19 August 2009

Your high school coach was right after all... The Lakers WOULD be better off without Kobe!

Wired magazine has this compelling article for those of us who've been waiting for basketball stats-keepers to catch up to baseball's cutting edge analysts. I know this has been keeping you up at night too, right? Um, Ok. But I really have been waiting for this ever since, jaw agape, I received the gospel of selflessness on my father's knee as he regaled me with tales Wooden's UCLA teams. In truth, those of us lucky enough to grow up during the best NBA decade ever (the '80's and you're a tool if you disagree) saw Doug Collins' Bulls get beat every May and Phil Jackson's Bulls win every June. It takes Allen Iverson not to draw the appropriate conclusion.

Now someone has finally gone to the trouble of doing some quantitative analysis as to why good ole' fashioned teamwork really does, well, uh, work.

Now the conclusion seems to follow from the general application of systems analysis principals. It gets more gritty when you factor in defensive matchups, whether Kobe, Allen, whomever, is playing with a big/small lineup, against a big/small lineup, whether possessions are at the end of a quarter, who's in foul trouble, etc. etc. This would seem to be where good coaching comes in. However, the obvious conclusion seems to be that, ceterus pluribus, playing five-man basketball is more effective than playing matchups. Here's the article.



Sorry, Lakers fans, Kobe could be holding your team’s offense back.

Elite players could be taking too many shots for optimal offensive efficiency, according to new mathematical analysis using network theory.

Treating each player like a pathway to get the ball into the basket, a physicist has deduced that the most efficient path to a basket does not always run through star players like Kobe Bryant, Lebron James, or Ray Allen, even though they are better shooters than their teammates.

“The idea that a team could improve after losing one of its best players may in fact have a network-based justification, and not just a psychological one,” wrote Brian Skinner, a physicist at the University of Minnesota, Minneapolis, in a paper posted to the arXiv.org. (Skinner is no relation to the other Brian Skinner, Baylor standout, Los Angeles Clippers power forward and 22nd pick in the 1998 NBA Draft.)

First, Skinner explains how people making the best decisions for themselves can hurt the efficiency of a total system. Let’s say that there are two roads, a highway and an alley shortcut. The alley takes up to ten minutes, but sometimes less depending on traffic, and the highway always takes ten minutes. Individuals realize they could save time by taking the alley, so they do. Unfortunately, when everyone takes the shortcut, it ends up taking the full ten minutes.

By analogy, perhaps, getting rid of Kobe Bryant could actually make things better by dispersing the “cars” (i.e. possessions) more evenly. Offensive balance could reduce “traffic,” making putting the ball in the basket easier.


By analogy, perhaps, getting rid of Kobe Bryant could actually make things better by dispersing the “cars” (i.e. possessions) more evenly. Offensive balance could reduce “traffic,” making putting the ball in the basket easier.

The key assumption is that a player’s real shooting percentage goes down as they take a greater percentage of a team’s shots. Skinner’s stats show this appears to be the case with Allen — and it stands to reason, too. As a player dominates an offense more, the defense adjusts. They double the player, devote more attention to him, and likely deny him high quality shots that are likely to go in. (We might call this the Iverson effect.)

So, if one were to distribute the number of shots a player takes on the basis of their shooting skill, the math says the team’s overall shooting percentage would go down. If Ray Allen takes only as many shots as the rest of his teammates, he will make more of them than he would if he put it up on 40 percent of the possessions.

By distributing shots more evenly, then, the team’s overall shooting efficiency could go up, even if the other players on the team are only average shooters. For the star player, it’s a bit like that old adage, “You’re promoted until you’re incompetent.”

Of course, Skinner’s analysis doesn’t take defense into account and the interplay between the shooting skills of the best players versus the worst players could change the results somewhat, but it will probably add fuel to the barbershop debates of Brooklyn over whether or not the Knicks really would have been better without Patrick Ewing.

5 comments:

  1. Kirk, I'm not buying any of this. I'm a big fan of statistical analysis, but this seems to be one of those examples of stats not holding up to common sense.

    I mean, I really do like stats. I love all these new baseball stats that prove what we've known all along: things ranging from Derek Jeter being a massively overrated defensive shortstop to the actual quantifying of how many runs good defense is worth.

    We go to stats to disprove common assertions of the ignorant fan. And despite the wonderful self-congratulation that we statheads revel in, the fact remains that this is a load of crap.

    Five Bruce Bowens wouldn't beat the Lakers or the Magic this year. Not in a million years.

    ReplyDelete
  2. No one is advocating stacking rosters with five Bruce Bowens. Perhaps this vignette will provide a better contrast; when the USA basketball team lost in the medal round to Argentina in 2004, Greg Knowlden quipped to me that Chauncey Billups, Rip Hamilton, Tayshaun Prince, Rasheed Wallace, and Ben Wallace could have sleepwalked to the gold medal. This was, of course, the NBA championship Detroit Pistons starting lineup (a lineup, incidently, that beat Kobe that year in the finals).

    ReplyDelete
  3. Why, then, could those five have easily done what a Dream Team could not?

    ReplyDelete
  4. The 2004 Olympic team was hardly a Dream Team.

    I'm not arguing that pure offensive talent can get the job done. There's a lot to be said for chemistry, coachability, and a basketball mind.

    The second two things are what separate Allen Iverson, Stephon Marbury, and J.R. Smith from Kobe Bryant and LeBron James.

    ReplyDelete
  5. I'm going to have to disagree with this as well. I'd like to see the math behind the analysis. I have a gut feeling you could work the math to make it appear as if the more a team's star shoots the better the team does.

    My biggest issue is that the author's example uses cars and computers. Cars on the highway or alley and computers in a network will be roughly the same. Not all NBA players are equal and this is where I think the argument falls apart.

    ReplyDelete