Mavs 108, Rockets 133

[KillerLeft]

"Bad" is very subjective here and depends on what people may mean by that. It is very accurate to say this though: 

"THJ is a negative defender whose presence consistently hurts his team's defense."

Defensive On/Off...
20-21: -1.7
19-20: -0.6
18-19: -2.4 w/ DAL, -6.6 w/ NYK
17-18: -1.7

He is a consistent liability on the defensive end.

Just curious, what are these numbers (for 20-21 only) for Richardson? I know it's a limited sample size, but I'm curious.

[Kammrath]

Luka improved his defense and his on/off impact is worse than ever.

I actually think Luka's defense has not been improved this year other than those couple games where he really put in the effort (as of late in particular guys have just been going around him super easy and getting in the paint). I am not surprised to see his defensive impact still a big negative. I don't think that is a statistical fluke. But of course 16 games of data with crazy lineup inconsistency is going to be noisy.

[KillerLeft]

But of course 16 games of data with crazy lineup inconsistency is going to be noisy.

Man, that's the truth. So frustrating.

[omahen]

DFS would still be the second biggest guy on the floor, for sure

Luka is actually same height but a bit heavier. I would say it is a bit more difficult to push him around.

[Kammrath]

Just curious, what are these numbers (for 20-21 only) for Richardson?

+1.2 on/off defensively in 260 on court minutes.

[KillerLeft]

So @"Kammrath", can you help me understand THIS model, then?

https://twitter.com/NBA_Math/status/1351...36769?s=20

[Kammrath]

This stat shows a lot why Mavs are 8-8. Mavs are eleventh in number of open three pointers shot in the league. Yet they are second to last in accuracy of this shots converting less than 35 % of them. They are 5th in open shots but convert just 33 % of them (18th in the league). In both cases some 10 percentage points behind league leaders. Mavs are creating well but not executing. If they improve that part, they will be great.

The big question:

Is this a talent/personnel issue?

OR 

Is this an anomaly that will correct itself in time?

[omahen]

I actually think Luka's defense has not been improved this year other than those couple games where he really put in the effort

I think it is very simple. Luka can play defense but he can't play it, because he is too tired of doing too much on the offensive end. When Mavs bring in another guy able to take part of the offensive load, Luka will have more energy to play defense. But with this tempo of games and the amount of stuff he is asked to do, don't expect him to be able to play it.

[dirkfansince1998]

I actually think Luka's defense has not been improved this year other than those couple games where he really put in the effort (as of late in particular guys have just been going around him super easy and getting in the paint). I am not surprised to see his defensive impact still a big negative. I don't think that is a statistical fluke. But of course 16 games of data with crazy lineup inconsistency is going to be noisy.

The fluke is that some of our guys played a lot of minutes with mediocre teammates that wouldn´t see the floor in most cases. When everyone was still healthy Richardson had a -10 on/off. While he was out the team continued to play and the "off" number changed. Not like Johnson or Iwundu have a better on/off number than Doncic. They are just as bad on defense. Even WCS is only at +1.1 on defense at this point.

[omahen]

The big question:

Is this a talent/personnel issue?

OR 

Is this an anomaly that will correct itself in time?

I think it is anomaly. We need to see 15 games of our full lineup to know with higher certainty. As long as we have constant lineups with complete non shooters, this will be bad.

[Kammrath]

So @Kammrath, can you help me understand THIS model, then?

So TPA (Total Points Added) model bases itself on BPM (Box Plus Minus) numbers. Those numbers are derived from boxscore contributions like rebounds, steals, blocks and does not take into account the +/- and on/off numbers.

I think the offensive side of those numbers are helpful, I tend to only like +/- and on/off numbers for defense (since I think defense cannot really be quantified in the boxscore for individual players).

[Kammrath]

The fluke is that some of our guys played a lot of minutes with mediocre teammates that wouldn´t see the floor in most cases. When everyone was still healthy Richardson had a -10 on/off. While he was out the team continued to play and the "off" number changed. Not like Johnson or Iwundu have a better on/off number than Doncic. They are just as bad on defense. Even WCS is only at +1.1 on defense at this point.

That is a fair point, the data is a mess for the Mavs right now because of so many variables and inconsistencies with lineups.

[KillerLeft]

Those numbers are derived from boxscore contributions like rebounds, steals, blocks and does not take into account the +/- and on/off numbers.

I think the offensive side of those numbers are helpful, I tend to only like +/- and on/off numbers for defense (since I think defense cannot really be quantified in the boxscore).

So just let me see if I understand this point. You're saying that these defensive rankings are sort of like ranking prime Deon Sanders low on a list of NFL corners because he doesn't have a lot of pass deflections or interceptions, only he should be much more highly ranked because he's so good the quarterback almost never targets the receiver he's covering?

[dirkfansince1998]

So @"Kammrath", can you help me understand THIS model, then?

https://twitter.com/NBA_Math/status/1351...36769?s=20

BPM (per 100) adjusted for possessions. Luka gets a lot of rebounds and that leads to a high DBPM. Looks really fancy and can tell us something about the overall production of the mentioned players. Not really helpful to evaluate non boxscore based things.

[Kammrath]

So just let me see if I understand this point. You're saying that these defensive rankings are sort of like ranking prime Deon Sanders low on a list of NFL corners because he doesn't have a lot of pass deflections or interceptions, only he should be much more highly ranked because he's so good the quarterback almost never targets the receiver he's covering?

EXACTLY. Guys who get lots of rebounds, steals, and blocks MAY be good at defense, but those numbers do not capture the whole picture of what defense is in the NBA. 

Great analogy.

[KillerLeft]

Lol, so basically, that's a fancy ass way of rewarding people for stuffing the box score...from an analytics community that constantly tells us we need to look past the box score???

