OFFICIAL BRANDON INGRAM THREAD
Goto page Previous  1, 2, 3 ... 310, 311, 312 ... 1883, 1884, 1885  Next
 
Post new topic    LakersGround.net Forum Index -> General Basketball Discussion Reply to topic
View previous topic :: View next topic  
Author Message
jonnybravo
Retired Number
Retired Number


Joined: 21 Sep 2007
Posts: 30621

PostPosted: Sun Jan 22, 2017 6:22 pm    Post subject:

tox wrote:
jonnybravo wrote:
How about Drapm?
IMO it's the only good all-in-one defensive stat, but even then you need to supplement it with your eyes.

Its biggest is issue is that it's really hard to come by. The guy who usually publishes the numbers, Engelmann (the creator of RPM & RAPM), hasn't done so in '16-'17 despite having the numbers available (he tweets some info about it). I suspect ESPN won't let him, but that's of course just speculation.


Yeah I remember you mentioning it being a better stat but had no luck either trying to find the #'s for it.
_________________
KOBE
Back to top
View user's profile Send private message Reply with quote
dao
Star Player
Star Player


Joined: 02 Jan 2013
Posts: 5572

PostPosted: Sun Jan 22, 2017 6:26 pm    Post subject:

tox wrote:
jonnybravo wrote:
tox wrote:
dao wrote:
tox wrote:

A really big part of DBPM is just this "rebound x assist" term: if you get a lot of rebounds and a lot of assists, it'll boost your DBPM score a lot. This is why DBPM thinks Westbrook is such a good defender when he's blatantly not, and it's also a big reason LeBron's DBPM spikes to a ridiculous amount in the playoffs (he gets a (bleep) ton of assists, and he starts getting more uncontested rebounds so that he can start a one man break... obviously he plays better defense and gets more blocks & steals as well).

I guess this explains why Randle has a DBPM of +1.0, which is indicative of a solid defensive player despite his noticeably poor defense. Why on earth would they give heavy weight to assists and rebounds...well defensive rebounds makes sense. But yeah, I always wondered why Randle's DBPM was so good. Even last year it was 0.0 despite a terrible defensive season from him.

Bingo.

And this is why DBPM sucks. Think about dudes like LeBron, Draymond, KG. What do they have in common? Lots of boards, lot of assists. And excellent defense (well YMMV with LeBron since it varies based on season).

What happens is there just aren't enough box score stats about defense, so when they tried to fit RAPM onto box score stats (this is exactly how BPM was created btw), then the defensive portion got pretty (bleep) up. They found mostly arbitrary correlations (like players who played good defense also tended to get a good amount of rebounds x assists).

Side note: DRPM also has problems with (bleep) like this. Did you know the taller you are, the better DRPM thinks of you? That's why Randle can be the 60th best defensive power forward and still have a DRPM of +0.22. But it's more obvious that he's not good defensively in RPM because you can compare him to other PFs.


Wait, why does it even factor in assists? I guess what the rationale is in including assists?


It's what levon said.

Basically, there's this stat RAPM (regularized adjusted plus/minus) that tells you how much better lineups get when some given player is in that lineup (vs. a league average player). It does this using a statistics called regularized regression.

But to get updated values, you have to continually re-run the regression process which takes time with a (bleep) ton of data. So instead, basketball reference tried to find the best box score approximation for RAPM, and they called it BPM (box [score] plus minus).

In the process of finding a box score approximation of RAPM, they found this rebound x assist term works very well. Here's a quote from basketball reference:

Quote:
Finally, a positive interaction term between rebounding and assists is included. This can be interpreted a number of ways – athleticism interacting with basketball awareness, size interacting with basketball skills, etc. This term was highly significant, and helped the overall fit of the regression quite a bit. Using the square root of the interaction both makes sense theoretically (maintaining a denominator of opportunities or possessions) and empirically (it is more significant and helps the overall fit of the regression more).


