The Rockets narrative— i.e., "Tracy McGrady is a superstar and the Rockets can’t win in the playoffs without him, but he’s not the type of clutch superstar who can win in the playoffs, blah, blah blah." — continues to be the dominant theme among most of the mainstream media in regard to the local NBA team.
In reality, McGrady is long past being a bona fide NBA superstar and really is not much more than a bit above-average NBA player at this point in his career.
However, despite McGrady’s and newcomer Ron Artest’s relative mediocrity this season, the Rockets are muddling along with a 29-19 record and — barring further injuries — are in the thick of the race for an upper division Western Conference playoff spot.
Given the Rockets’ narrative, how is that possible? Let’s take a look at the numbers.
Dave Berri doesn’t rely on subjective narratives and instead continues to provide some of the best objective analysis of what is really happening in the NBA over at his Wages of Wins blog. Here are his mid-season player rankings (organized by team here), which are much more revealing than the Rockets narrative.
Berri’s ratings, which he calls the "Wins Produced Model," begins with a player’s statistics — his points scored, shot attempts, rebounds, steals, turnovers, etc.. — and translates them into how many wins those statistics create over the course of a season.
Under Berri’s system, players who do many things well — such as former stars Oscar Robertson, Bill Russell, Wilt Chamberlain, Magic Johnson, Larry Bird, Michael Jordan and Hakeem Olajuwon, as well as current stars LeBron James, Shaquille O’Neal, Tim Duncan, and Kevin Garnett — are among the most productive players, just as most of us know intuitively.
On the other hand, players who only score a lot of points but who have deficiencies such as low shooting efficiency, high turnover rates, and poor rebounding skills are not as productive as many folks believe based on their high scoring. Current examples of that type of player are Allen Iverson, Antoine Walker, and Carmelo Anthony.
Finally, players who are excellent rebounders — such as the aforementioned Russell or Dennis Rodman, or current players Ben Wallace and Marcus Camby — are usually much more productive than most folks give them credit for.
So, players who do not shoot efficiently, or who turn the ball over frequently, do not help their team win many games. On the other hand, players who generate large numbers of rebounds, or who score consistently and efficiently, or who create steals — they generally help their teams win more games than average.
Thus, Berri’s Wins Produced statistic tells us how productive a player has been. What it does not tell us is why a player shoots inefficiently, commits turnovers, rebounds well or creates steals. That’s a question for team management and coaches to figure out.
Having said all that, the following are few observations on the mid-season statistics:
The most productive players by position are as follows:
The Rockets most productive player so far this season is Yao Ming, who is the NBA’s 17th most productive player at 7.1 WP, which is only three spots below the Lakers’ Kobe Bryant (7.4 WP).
The Rockets second most-productive player this season is not McGrady, or Artest (120th-ranked at 2.0 WP), or PG Rafer Alston. It’s PF Luis Scola (62nd most productive at 3.8 WP).
Even with all his physical problems, McGrady (73rd-ranked at 3.2 WP) is the third most productive player on the team. Having said that, the Rockets aren’t any more productive by playing McGrady than SF Carl Landry (74th-ranked at 3.2 WP).
Are the Rockets ever going to have a highly productive point guard again (Alston — 134th-ranked at 1.6 WP; Aaron Brooks — 196th-ranked at .7 WP)?
Also, SF Shane Battier (157th-ranked at 1.3 WP) is a marginal starter at this point in time, although fan favorite Von Wafer (173rd-ranked at 1.1 WP) really isn’t a better alternative.
Although the Rockets do not have many highly-productive players, they also do not have any players who are actually counter-productive — i.e., who have a negative WP. Most teams have at least a few counter-productive players.
If McGrady and Artest ever get healthy, then the Rockets’ best chance of finally winning a playoff series (it has been 12 years now) may well be playing a lineup of McGrady at the point with Artest at shooting guard, Landry at SF with Scola at PF, and of course Yao at C.
By the way, in view of all this, why do so many folks continue to expect so much from McGrady?
Call it the curse of the big contract — McGrady is pulling down a total of $40 million guaranteed over this season and next. Many folks just can’t come to terms with the fact that sometimes the player gets the better of management in contact negotiations.
Me, I just think McGrady has a good agent. ;^)
1) I’m skeptical that rebounding stats are as important as some of these statistical analyses claim. A rebound isn’t quite like a steal or block or a football interception: someone like Dennis Rodman was grabbing a lot of rebounds at the expense of his teammates’ rebound statistics, just like Iverson’s scoring is at the expense of his teammates’.
2) While I agree that the ppg of an Iverson or C. Anthony tends to be overrated because of his ball-hogging and low shooting percentage, a number of the statistical measures fail to account for the value of being able to create a shot. To take an extreme example, G. Muresan is overvalued by these measures, because he had a ludicrously high shooting percentage, but he was a liability because of his lack of mobility, or the Bullets would’ve been able to feed him for more than 12 high-percentage shots per 30 minutes.
3) NBA assist counts aren’t quite as subjective as MLB “error” counts, but they’re still somewhat scorer-dependent, leading to GIGO problems.
4) And all of the statistical measures fail to adequately account for defensive value, the difference between being able to clog passing lanes or alter a shot and being a matador waving people driving into the basket as they score–in part because doing so is so team-dependent.
Ted, comments parallel your paragraphs:
1) Rebounding stats involve a change of possession and, thus, are more important than blocks, which may or may not result in COP. My sense is that the best way to measure a rebounder is how he compares to a league-average rebounder playing in the NBA during the same time as the rebounder. Based on that measure, both Russell and Rodman were off the charts good, which is probably one of the main reasons why their respective teams were generally quite successful.
2) Interesting point on Muresan. I’ll watch to see if any current players are comparable and how they rank under Berri’s system.
3) NBA stats on assists and steals have been normalized over the past 10 years or so. Thus, these stats are pretty accurate. You are correct that such stats were much less accurate in prior eras.
4) I’m not sure whether Berri has a defensive component to his rating beyond steals and rebounds. If not, you are correct that a player such as Battier is undervalued because of the defensive skills he brings to the table.