Tuesday, August 31, 2010
How does defense win championships?
By Henry Abbott
A few days ago I blogged about defense winning championships. Then I blogged about it again. Now that I'm getting a heck of a lot of e-mail, phone calls and comments about it, I can't stop!
To recap: We're not really just talking about the old adage here. Instead, we're talking about newly unearthed numbers showing that being an elite defensive team has, historically, make it more likely that you'd win a title than a team that was similarly good at offense. Neil Paine at Basketball-Reference looked at a half-century of title-winning teams. They're pretty much all good at offense and defense. However, he noticed that small improvements to defense increased a team's chances of winning titles a lot more than the same improvement to offense.
In general, especially when you measure by possession, the game is half offense and half defense. Right? Makes sense. Possessions alternate just about all game -- you get a chance to score, and then its our turn. All game long you're playing offense or defense, in alternating stints. They ought to both be important.
But when it comes to winning titles, the numbers suggest what coaches have been harping on since the dawn of timeouts -- the defensive part matters a little more than the offensive part.
I have already offered two quick theories as to why:
- Maybe defensive efficiency is a marker for which teams have good coaching and team cohesion. In other words, of course the defense itself helps, but maybe showing you can play good team defense is also a way of showing that you're a good team, a team that's prepared to hang together when the going gets tough, and maybe that's really what's winning those titles.
- My first guess was that offense is more volatile than defense, because it's based on shooting, which is an iffy proposition for even the best shooters in the world. What that means to me is that if your team's main thing is being able to score like crazy, and that's how you intend to win, you will still have games -- every team does -- where the ball does not go in the hoop as much as normal. Those games will be losses for you against good opponents.
But now there are many more theories to consider:
David Thorpe has been reading all this, and loves the idea that good defense may be a marker for team cohesion. However, he also thinks we have been missing a major point, which is that good defense leads to good offense, and the opposite is less true.
One of the ways this is true is that playing really active team defense is a way to make the referee blow the whistle in your favor. Illegal screens, over-the-back, frustration fouls, even technicals -- intense defense drives all of these things, which can in turn help your team with free throws. (It occurs to me those things can also, of course, get opposing players in foul trouble.)
But the most obvious way a good defense can help your offense is in creating live turnovers. A few years ago, Sandy Weil and John Huizinga, as part of their hot hand research, found something pretty important. As I wrote after they presented their findings at MIT: "How you get the ball matters tremendously to how likely you are to score with it. If you get the ball in a live ball turnover (a steal for instance) your shooting percentage is 12% higher than normal. If you get the ball from an offensive rebound, it's 8% higher. Home teams, meanwhile, shoot 1.5% better -- so you can see these advantages are major."
"I don't care what kind of offense you run," says Thorpe, "it works better if you start it by forcing turnovers."
Meanwhile, coaches often say that offense can help defense, because you can set up your defense more easily as opponents inbound after your made shot.
Thorpe also believes that teams with a reputation for playing good defense have multi-faceted psychological advantages. For instance, playing against the Celtics, Kobe Bryant or LeBron James might tweak his normal shot selection, knowing that there won't be any easy opportunities. That may earn the Celtics some key stops -- thanks to poor shot selection by opponents -- before they even start playing defense.
Paine's numbers showed that elite defensive teams won more championships than you'd expect. And I think what we're finding is that if you're one of those really great defensive teams we might also know that you're likely to:
- Be well coached.
- Have good teamwork.
- Get more easy offensive possessions than an average team.
- Draw more offensive fouls and violations.
- Have certain psychological advantages.
Some other e-mailers have additional thoughts:
- One of the best offensive performances in NBA playoff history was Michael Jordan's 63 points against the Celtics. But it is interesting to note how even though the Celtics could do nothing at all to stop the most potent offensive force in the game, they still played good team defense generally (and offense) and marched on to a title while the Bulls were swept in the first round.
- The numbers may be a bit skewed, because good teams allow few points and score a lot. That means a point or two of improvement in defensive efficiency is a bigger deal than the same improvement in offense. On Basketball-Reference.com, commenter Jax points out: "Allowing 10 fewer points on defense is a lot harder to do than adding 10 points on offense. The percentages are also different. 80 -> 70 is a 14.2% change. 80 -> 90 is a 11.1% change."
- A few other e-mailers pointed out that in the playoffs referees have the reputation of allowing more physical play, which favors aggressive defenses more than aggressive offenses. (This is suspect, however, as a way to explain titles. Almost every year, teams shoot more free throws in the playoffs than in the regular season. This is a topic for another post ...)
- Wayne Winston, the business professor who supplied the Mavericks with statistical insight for nearly a decade, e-mails: "I think you are right on team concept. In playoffs teams will always find your weakness. A chain is as strong as its weakest link so a weakness on defense will usually be more important than an offensive weakness. ... You have to win four series to win the title and any serious weakness is bound to show up."