David Biderman – of the Wall Street Journal – reports today on The Teams that Make the Nets Look Good.
Biderman’s story is based on the following question he asked me to consider: Across the past 10 years, have their been other teams in professional sports as bad as the Nets?
To answer this question – as Biderman reports — I looked at how many standard deviations each team is below the average performance in the league. For example, New Jersey’s current winning percentage of 0.111 is 2.4 standard deviations below the average mark of 0.500. As Biderman notes, this marks actually trails the performance of the Detroit Tigers in 2003, the Atlanta Thrashers of 1999-00, the Kansas City Royals of 2005, and the Detroit Lions of 2008 (yes, two Detroit teams make this list). So the Nets – by this measure – are indeed the worst team in the NBA across the past 10 years. But they are not the worst team in professional sports.
The focus on standard deviations is necessary if one wishes to make comparisons across sports. If one wishes to focus solely on the NBA, though, one can use a different measure that makes the Nets look slightly better. New Jersey currently has an efficiency differential (offensive efficiency minus defensive efficiency) of -10.88. Since 1973-74 (the first year one can calculate efficiency differential) the following teams have posted a differential below what we see from the current Nets.
- Dallas Mavericks [1992-93]: -14.70
- Denver Nuggets [1997-98]: -12.63
- LA Clippers [1999-2000]: -11.89
- Vancouver Grizzlies [1996-97]: -11.17
- Houston Rockets [1982-83]: -10.95
If we maintain our focus on the just the last 10 years, we see that only the Clippers of 1999-00 are doing worse. So across the last decade – if we focus on efficiency differential – the Nets are only the second worse team in the NBA.
One might wonder how the Nets fell so far. To address this issue we need to consider the performance of the individual players. As Table One reports, the Nets do have four above average players this season (Brook Lopez, Kris Humphries, Josh Boone, Courtney Lee). None of these players, though, are far above the average WP48 [Wins Produced per 48 minutes] mark of 0.100. And nine players employed by the Nets this season have posted WP48 marks in the negative range.
If we look at how the veterans this team has employed performed last year, we see this team shouldn’t be this bad. The team’s current efficiency differential (and Wins Produced) is consistent with a team that should have already won 10 games (so the Nets won-loss record is a bit misleading). The 2008-09 performance of these veterans, though, suggests this team should have already won 20 games.
When we look at the individual players we see that much of the team’s decline is tied to the play of Rafer Alston, Devin Harris, and Keyon Dooling. What do these players have in common? All three log time at the point guard position. So that one position has been the reason why the Nets have moved from “bad” to “all-time horrible.”
Let me close by noting – as the following list indicates — that this is the fifth time Biderman and the Wall Street Journal have referenced my work in recent months.
The sports section at the Wall Street Journal has focused tremendously on how numbers inform our understanding of sports. So if you are interested in this aspect of sports (and since you are here I suspect you are), you might want to think about reading the WSJ sports section on a regular basis.
And one last note (this just came in as I wrote this post)… Martin Schmidt says a copy of Stumbling on Wins arrived in the mail today (I haven’t checked my mail yet). The book is officially released on March 26, so it won’t be long until everyone is able to read our latest.
The WoW Journal Comments Policy
Our research on the NBA was summarized HERE.
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.