A Substitute for War

Basketball philosophy

Team Volatility

with 2 comments


I typically like to have some kind of tight argument in my posts here, but this one is loose. I welcome any feedback from people who’d like to see the data presented in a different light. With the major trades occurring at the trade deadline, I thought it of interest to to take a look at the data surrounding how volatile teams are in terms of player movement. In the above graph, the vertical represents team wins this season, the horizontal represents the amount of players who’ve played for the team in the past 3 years, and the size of the dot represents the payroll of the team right now.

The color of the dot does relate to the team’s uniform, so hopefully you can tell a lot just by looking at the graph, but here’s some specific pieces of info:

1) At the top, we have the Spurs because they have the best record in the league.

2) On the right, we have the Knicks, who win the “most volatile” franchise award by a country mile.

3) On the left, we have the two-time defending champion Lakers, the most stable franchise in the Association.

4) At the bottom, we have those poor Cavaliers.

In general we see a negative correlation between record and volatility, which is no great insight since we would expect poor performance to cause volatility. At the same time, this makes the franchises on the bottom left (Pistons and Pacers) standout – and all the more because both teams are over the salary cap. Is such a combination of attributes ever defensible? Well, with the Pacers most of their bigger salaries come off the books after this season, so I’d say this does make sense. The Pistons though, I don’t see the logic.

One surprise to me was seeing that the Spurs actually aren’t all that stable by this metric. While their core is arguably the most stable in the league, role players come and go on a regular basis for them, which makes it all the more impressive that everyone who plays in San Antonio seems to have a specific niche that contributes a net positive value.

Here’s the data I used to create the graph (thanks to basketball-reference.com and shamsports.com):

Written by Matt Johnson

March 13, 2011 at 12:22 am

2 Responses

Subscribe to comments with RSS.

  1. Does the correlation improve if we filter by players who average 10 or 15 minutes a game (or some other threshold)? A lot of the top teams sign veterans and will change the end of their bench around, but the core remains the same…(maybe)

    ElGee

    March 14, 2011 at 12:47 pm

    • Haven’t done the full analysis (data would require significantly more effort), but this will clearly be true for a team like the Spurs.

      Matt Johnson

      March 14, 2011 at 1:08 pm


Leave a reply to ElGee Cancel reply