≡ Menu

The Houston Rockets bench revisited


Click for a full-size interactive version

Last year I took a look at the Houston Rockets bench. The data showed that the Rockets had one of the most productive benches in the NBA, but also one of the least utilized. Now I’m revisiting the Rockets bench with updated data.

Technical information

The chart above is very similar to the one from my previous post. The only differences are that I’ve color coded the teams, made the sizes of the circles correspond to the teams’ winning percentages, and added a filter at the top right that allows you to toggle between when I collected data.


Data is calculated from the beginning of the season through specified dates, not in separate segments. In other words, the data for February 10th is not from December 19th to February 10th, but from the beginning of the season to February 10th.

A player’s status as starter/bench is determined by where he is when the game begins. Thus, over the course of a season, the same player’s status changes from game to game (like Darren Collison), and the stats he accumulates will count towards starters or bench depending upon the game.


The Rockets bench has declined in performance a bit, but is still strong compared to the rest of the league. The Rockets bench currently ranks:

  • 6th in points per possession
  • 9th in net points per possession
  • 11th in true shooting percentage

These ranks are all lower than what they were when I first collected data on December 19th. Additionally, the Rockets bench now ranks 28th in minutes played. That is also lower than on December 19th, when the Rockets bench ranked 25th in minutes played.

Bench usage across the NBA didn’t budge very much between December 19th and now, decreasing about 0.5%. Net rating moved even less, increasing from -0.87 to -0.85. What did change is the range of both measures. Extreme outliers from December 19th, such as the Wizards and Heat, are now closer to the league average. Even the Spurs and the Thunder benches, while still markedly more productive than everyone else’s, are a little closer to the average.  Basically, teams are clumping towards the middle of the graph.

Injuries are certainly a part of the conversation, since the Rockets have quite a few. But, like I mentioned in the forums, it’s not like other teams haven’t suffered injuries. What I find interesting is that despite injuries increasing as the season progresses, bench data hasn’t changed much. Usage and productivity have both, for the most part, stayed constant. This means that teams are either playing their remaining bench players more minutes (to make up for minutes lost to injured players), or are playing new players from the practice squads or D-League teams. And these new bench configurations aren’t losing productivity, though this might be due to the fact that everyone’s weaker benches are playing each other, so it’s a zero sum game.

I am even more skeptical now of the Warriors and Trailblazers chances for sustained success. Despite winning, they are the only two teams to play their benches less than the Rockets, and their benches aren’t particularly good. They are one injury away from being a footnote.

What’s interesting is how different teams have responded to a similar set of circumstances. In this respect, the Rockets are interesting. From December 19th to February 10th, the Rockets bench was on the floor 2.1% less of the time, or more than 4x the NBA average. Compare that to the Clippers, whose dot overlapped the Rockets on December 19th. They suffered their own injuries, but their bench minutes only decreased at the NBA average of 0.5%.

There are certainly different philosophies about how to handle injuries and new players. I’ve made it known that, for lots of reasons, I think it’s better to play bench players more. A legitimate counter-argument is that teams need to rely on the players they’ve come to trust instead of introduce new and untested ones. But to me it almost seems dangerous to respond to injuries by running uninjured players even harder, thus increasing the chances that they then become injured.

For four straight games (San Antonio through Phoenix), the Rockets played with an eight man rotation. Yes, people were hurt, but Covington and Brewer dressed for each game and logged zero minutes total. Brooks didn’t play in (I think) two of those games, despite being healthy. On a team with this much depth and injury concerns, there is no reason for Parsons and Harden to approach 40 minutes per game. At this stage, prolonged health is more important than individual wins. And who knows, some players might surprise given the opportunity and the trust.

View this discussion from the forum.

About the author: Richard Li is an independent researcher and consultant. He likes numbers and pictures.

in essays
Follow Red94 for all new post updates
Read previous post:
The Rockets Daily – February 10, 2014