So that's how they got BPM, and that's why there's this rebound x assist term.

The reason why it's part of DBPM is circumstantial. They reran the BPM process but on offensive RAPM only, and the best fit they called "OBPM." Because there aren't enough numbers for defensive stats, though, they just made DBPM = BPM - OBPM. As it just so happened, DBPM ended up including a heavy rebounding x assist interaction term.
wow, can't really argue with their reasoning. Players with the physical tools to rebound and the IQ to get high assist numbers generally will be good defenders. A guy like Randle who gets assists despite low awareness will be an outlier, but for most players this is probably pretty predictive I imagine. But getting high assists requires more than just awareness, so guys like klay thompson who have high IQ but low playmaking skill get underrated defensively by the stat.
Back to top
View user's profile Send private message Reply with quote
tox
Franchise Player
Franchise Player


Joined: 16 Nov 2015
Posts: 17835

PostPosted: Sun Jan 22, 2017 6:28 pm    Post subject:

jonnybravo wrote:
tox wrote:
jonnybravo wrote:
How about Drapm?
IMO it's the only good all-in-one defensive stat, but even then you need to supplement it with your eyes.

Its biggest is issue is that it's really hard to come by. The guy who usually publishes the numbers, Engelmann (the creator of RPM & RAPM), hasn't done so in '16-'17 despite having the numbers available (he tweets some info about it). I suspect ESPN won't let him, but that's of course just speculation.


Yeah I remember you mentioning it being a better stat but had no luck either trying to find the #'s for it.


Yeah. If I could figure out his exact process, I would do it myself and post the results somewhere. The best probably is to use DRPM while knowing its weaknesses... but I'd take an informed eye test over DRPM any day. Defensive stats are just really limited.

I'm waiting to get more useful tracking stats... something as simple as "distance moved per defensive possession" on ball and off ball would be amazing. Or even "average time to closeout shooter on the weakside corner" or whatever.
Back to top
View user's profile Send private message Reply with quote
tox
Franchise Player
Franchise Player


Joined: 16 Nov 2015
Posts: 17835

PostPosted: Sun Jan 22, 2017 6:31 pm    Post subject:

dao wrote:
wow, can't really argue with their reasoning. Players with the physical tools to rebound and the IQ to get high assist numbers generally will be good defenders. A guy like Randle who gets assists despite low awareness will be an outlier, but for most players this is probably pretty predictive I imagine. But getting high assists requires more than just awareness, so guys like klay thompson who have high IQ but low playmaking skill get underrated defensively by the stat.


Yeah, the reasoning makes sense when you think about it.

But in practice it doesn't work all that well. Not just Randle, but we're finding guards rebound more so they can push the ball up more. Westbrook & Harden both have obscenely bloated DBPM numbers. And as you noted, plenty of defenders don't rebound much or don't get many assists.

And of course, it has a large dependence on steals and blocks, which helps people who gamble a lot (again like Westbrook).
Back to top
View user's profile Send private message Reply with quote
fiendishoc
Star Player
Star Player


Joined: 23 Jun 2005
Posts: 8488
Location: The (real) short corner

PostPosted: Sun Jan 22, 2017 6:31 pm    Post subject:

The box score prior for RPM is formulated similarly to BPM, so likely they share a lot of the same flaws.
Back to top
View user's profile Send private message Reply with quote
tox
Franchise Player
Franchise Player


Joined: 16 Nov 2015
Posts: 17835

PostPosted: Sun Jan 22, 2017 6:36 pm    Post subject:

fiendishoc wrote:
The box score prior for RPM is formulated similarly to BPM, so likely they share a lot of the same flaws.
At least the actual regression process can sift past some of that nonsense. Westbrook at least isn't a top defender, and Harden is way at the bottom by position.

