Why I don't like individual Defensive rating

[deanvogs]

Yes, there will be lots of Math.

Instead of boring everyone in different threads, lets consolidate this. Now I am a math guy, I love stats, they mean something. Defensive rating just isn't a good stat though, and I'll show you why.

DRate is a formula, so we can calculate things from it. This all come from the Sports reference site.

Individual Defensive rating is calculated like this:

Drtg = Team Defensive Rating + .2 (100 * D Pts per ScPoss * (1 - Stop%) - Team Defensive Rating)

Now since we know that Iowa's Team Defensive rating per the Sports reference site is 103.5, we can back calculate some things from Jordan's 108.7 individual Drate, we can find X where X equals Dpts per ScPoss * (1-Stop%)

103.5 + .2 (100 * X - 103.5) = 108.7
So
x = 1.295

Lets see what happens if we put Jordan on Say California team, and calculated his Drate.

Since Cal team D is 92.2 lets plug that in and see what we get:

Jordan now has a Drate of 97.6

WOW, but how can that be? How can a player with a Drate of 108.7 have a Drate of 97.6 just by going to a different team?

Well it is easy math actually. When a formula relies so much on team D, that is going to make huge swings in an individuals rating, the team component that is. So it is very unfair to say that an individual's D is good or bad, when you tie into Team D so heavily in factoring an individuals D.

Lets see how Charlie Moore would fare if we put him on Iowa's defense. Right now Charlies defensive rating is 96.8 and the Cal team D rating is 92.2, so lets find X

92.2 + .2(100 * X - 92.2) = 96.8
So
X = 1.152

Lets put Charlie's 1.152 factor on Iowa's team.

103.5 + .2 (100 * 1.152 - 103.5) = 105.84

Now the math would actually get way more complicated, and more in depth than I'm willing to go. Part of the formula to find Dpts per ScPoss * (1-Stop%) includes yet again even more team defensive stats. I'm not willing to go all the way down that rabbit hole, and frankly that math is above my pay grade. Yet from looking at the formulas, putting Jordan on a good defensive team might even make his ranking even better when you factor in yet again more good team numbers that way, and vice versa taking a player out of a good team D situation and putting them on a bad defensive team, his numbers are going to look even worse yet factors this in.

 

[deanvogs]

The thing is the above portion just touches on the Team aspect of this "individual" rating. The other troubling part of the stat is how they calculate "stops".

Stops are awarded for a Steal or block or defensive rebound. I don't see any component to it that rewards a player for playing good position D, playing good help D, for staying in front of his man, for denying his man the ball, for basically a hundred important defensive things that happen on a court. Defense is SOOOO much more than just steals, blocks and rebounds.

56, I gotta say this has nothing to do with Jordan, or other 4* PG that you salivate over. It is simply a math thing for me, there are WAYYYYYYYYY to many holes in the formula for me to take it seriously as a stat.

 

[PCHawk]

There is the argument that Jordan going to a better team would bring down their team defensive rating. But there is also the argument that a point guard rarely gets defensive rebounds. So when you say he has the worst defense on the team, a lot of that is because he doesn't get rebounds. That's pretty dumb.

 

[deanvogs]

There is the argument that Jordan going to a better team would bring down their team defensive rating. But there is also the argument that a point guard rarely gets defensive rebounds. So when you say he has the worst defense on the team, a lot of that is because he doesn't get rebounds. That's pretty dumb.

There is some of that. But if you look at Cal for instance, Charlie Moore is the worse defender on his team as well. Much of Charlie's stats look better, because the team plays good team D, and that helps his individual rating more than he hurts their team rating.

 

[PCHawk]

There is some of that. But if you look at Cal for instance, Charlie Moore is the worse defender on his team as well. Much of Charlie's stats look better, because the team plays good team D, and that helps his individual rating more than he hurts their team rating.

But is he really the worst defender? Or does the stat just show he is because he is the point guard so he doesn't get any rebounds? Like you said, the stat sucks.

 

[deanvogs]

But is he really the worst defender? Or does the stat just show he is because he is the point guard so he doesn't get any rebounds? Like you said, the stat sucks.

