Bucks Diary

Wednesday, March 05, 2008

A new way to measure individual defensive performance in basketball


Alright Bucks fans, I think I have developed a way to tell the full story about the debacle that is the 2007-08 Milwaukee Bucks... on both sides of the court. At last, with my easy new defensive metric individual responsibility can be meted out where it belongs and the weak links can be exposed. No more hiding behind the cloak of "collective responsibility."

Introducing "Counterpart Win Score" to measure individual defensive performance

I wasn't satisfied with the "Defensive Rating" metric used by basketball-reference.com to grade the defensive performance of each player. It seemed to be too dependent on the performance of the entire team.

So instead I referenced the "counterpart" data provided by 82games.com (a "counterpart" is the guy you are guarding) and did a "Position Adjusted Win Score" analysis on each Bucks collective counterparts. Sort of a Win Score in reverse.

Here's my thinking. If "Win Score" is a formula that accurately measures the way any given player's statistical output translates into victories... and I am convinced it is... then I figure the degree to which each Buck prevents or allows his counterparts from accumulating said statistics provides a good indication of the degree to which that Buck has prevented his counterparts from contributing to victories for their teams. And isn't that the definition of defense?

The Data


When you look at my chart, here is what is going on: the first column is the player's "Position Adjusted Win Score / 48", what I will call his offensive contribution to victory. The second column is the cumulative "Position Adjusted Win Score / 48" of every guy that Buck has guarded this season. The third column is the player's "True Win Contribution" which is the sum of those two numbers divided by 2 and then multiplied by the player's percentage of overall playing time. (Note: Since I'm doing this entirely from the Bucks perspective, all "+" numbers are good. Also, keep in mind that a Win Contribution of .000 is exactly average, and a Team total of .000 will indicate a .500 team, so any contribution of a positive number indicates the player has contributed to a winning season, and vice versa.)

Discussion of results

1. Pressey25 was right!! The Bucks backcourt, especially Mo Williams, is awful defensively. Mo has wiped away his good work on offense with his poor work on defense.

2. I was stunned to see that Yi Jianlian is a good defender. Remember, "defense" under my definition is more than just "checking your man". Its keeping him off the boards, its getting him into foul trouble, its forcing him into turnovers, its preventing him from accumulating assists, ... Of all the starters Yi the most to disrupt his counterparts from making a positive contribution. I feel much better about him now. I still think his best spot is SF because of his offensive tendencies.

3. I was disappointed but not shocked to see that Andrew Bogut is not that effective defensively. Granted, he sometimes has to leave his man to help, but those instances should have come out in the wash, so to speak, by now. I think he needs to do a better job of scoring prevention, and rebounding prevention.

4. Royal Ivey, you are exposed! Get out now!

5. Charlie V and Bobby Simmons are basically wastes of space.

6. Oh, I have to close by giving respect to Desmond Mason, the only Buck above average on both sides of the court. I was sure wrong about him.

7 Comments:

At March 5, 2008 at 4:18 PM, Blogger frank said...

Very cool, Ty.

 
At March 6, 2008 at 9:27 AM, Anonymous J Rock said...

Great work! Hope you forwarded it to the WOW guys.

I'm curious as to what kind of numbers Brian Skinner and Ruben Patterson put up last year, in this formula? They seem to be the two players the team misses the most.

These stats also gives credence to my opinion that Charlie Bell should have been starting since the beginning of the year, for his defense at the point of attack. That way Williams gets to defend against a bench player (really don't know how much that would matter, though).

Bogut is somewhat surprising, but I guess he does tend to let the guy he is directly defending, do whatever. Bogut seems like a better weakside defender, than a man-to-man defender.

Kudos to Desmond for being a two way player.

I think with Des at the 3 and Yi at the 4, Yi should play more of the 3 role offensively and Des the 4. That makes Yi more of small forward on offense, but a power forward on defense.

Anyway, keep up the good work.

 
At March 6, 2008 at 2:18 PM, Anonymous DD said...

The problem for Bogut as the last line of defence is that he'll regularly be out of position trying to remedy problems caused by defensive breakdowns on the perimeter.
He doesn't get scored on so much when he gets to play his man straight up. However, he is forced to give up a ridiculous amount of bunnies and offensive rebounds when he steps off to help.
Embarrassing really, but at least this year he just keeps trying without complaining or sulking.

 
At March 8, 2008 at 1:26 AM, Anonymous paulpressey25 said...

This is some good stuff.

But I have to reconcile what I see during the games with Bogut's low reverse win score. Why does he score so low defensively?

If I understand your formula, you are taking the win scores of the guy he guarded. But I can't recall centers going off against him this season. In most cases he seems to contain guys pretty well. Or does this win score formula also account for rebounds?

 
At March 10, 2008 at 9:30 AM, Anonymous Anonymous said...

There are times when stats just make absolutely no sense and should be treated as such. Sorry Ty but this is one of those times, any stat that ranks Bogut below several bucks players defensively and has Yi as our best defender (??????) is clearly flawed beyond repair. Sometimes we've just got to trust our eyes as the stats clearly give a false impression.

 
At March 11, 2008 at 1:11 PM, Anonymous Glenn said...

"Sometimes we've just got to trust our eyes as the stats clearly give a false impression." Isn't that what has everyone at the Journal Sentinel and in the Bucks broadcasting booth gushing over Royal "7-10" Ivey? He sucks big time and has no stats to back up all of the wonderous things that are said about him. But hey, he looks like a spaz on court, so he must be a good defender. And I love Bogut, but has anyone noticed how tall guys seem to effortlessly sink their close range jumpers right in his face?

 
At March 11, 2008 at 5:03 PM, Blogger frank said...

The Bogut/Yi thing could very well be because Bogut has been defending the better post players while Yi has generally taken the more perimeter-oriented bigs, even if the PF/C designations are reversed (ie against Utah, Yi takes Okur and Bogut takes Boozer).

If the post guys are typically the better players then Bogut's score would likely be worse even if he's a better defender. I'm not sure how 82games does defensive assignments, but that would be one explanation.

 

Post a Comment

Links to this post:

Create a Link

<< Home