Did ESPN ever release what the box score prior is exactly? I always thought they kept it as a black box.
Back to top
View user's profile Send private message Reply with quote
fiendishoc
Star Player
Star Player


Joined: 23 Jun 2005
Posts: 8488
Location: The (real) short corner

PostPosted: Sun Jan 22, 2017 6:43 pm    Post subject:

It's a black box, but I've heard it's a SPM process similar to BPM. Nylon Calculus showed that BPM predicted differences between RPM and RAPM quite well.
Back to top
View user's profile Send private message Reply with quote
fiendishoc
Star Player
Star Player


Joined: 23 Jun 2005
Posts: 8488
Location: The (real) short corner

PostPosted: Sun Jan 22, 2017 6:50 pm    Post subject:

I brought up the prior actually in response to the Ingram and Randle discussion, not the Westbrook and Harden one which I didn't see.
Back to top
View user's profile Send private message Reply with quote
GoldenThroat
Moderator
Moderator


Joined: 25 Jul 2005
Posts: 37474

PostPosted: Sun Jan 22, 2017 6:51 pm    Post subject:

Back to top
View user's profile Send private message Reply with quote
tox
Franchise Player
Franchise Player


Joined: 16 Nov 2015
Posts: 17835

PostPosted: Sun Jan 22, 2017 6:53 pm    Post subject:

fiendishoc wrote:
It's a black box, but I've heard it's a SPM process similar to BPM. Nylon Calculus showed that BPM predicted differences between RPM and RAPM quite well.
I just read that article. That is some seriously excellent stuff. Like legitimately excellent. Interesting that it's a "guest post"

Quote:
BPM is actually a better predictor than RAPM.



So it confirms my first thoughts: DRPM is the best we've got, since it has the regression part. But I underestimated how corrupted it was by box score stats. Man, I really hate BPM as a prior. You'll never be able to convince me this is an advancement and not a cheap way to get more palatable.

Thanks, fiendish.
Back to top
View user's profile Send private message Reply with quote
tox
Franchise Player
Franchise Player


Joined: 16 Nov 2015
Posts: 17835

PostPosted: Sun Jan 22, 2017 6:56 pm    Post subject:

fiendishoc wrote:
I brought up the prior actually in response to the Ingram and Randle discussion, not the Westbrook and Harden one which I didn't see.
Hmm, well in the case of both Ingram & Randle, DBPM and DRPM diverge a good amount, even given DBPM being a prior for DRPM.

DBPM likes both of them better. The Randle part doesn't surprise me, but the Ingram part does on first glance.
Back to top
View user's profile Send private message Reply with quote
fiendishoc
Star Player
Star Player


Joined: 23 Jun 2005
Posts: 8488
Location: The (real) short corner

PostPosted: Sun Jan 22, 2017 6:56 pm    Post subject:

When in doubt, I think it's still best to go by on/off and lineups. Keep the caveats where we can see them.
Back to top
View user's profile Send private message Reply with quote
tox
Franchise Player
Franchise Player


Joined: 16 Nov 2015
Posts: 17835

PostPosted: Sun Jan 22, 2017 7:00 pm    Post subject:

fiendishoc wrote:
When in doubt, I think it's still best to go by on/off and lineups. Keep the caveats where we can see them.
Without doubt. In general, I find it more useful to know, "Oh hey, so no matter what you think of Russell and Randle's defense, they can be part of a league average defense, i.e. the starters." Lineup data is great, since it really doesn't matter as much how good a player is in a vacuum.

But raw on/ off I don't really like. Those still have their own issues... like who the replacement is as well as opposition quality. Ideally you'd sift past some of that, which lands you right back at RAPM.
Back to top
View user's profile Send private message Reply with quote
fiendishoc
Star Player
Star Player


Joined: 23 Jun 2005
Posts: 8488
Location: The (real) short corner

PostPosted: Sun Jan 22, 2017 7:07 pm    Post subject:

tox wrote:
fiendishoc wrote:
I brought up the prior actually in response to the Ingram and Randle discussion, not the Westbrook and Harden one which I didn't see.
Hmm, well in the case of both Ingram & Randle, DBPM and DRPM diverge a good amount, even given DBPM being a prior for DRPM.

