basis. Why? Just look at the Warriors. Does Golden State really have the best offense in the league? Don’t get me wrong. Ellis, Jackson, Maggette and crew can put the ball in the hole, shooting 49% overall and 39.5% from deep, but those 112 points per game are somewhat due to the fact that they get up and down the floor more than any other team – about 108 times a game.

Would you say, for instance, that the Warriors have a better offense than the Nuggets since they’re scoring 5 more points per game than Denver? In fact, since Denver slows it up a little more, getting only 102 possessions per game, Denver is actually scoring slightly more points per possession than Golden State. Since certain games are played at different paces—depending on the teams involved, the refereeing, which team is home, and whether or not a team played the night before— purely rating teams on points scored and allowed per game is a flawed method for estimating efficiency.

It also seems apparent that faster-paced teams tend to also give up more points per possession. (Obviously these faster-paced teams give up more points per game because the opposing teams get the same number of possessions—it’s a rule or something—and therefore more chances to score.) Looking at the difference between teams’ points scored and points allowed per possession, the only grouping of teams in the positives is the slower-paced teams. The medium and fast-paced teams actually average negative differentials as a whole. It will be interesting to see if this trend holds for the whole season, since it’s hard to remember a really fast-paced team that won an NBA championship recently.

What else can we use this possession analysis for?

We can see how teams play at various paces. Teams that can adjust well to slow and fast games are likely to have a good chance at the title, since they will see different styles of play in the 17 million rounds (4) that comprise the NBA Playoffs.

We can better compare players from two different teams. Many argued two years ago that Golden State’s Baron Davis was much more deserving of an All Star spot than Portland’s Brandon Roy. However, Portland’s grind-it-out approach limited Roy’s total possessions and easy basket opportunities, and his production likely made up a larger percentage of his team’s overall production than Davis’.

This season I will follow the Portland Trailblazers and their ability to play effectively at various paces. Thus far this season, Portland is allowing the least points per possession of any NBA team. While their schedule has not been too tough, the offenses they have played against are actually scoring at the league average rate, indicating that this could be a legitimately good defense. Oden is quicker this season, and the Vanilla Gorilla is still one of the best defensive centers in the league. Not surprisingly, the Blazers lead all teams in allowing the least points per field goal attempt, indicating that teams are getting difficult looks against their defense. If it is true that defense wins championships, then there’s no reason not to quit despairing for Rip City fans. There are four negatives, figure it out.

And finally, the tops teams offensively using this possession system are (1 through 5) Toronto, Phoenix, Atlanta, San Antonio and Boston, while the worst five teams go (in getting-worse-order)  Golden State, New York, Memphis, New Orleans and Toronto. So while some fast-paced teams are in the bunch, pace doesn’t necessarily correlate with points per possession. Boston and San Antonio don’t put the scoreboard operator on the DL with Carpal Tunnel, yet they score more efficiently than most teams. In fact the correlation coefficient between scoring efficiency and pace of play is 0.14. If correlation coefficients aren’t your thing, just know that it wouldn’t be worth writing home about unless it was much closer to 1.0. Basically pace and offensive efficiency have very little to do with one another.

On the other hand, the correlation to defensive efficiency (allowing less points per possession) is a much stronger 0.42. Perhaps knowing when to slow the game down prevents easy baskets for the opponent… It will be interesting to see if these trends continue throughout the season. I’ll keep you posted.

*Again, as with football, I count possessions based on how many ways a team can end a possession. Teams can end a possession with a field goal attempt, a free throw attempt, or a turnover. I estimate that an average of 1.8 free throws are shot per trip to the charity stripe, so every 1.8 free throws equates to one possession. The formula for possessions looks like this: FGA + FTA/1.8 + TO – OFFREBS. Why minus offensive rebounds? Because I believe it’s still part of the same possession, and offensive rebounding should be included in measuring offensive efficiency.