Interesting.

[Kammrath]

BPM (per 100) adjusted for possessions. Luka gets a lot of rebounds and that leads to a high DBPM. Looks really fancy and can tell us something about the overall production of the mentioned players. Not really helpful to evaluate non boxscore based things.

Yep. Luka's rebounds ARE helpful to the team, but they don't actually tell you how effective he is in impacting team defense positively. They tell you something, but not all you need to know.

[Kammrath]

so basically, that's a fancy ass way of rewarding people for stuffing the box score...from an analytics community that constantly tells us we need to look past the box score???

I like NBA Math's numbers better for teams rather than individuals. I prefer individual numbers to incorporate on/off numbers at least in some form. But overall the reality is the numbers all tell us SOMETHING but it can be really, really hard to interpret them in some circumstances. There is no single "holy grail" number, the more numbers you can look at from different angles the better.

[DanSchwartzgan]

I have a question for @"DanSchwartzman":

You have consistently made a compelling argument that illuminates the positives Powell brings to the table, and I honestly agree with 99% of it. 

When you imagine Powell through the frame of what Burkemede and I are discussing, where does he fall? Like, forget synergy for a second and just rank the Mavs' players in a way that makes sense to you in a vacuum. Where is he on that list for you?

Don’t you guys know there are some good football games on?  Not meaning to shrink from the question, but in the current season context matters and we don’t have enough context.  

For instance, a year ago I would have ranked Powell ahead of Maxi (the team did too in deciding what to pay and for how long...Maxi’s 2022 in non-guaranteed).  This season, I think the answer is Maxi.  But Powell has spent much more time on a defensive island than Maxi has.   So, the comparison isn’t apples to apples.  With that said, Maxi based lineups have had dramatically more offensive success than Powell based lineups.  Dare I say Maxi is putting up a “Powell-like” O-Rating so far this season.  What we have no idea about is what Powell will look like with KP next to him and whether we will ever see .650 TS% Powell again.

Another way context (which we don’t have) matters is who plays with who.  This is a team largely made up of specialists.  When forced to do more than they should, things don’t go well.  Right now we have a lot of people doing more than they should.  It is pretty clear (at least to me) that THJ is better than DFS or JRich.  But, when dealing with specialists, what is meant by “better”.  DFS and JRich provide needed D.  THJ provides needed spacing and scoring.  The team is kind of a fragile eco-system and getting to play your best role is really important.  Put another way, when you have multiple defensive specialists, someone who can provide “O” is important.  Same thing in reverse if you have a bunch of “O” oriented specialists in your lineup.

I liked what Kamm wrote about who he would try in terms of big man pairings.  The only thing I would add is if you are going to start Powell, THJ probably needs to start also.  If you start Maxi, you can better afford a non-shooter in the starting lineup.  But, it is simply too early to tell what level Powell will return to and whether KP will make him look better than he looked w/o KP.  I think (or thought before the last two disastrous WCS games) that we’d have one shooting big and one rim running big together until the final few minutes when we would go to KP/Maxi to close.  I’ve always said 5-out with KP at center is probably best used as a gimmick lineup...probably in the w/o Luka minutes in the early 2nd/4th quarters.  

Finally, watch out for Brunson.  He’s got it going on right now and threatens to join the THJ/DFS/JRich/Maxi tier behind Luka and KP.  I’ve got Powell in a TBD slot after that.  I’m kind of resigned to the idea WCS will have nights where he’s in our top 5 players and games where he’s in the bottom five.

[KillerLeft]

For instance, a year ago I would have ranked Powell ahead of Maxi (the team did too in deciding what to pay and for how long...Maxi’s 2022 in non-guaranteed).  This season, I think the answer is Maxi.  But Powell has spent much more time on a defensive island than Maxi has.   So, the comparison isn’t apples to apples.  With that said, Maxi based lineups have had dramatically more offensive success than Powell based lineups.  Dare I say Maxi is putting up a “Powell-like” O-Rating so far this season.  What we have no idea about is what Powell will look like with KP next to him and whether we will ever see .650 TS% Powell again.

Another way context (which we don’t have) matters is who plays with who.  This is a team largely made up of specialists.  When forced to do more than they should, things don’t go well.  Right now we have a lot of people doing more than they should.  It is pretty clear (at least to me) that THJ is better than DFS or JRich.  But, when dealing with specialists, what is meant by “better”.  DFS and JRich provide needed D.  THJ provides needed spacing and scoring.  The team is kind of a fragile eco-system and getting to play your best role is really important.  Put another way, when you have multiple defensive specialists, someone who can provide “O” is important.  Same thing in reverse if you have a bunch of “O” oriented specialists in your lineup.

I liked what Kamm wrote about who he would try in terms of big man pairings.  The only thing I would add is if you are going to start Powell, THJ probably needs to start also.  If you start Maxi, you can better afford a non-shooter in the starting lineup.  But, it is simply too early to tell what level Powell will return to and whether KP will make him look better than he looked w/o KP.  I think (or thought before the last two disastrous WCS games) that we’d have one shooting big and one rim running big together until the final few minutes when we would go to KP/Maxi to close.  I’ve always said 5-out with KP at center is probably best used as a gimmick lineup...probably in the w/o Luka minutes in the early 2nd/4th quarters.  

Finally, watch out for Brunson.  He’s got it going on right now and threatens to join the THJ/DFS/JRich/Maxi tier behind Luka and KP.  I’ve got Powell in a TBD slot after that.  I’m kind of resigned to the idea WCS will have nights where he’s in our top 5 players and games where he’s in the bottom five.

Well, you didn't follow the instructions laid out in the assignment, but I'll give you an "A" anyway. Good stuff.