DBPM likes both of them better. The Randle part doesn't surprise me, but the Ingram part does on first glance.


Yeah it's not exactly the same as BPM, but I would guess that it shares the same flawed interactions and there's some other stuff in there like height as you've noted. I can't see the numbers at the moment- what is it about Ingram that's changed the most?
Back to top
View user's profile Send private message Reply with quote
anpherknee
Franchise Player
Franchise Player


Joined: 14 Mar 2014
Posts: 16933

PostPosted: Sun Jan 22, 2017 7:09 pm    Post subject:

GoldenThroat wrote:
https://media.tenor.co/images/fb3f2d1e814190100a4ae401b1660d5b/tenor.gif


it me
Back to top
View user's profile Send private message Reply with quote
fiendishoc
Star Player
Star Player


Joined: 23 Jun 2005
Posts: 8488
Location: The (real) short corner

PostPosted: Sun Jan 22, 2017 7:09 pm    Post subject:

tox wrote:
fiendishoc wrote:
When in doubt, I think it's still best to go by on/off and lineups. Keep the caveats where we can see them.
Without doubt. In general, I find it more useful to know, "Oh hey, so no matter what you think of Russell and Randle's defense, they can be part of a league average defense, i.e. the starters." Lineup data is great, since it really doesn't matter as much how good a player is in a vacuum.

But raw on/ off I don't really like. Those still have their own issues... like who the replacement is as well as opposition quality. Ideally you'd sift past some of that, which lands you right back at RAPM.


Yeah I'd rather do that in my head than worry about whether a tiny sample size is throwing off RAPM, even with the 0 prior.
Back to top
View user's profile Send private message Reply with quote
tox
Franchise Player
Franchise Player


Joined: 16 Nov 2015
Posts: 17835

PostPosted: Sun Jan 22, 2017 7:10 pm    Post subject:

fiendishoc wrote:
tox wrote:
fiendishoc wrote:
I brought up the prior actually in response to the Ingram and Randle discussion, not the Westbrook and Harden one which I didn't see.
Hmm, well in the case of both Ingram & Randle, DBPM and DRPM diverge a good amount, even given DBPM being a prior for DRPM.

DBPM likes both of them better. The Randle part doesn't surprise me, but the Ingram part does on first glance.


Yeah it's not exactly the same as BPM, but I would guess that it shares the same flawed interactions and there's some other stuff in there like height as you've noted. I can't see the numbers at the moment- what is it about Ingram that's changed the most?

Ingram's DBPM is okay. -0.9. Definitely not worst in the league. His DRPM is something like -2.8 (probably going down after tonight), bottom 7 in the league and worst on the Lakers (worse than Lou, Clarkson, Young, etc.). I know why RPM doesn't like him, even beyond his own limitations -- lineup data isn't in his favor. But it still surprises me a bit regardless.
Back to top
View user's profile Send private message Reply with quote
tox
Franchise Player
Franchise Player


Joined: 16 Nov 2015
Posts: 17835

PostPosted: Sun Jan 22, 2017 7:14 pm    Post subject:

fiendishoc wrote:
tox wrote:
fiendishoc wrote:
When in doubt, I think it's still best to go by on/off and lineups. Keep the caveats where we can see them.
Without doubt. In general, I find it more useful to know, "Oh hey, so no matter what you think of Russell and Randle's defense, they can be part of a league average defense, i.e. the starters." Lineup data is great, since it really doesn't matter as much how good a player is in a vacuum.

But raw on/ off I don't really like. Those still have their own issues... like who the replacement is as well as opposition quality. Ideally you'd sift past some of that, which lands you right back at RAPM.


Yeah I'd rather do that in my head than worry about whether a tiny sample size is throwing off RAPM, even with the 0 prior.