Yep, and as I really broke the stat down and looked at it, everything flat out admitted that Big men were always going to rate better. They simply wrote that off and said that well of course it only makes sense and guys in the middle make better defenders. That is complete bullshit though, if perimeter defenders stopped penetration, then the big men have an easier time, and become better defenders as well. Heck here is the disclaimer on their own site after describing the formula for Drate (I have bolded the very telling admission).

• Out of necessity (owing to a lack of defensive data in the basic boxscore), individual Defensive Ratings are heavily influenced by the team's defensive efficiency. They assume that all teammates are equally good (per minute) at forcing non-steal turnovers and non-block misses, as well as assuming that all teammates face the same number of total possessions per minute.
• Perhaps as a byproduct, big men tend to have the best Defensive Ratings (although Oliver notes that history's best defensive teams were generally anchored by dominant defensive big men, suggesting that those types of players are the most important to a team's defensive success). A corollary to this is that excellent perimeter defenders who don't steal the ball a lot — for instance, Joe Dumars or Doug Christie — are underrated defensively by DRtg, and are prone to look only as good as their team's overall defense performs.

I also just couldn't believe just how much of a team aspect it had to it. Simply put it is impossible for Charlie Moore to have an above 100 Drate while playing for Cal, almost no matter how bad he is. At the same time if you put him on Iowa, there is almost no way possible he could have a sub 100 Drate while on Iowa's team.

 

[PCHawk]

Yep, and as I really broke the stat down and looked at it, everything flat out admitted that Big men were always going to rate better. They simply wrote that off and said that well of course it only makes sense and guys in the middle make better defenders. That is complete bullshit though, if perimeter defenders stopped penetration, then the big men have an easier time, and become better defenders as well. Heck here is the disclaimer on their own site after describing the formula for Drate (I have bolded the very telling admission).

• Out of necessity (owing to a lack of defensive data in the basic boxscore), individual Defensive Ratings are heavily influenced by the team's defensive efficiency. They assume that all teammates are equally good (per minute) at forcing non-steal turnovers and non-block misses, as well as assuming that all teammates face the same number of total possessions per minute.
• Perhaps as a byproduct, big men tend to have the best Defensive Ratings (although Oliver notes that history's best defensive teams were generally anchored by dominant defensive big men, suggesting that those types of players are the most important to a team's defensive success). A corollary to this is that excellent perimeter defenders who don't steal the ball a lot — for instance, Joe Dumars or Doug Christie — are underrated defensively by DRtg, and are prone to look only as good as their team's overall defense performs.

I also just couldn't believe just how much of a team aspect it had to it. Simply put it is impossible for Charlie Moore to have an above 100 Drate while playing for Cal, almost no matter how bad he is. At the same time if you put him on Iowa, there is almost no way possible he could have a sub 100 Drate while on Iowa's team.

And PER is influenced greatly by that stat. Lol

 

[BryceC]

Similar to defensive metrics in baseball. A little spotty imo.

 

[Robowe]

The thing is the above portion just touches on the Team aspect of this "individual" rating. The other troubling part of the stat is how they calculate "stops".

Stops are awarded for a Steal or block or defensive rebound. I don't see any component to it that rewards a player for playing good position D, playing good help D, for staying in front of his man, for denying his man the ball, for basically a hundred important defensive things that happen on a court. Defense is SOOOO much more than just steals, blocks and rebounds.

56, I gotta say this has nothing to do with Jordan, or other 4* PG that you salivate over. It is simply a math thing for me, there are WAYYYYYYYYY to many holes in the formula for me to take it seriously as a stat.

Not to mention teams playing almost exclusively zone defense - like Syracuse. Defensive ratings have to calculate stops as related to zone defense (who was responsible when a shot was made). It's not always the closest defender being slow on a close out. Sometimes, a defender over commits (helps) and leaves a place in the defense where another defender has to then cover up leaving an offensive player open to make a shot. As I have asked before, who is really responsible? It's not always easy to determine.