The Bratwurst - Milwaukee Bucks Blog

I am pleased to announce that I have made an upgrade to my IPM rankings — thanks to the geniuses at www.82games.com, they kept track of charges this season and so I was able to add them into the ranking. I’ve been waiting for this. The rankings don’t change much (Andrew Bogut gets a little boost), but I still feel it helps with accuracy a bit. The rankings are available here.

But the real improvement is that I have added a defensive component to IPM, making this ranking truly unique — not only does it include charges, but also defense. I’m mighty proud of it.

The idea for a defensive ranking came from this article by Jon Nichols at 82games.com. His method — called Defensive Composite Score — ranked players by three categories, summed the rankings, and then ranked them by percentage to come up with a 0-100 ranking. His categories were athleticism ratio (if you are familiar with my draft preview, I also used it there. It is blocks+steals/fouls), Dean Oliver’s Defensive Rating (a single number developed and described in the book Basketball on Paper) and on court/off court differential.

I took Nichols’ DCS, tweaked it and took it a step further. First, I added charges to the athleticism ratio (blocks+steals+offensive fouls drawn/fouls committed). After getting the individual ratings, I took my team defensive IPM score in order to add a “team adjustment” to each individual. Then I adjusted for how much each player was responsible for their teams’ defensive IPM, by backing out each player to determine how much better or worse each team would be had he not been on the team. Finally (and I wish I didn’t have to do this, but the data demanded it) I inputted a position adjustment to reflect the fact that big men have much more difficult demands on their defense.

I absolutely hate position adjustments, and feel that if you have to use them then there is a flaw in your methodology because you can then just change a players’ ranking by changing their position — even though they don’t always play the same position. It’s a major flaw that also crops up in the “Wages of Wins” school of thought, but that’s a whole other story.

Anyway, I then had an adjusted score that I could convert to IPM, and by combining IPM with the Defensive IPM I obtained a final, all-around ranking. Because the rules of basketball are slanted toward offense I weighted the combined rankings as 75% IPM and 25% defensive IPM. Basically, I figured that since a 25 ppg scorer would likely score 35 points against the worst defender but about 20 against the best defender, you can’t truly assume the two systems are equal. A great defensive player will still allow enough scoring that he has to at least be able to pick up a little slack at the offensive end.

The ranking are available here.

What does the defensive ranking component tell us?

The most interesting thing is that it says that in the case of top players, they usually bring the goods at both ends of the floor. It makes some sense — logically great players like LeBron, KG and Kobe are so superior athletically that they can dominate at both ends. So the players who don’t have their overall rankings changed very much when defensive IPM is included tend to be the best players (or, like Matt Carroll, they just suck at both offense and defense).

What was also interesting was looking at the players who had their rankings changed the most by adding defensive IPM. It clearly shows that there are some “winning”, “intangible” skills that some players bring and that some lack.

It doesn’t come as much surprise that the players who had their rankings improved the most were the known offensively-challenged defensive specialists: Shane Battier, Ben Wallace, and Anderson Varejao types. But what was interesting was that of the 40 players who had their rankings improved by the most (among those who averaged 20 mpg), 30 of them made the playoffs this season. At the other end of the spectrum, among the 40 20mpg+ offensive specialists who were hurt the most by incorporating defensive rankings (Eddy Curry, Ben Gordon, Hakim Warrick …. Charlie Villanueva) 30 of them missed the playoffs. Coincidence? I think not.

What does this ranking system tell us about some of the Bucks?

Charlie Villanueva must go. While Charlie V’s offensive IPM was decent (0.7756, #116 overall) his team-worst 0.508 defensive IPM dropped his overall ranking 58 spots to #176. Quite simply, Villanueva is a pretty average offensive player who doesn’t defend well enough to make him a net-positive player.

Michael Redd should stay. Redd suffered through an awful offensive season by his standards and had a pretty lousy defensive IPM to boot, but when it was all said and done his overall ranking dropped by 19 spots. However, Redd is good enough offensively to score 25 a night and get to the line 10 times, while attempting to replace him would likely lead to a replacement who is moderately better defensively but much, much worse offensively. And Redd is likely to be better on offense next season.

Mo Williams should go. Mo’s 0.625 defensive IPM lowered his overall ranking by 25 spots, which pretty much negates all of the strides his offensive game has made. Added to that is the way Mo’s decision making seems to make the game harder on his teammates, and he winds up looking like a real drag on the team.

Andrew Bogut should sign that extension. When charges were added, Bogut’s IPM rose to 0.910, which is borderline-all-star level. At his age, there is still plenty of time for Bogut to mature into a solid contributor for a good team. Bogut’s defense was also the best on the team, which suggests that once Scott Skiles beats some defensive responsibility into his teammates, he should be able to become the centerpiece of a pretty good defense.

Yi was … good defensively? I don’t know about this one. Yi’s defensive IPM was pretty good, but I think that Yi gets a big assist from Bogut here. Bogut would always guard the oppositions’ best big, and Yi’s on/off court differential looks a little better than it is because he got hurt a little before the team really cratered. Yi has a lot of work to do next year — and he owes it to Bogut for making his numbers look good this season.

Thank God Royal Ivey is gone. Having a defensive specialist can be a good strategy. However, Ivey was so bad offensively that even great defense couldn’t make him a net-positive player. But Ivey, despite maybe looking good here and there, was lousy defensively as well — his 0.523 DefIPM was the second worst on the team. Charlie Bell’s DefIPM was awful as well, but he gets a pass because he was thrown to the wolves at small forward way too often. How can a 6′3″ player be expected to guard LeBron James straight up?

Ramon Sessions isn’t as good as the numbers say. While Sessions had the best offensive and defensive IPM on the team, don’t forget that his 17 games played amounted to 450 minutes of extended garbage time. Nobody was very serious about guarding him and there wasn’t much incentive for him to play much defense beyond gambling for steals. He’ll be a good backup, but he’s not nearly as good as the numbers think.

And there’s my two cents for John Hammond.