Mmm, fair point. You also see what's going on with your eyes much better than I can... I tend to use numbers as a sanity check. Besides you can often figure out what's throwing off RAPM (or RPM) and how much to qualify its measure.
Back to top
View user's profile Send private message Reply with quote
fiendishoc
Star Player
Star Player


Joined: 23 Jun 2005
Posts: 8488
Location: The (real) short corner

PostPosted: Sun Jan 22, 2017 7:16 pm    Post subject:

tox wrote:
fiendishoc wrote:
tox wrote:
fiendishoc wrote:
I brought up the prior actually in response to the Ingram and Randle discussion, not the Westbrook and Harden one which I didn't see.
Hmm, well in the case of both Ingram & Randle, DBPM and DRPM diverge a good amount, even given DBPM being a prior for DRPM.

DBPM likes both of them better. The Randle part doesn't surprise me, but the Ingram part does on first glance.


Yeah it's not exactly the same as BPM, but I would guess that it shares the same flawed interactions and there's some other stuff in there like height as you've noted. I can't see the numbers at the moment- what is it about Ingram that's changed the most?

Ingram's DBPM is okay. -0.9. Definitely not worst in the league. His DRPM is something like -2.8 (probably going down after tonight), bottom 7 in the league and worst on the Lakers (worse than Lou, Clarkson, Young, etc.). I know why RPM doesn't like him, even beyond his own limitations -- lineup data isn't in his favor. But it still surprises me a bit regardless.


Too bad Engelmann isn't posting the RAPM numbers anymore so we can't see just how much it's pulling him down. You have one of the most played lineups in the league with the starters doing pretty well and then a pretty decent sample of replacing Deng (who is no on off stud himself) with Ingram and it goes into the tank.
Back to top
View user's profile Send private message Reply with quote
tox
Franchise Player
Franchise Player


Joined: 16 Nov 2015
Posts: 17835

PostPosted: Sun Jan 22, 2017 7:45 pm    Post subject:

fiendishoc wrote:
tox wrote:
fiendishoc wrote:
tox wrote:
fiendishoc wrote:
I brought up the prior actually in response to the Ingram and Randle discussion, not the Westbrook and Harden one which I didn't see.
Hmm, well in the case of both Ingram & Randle, DBPM and DRPM diverge a good amount, even given DBPM being a prior for DRPM.

DBPM likes both of them better. The Randle part doesn't surprise me, but the Ingram part does on first glance.


Yeah it's not exactly the same as BPM, but I would guess that it shares the same flawed interactions and there's some other stuff in there like height as you've noted. I can't see the numbers at the moment- what is it about Ingram that's changed the most?

Ingram's DBPM is okay. -0.9. Definitely not worst in the league. His DRPM is something like -2.8 (probably going down after tonight), bottom 7 in the league and worst on the Lakers (worse than Lou, Clarkson, Young, etc.). I know why RPM doesn't like him, even beyond his own limitations -- lineup data isn't in his favor. But it still surprises me a bit regardless.


Too bad Engelmann isn't posting the RAPM numbers anymore so we can't see just how much it's pulling him down. You have one of the most played lineups in the league with the starters doing pretty well and then a pretty decent sample of replacing Deng (who is no on off stud himself) with Ingram and it goes into the tank.


Yeah. My gut feeling is RAPM wouldn't be favorable to him either. It's not just the starters. Most lineups do better defensively with Deng playing. Put Ingram in a small ball lineup with Deng and they do poorly defensively as well (to be expected). Forget about putting Ingram at the 1 or 2, which are disasters not just by the stats but by the eye test. The lineup data points to him being a sub-par defender on the worst team in the league, even if we can see some of the good things he does.

His -0.9 "BPM" prior (whatever it really is) might actually be dragging him up lol.
Back to top
View user's profile Send private message Reply with quote
MJST
Retired Number
Retired Number


Joined: 06 Jul 2014
Posts: 26085

PostPosted: Mon Jan 23, 2017 3:30 am    Post subject:

Had a bad game, let's see how he responds next game.

