Two shortish articles out today.
First in Slate we tackle the myth that NBA rule changes have made defense less important.
The in The American Prospect Online we examine the use and (mostly) abuse of demographic targeting as an electoral strategy.
« WMD Counterfactuals | Main | Harman Versus Hastings »
Articles Appearing Elsewhere
14 Nov 2006 03:13 pm
Comments (20)
Whaddaya mean, "we"?
The juxtaposition intrigues me: do you find your interest in politics operates in vaguely the same way as your interest in basketball? I have a theory that to really engage in american politics, you have to treat it as a sport; it's just too remote from everyday life, or even policy and governance.
Two factors determine who wins a basketball game: how many points your team scores and how many points the other team scores.
Good thing Slate is free.
Matt claims that "it stands to reason that offense and defense should have exactly the same importance," but this is faulty reasoning.
Imagine we played a game that consisted entirely of each team shooting 100 free throws, while the other team tries to distract them by yelling and jumping up and down. It would still be the case that the outcome would be determined by "how many points your team scores and how many points the other team scores." But the ability to yell and jump up and down would probably be much less important than the ability to shoot, simply because most of the variation in actual performance would be due to differences in shooting ability (offense), not due to differences in yelling ability (defense).
Oh god no, you're writing for Slate now? Where washed up hacks go to die? You haven't even turned 40.
I'm pretty sure he hasn't even turned 30.
Extending ed johnson's point, it is possible for one of offense or defense to have greater variability than the other.
If standard deviation for defensive performance was 1.5 point/100 possessions and the standard deviation for offensive performace was 3 points / 100 possessions than it would be much more valuable (in terms of wins and losses) to make a chance to a team that would lower its defensive performance by 1 SD and raised it's offensive performance 1 SD.
Similarly, in that case a team's W/L record could be predicted relatively accurately just by looking at the team's ranking in offensive efficiency.
This would be the case in ed johnson's example.
In the actual NBA I believe the variability of offense and defense are almost exactly the same (I calculated SD preparing this post (scroll about halfway down the page).
Have you started to refer to yourself with the royal "we", Yglesias?
NickS,
The SD for offense and defense has to be the same, because every possession is both offense and defense, right?
Matt,
I'm afraid that's not a very high quality piece of analysis. The theoretical argument about offense vs defense is much more complicated than what you describe, and your statistical analysis is worse.
You use 2 finals matchups as examples, a rather small sample set to begin with. But even those examples give no support to your argument. You claim that Miami was a more balanced squad because they were ranked 9th in defense and 7th in offense. Dallas was ranked 10th in defense and 1st in offense. Is 10th really that much worse than 9th?
Is 10th really that much worse than 9th?
No, but 7th isn't insignificantly lower than 1st. Yet despite this offensive disparity, they won. It's illustrating that defense isn't of secondary importance.
Okay, let's keep playing the theoretical game. Say that ed's game sees the shot moved from the free-throw line to the 3-point line. This unequivocally makes the shot harder and means less offense.
However, this change does not make offense less important; because it's relatively easy to find guys who can make free throws as compared to 3-pointers, the standard deviation of offense should go up, right?
This is relatively equivalent to what the NBA has done. The new rules interpretations favor offense over defense, but that doesn't necessarily mean they favor offensive teams over defensive ones.
In the non-theoretical realm, I'd be surprised if you'd find evidence that defense has become less important in the NBA since the new rules interpretations.
I think the problem is that "defense" as typically used in these discussion has more to do with the evil that Pat Riley begat with his '90s Knicks/Heat teams.
When most pundits (who aren't SABREmatic geniuses at the best of times. Except Joe Morgan. SuperGenius, just ask him) say that "defense is less important" they mean it's less useful to go out and get players whose primary skill is the ability to foul continuously and/or convince better players not to be aggressive on offense out of fear of being injured. By all that I mean Bruce Bowen. Bruce Bowen is less important under the new rules, and that is an empirically good thing.
You use 2 finals matchups as examples, a rather small sample set to begin with.
I wasn't using that sample as my evidence -- Bucher and Aldridge, who I quote, were. I was pointing out that their examples don't actually support their conclusion.
Extending ed johnson's point, it is possible for one of offense or defense to have greater variability than the other.
True. In practice, this isn't the case in any clear way.
What's more, insofar as we want to gloss "the variability of offense got higher" as meaning "offense has become more important," rule changes that make scoring easier should reduce the varaibility of NBA offenses.
"rule changes that make scoring easier should reduce the varaibility of NBA offenses."
That isn't clear at all---the effect could go either way. This is all an empirical question.
What the new rules change is not the importance of defense, but the importance of defensive players. Under the new rules, team defense is more important than individual defense. It used to be that you could throw a lockdown guy, like Bowen, on an offensive threat and he might be able to neutralize him, or at least slow him down. That's not how it works anymore. Now you need your whole team to play defense together. Often that's done by throwing erratic double teams and switching from man to man to zone. Defensive balance and teamwork is more valuable than star defensive players because star defenders can no longer do it on their own. That's what makes Ben Wallace less valuable.
If you built a team to succeed using Pat Riley type defense, you'd expect to have very little offensive benefits to the new rules since you have a bunch of mediocre offensive players. A team with good offensive players can take advantage of the rule change more. Hence variability should increase as the average score also goes up.
"The SD for offense and defense has to be the same, because every possession is both offense and defense, right?"
Imagine 3 teams (A,B,C) with offensive efficiency of 90,100, and 110.
If they each play 1 game against each other team their ratings will look like:
Team A: Offense 90, defense 105 (record 0-2)
Team B: Offense 100, defense 100 (record 1-1)
Team C: Offense 110, defense 95 (record 2-0)
The variability of offense would be twice that of defense.
But, as MY said, empirically this is not the case in the NBA.
Comments closed November 28, 2006.
| 
Heads in the Sand
Slate still points to your TPMCafe site.
Posted by ogged | November 14, 2006 3:18 PM