I think expecting 2-3 good performances out of Ingram followed up by a 1 for x wth happened performance is about where we are in terms of what to expect of Ingram's consistency at this point in his career.
Back to top
View user's profile Send private message Reply with quote
KeepItRealOrElse
Retired Number
Retired Number


Joined: 11 Oct 2012
Posts: 32767

PostPosted: Mon Jan 23, 2017 3:47 am    Post subject:

anpherknee wrote:
GoldenThroat wrote:
https://media.tenor.co/images/fb3f2d1e814190100a4ae401b1660d5b/tenor.gif


it me


"analytic" *cue gif* *CUE GIF*
Back to top
View user's profile Send private message Reply with quote
pio2u
Retired Number
Retired Number


Joined: 26 Dec 2012
Posts: 54520

PostPosted: Mon Jan 23, 2017 5:52 am    Post subject:

Quote:
Ben Golliver ✔ @BenGolliver
NBA's worst single-game +/-, last 3 seasons
2015: Lakers' Wesley Johnson -41
2016: Lakers' Lou Williams -47
2017: Lakers' Brandon Ingram -45
Back to top
View user's profile Send private message Reply with quote
yinoma2001
Retired Number
Retired Number


Joined: 19 Jun 2010
Posts: 119487

PostPosted: Mon Jan 23, 2017 6:18 am    Post subject:

Both the eye and analytical tests confirmed it was a historically bad game last night from Ingram.

Yet, he's 19, so chalk that up to growing pains, a bad role (starting PG) and missing a key player like DLO.
_________________
From 2-10 to the Western Conference Finals
Back to top
View user's profile Send private message Reply with quote
av3773
Star Player
Star Player


Joined: 11 Dec 2011
Posts: 3750

PostPosted: Mon Jan 23, 2017 6:49 am    Post subject:

I think ingram will be fine. I think everyone knew realistically it was going to take some time for him to be a consistent impact player, he just has to fill out a lot more. I do like that he has exhibited a lot of really great tools, and he seems to be a really humble, hard worker. It will be at least another 2 years most likely before have any real sense of what level player he will be.

For me I'm much more bullish on our young guys now than before the season in-spite of our record. I've come around a lot of Randle, still optimistic about D lo, very optimistic about Ingram. Still like what Nance brings to the table. I think in a few years Zubac is going to be another late draft laker gem. The one young player I'm not as big on is JC, but it's not like I think he is trash or something, it just seems he regressed a little bit this year.

I'm hopeful we keep the pick and can get in the top 2, because as much as I know this has gone back and forth on the boards, I'd really like to see D lo playoff ball consistently and turn over the primary ball handling duties to one of the top PG prospects in this years draft if we are able to get either of them. I think that still allows D lo to use his passing gifts, but I think it will do a lot of good for putting him in great shooting positions to take advantage of his shooting abilities. I know we aren't talking "traditional" PG, but Fultz or Ball as the primary ball handler along with d lo, randle and Ingram? For me dream scenario for next years starting line up.

PG - Fultz
SG - Dlo
SF - Ingram
PF - Randle
C - Zubac

I know that's not going to happen even if we got Fultz or Ball, but I would love to see it even if it means some real ups and downs, because in the long run I see those 5 being a really dynamic unit with Nance, JC off the bench
Back to top
View user's profile Send private message Reply with quote
Display posts from previous:   
Post new topic    LakersGround.net Forum Index -> General Basketball Discussion All times are GMT - 8 Hours
Goto page Previous  1, 2, 3 ... 310, 311, 312 ... 1883, 1884, 1885  Next
Page 311 of 1885
Jump to:  

 
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum






Graphics by uberzev
© 1995-2018 LakersGround.net. All Rights Reserved. Privacy Policy. Terms of Use.
LakersGround is an unofficial news source serving the fan community since 1995.
We are in no way associated with the Los Angeles Lakers or the National Basketball Association.


Powered by phpBB