APBRmetrics

The team data with him went from 2nd best to 2nd worst and he hasn't changed? You can believe that but it seems very, very unlikely to me. It seems far more likely to me that his defensive impact has diminished this season than that it has stayed the same. The raw on / off data, the raw on data, RPM, BPM and WinShares per 48 all report big declines in positive defensive impact this season.

so you completely ignore actual data - the charted data of synergy people who watch games that rate's leonard's defense as excellent - for calculations that determine how good a player is based on what happens when he does not play?...

Try to slant and ignore what I said all you want

you ignore actual data - the synergy defensive data - and ignore the RPM data where 4 spurs starters rank low in DRPM yet the team is still best in the league on defense. how does that DRPM add up to show the spurs are 1st in the league in defense?...

yet accuse others of slanting/ignoring you?...

but I am satisfied with what I said and am probably done

and this was just getting to be fun...

I am not ignoring the synergy data; I am saying it is not the only thing out there of merit (which is pretty much your position... except you also use team as a whole data and weaken your philosophical case against use of team data including the "without" data it contains considerably) and that is the outlier. Not invalid, just outlier to the rest- this season. They agreed for at least 3 years previously and probably his entire career, except this season when so many metric findings have changed.

Plays that end against him are only 15-25% of his total time / action on court. Mostly or entirely enough for you apparently, not for me.

"Discussions" like this aren't "fun"; but you got the best response I could give in the time I am going to give it and my responses were for others too. Addressing my points with the same words that came before my response and fundamentally ignore my points and repeat your original points isn't really an adequate or improving discussion for me. But it was what it was. Potential paradox and dilemma. With reason for more research and consideration of both sides.

If you want to advance the analysis, show all the detailed synergy / sports Vu data for Leonard this season vs. last season. Is it essentially the same or has it changed? Is not simply a matter of is he good, it is mostly about is he as good this season as last season. If you don't, I might look it up later. Or maybe someone else can handle this step. On my phone NBA.com tables are usually very slow and subject to repeated drops.

Here's what b-r.com claims to know about the Spurs with Kawhi and without him:With Kawhi on the floor, their offense ranks #1, while on D they're tied for 5th.
When he's not been on the floor, their offense shrinks to 24th, while their D has been some 6 pts/100 better than any other team.

This doesn't seem to include the last 2 games, including last night's debacle. Kawhi didn't play, and the gsw ORtg was some 26 points below their norm.
Here are the Spurs' other 6 opponents this season when KL has not appeared:Note each of 6 opponents' ORtg has been sub-par vs the Kawhi-less Spurs; on average 7.4 pts/100
These 6 games account for 1/4 of the 'without KW' minutes.
Players other than Kawhi Leonard seem to move the Spurs' DRtg from 5th best to 1st.

Kyle Anderson is listed as SG but plays 2/3 of his minutes at SF for the Spurs.
The Spurs have only Kawhi Leonard listed at SF. He leads the team in minutes at 33.6 mpg; Anderson is 11th at 13.4
Anderson has started 5 of the 7 games Kawhi hasn't played. In those, he gets 28 mpg on avg. So when backing up Kawhi, he goes not quite 12 mpg (when he plays).

He's also listed #1 in DRPM, again under SG, at 2.35. Danny Green and Jonathon Simmons are #3 and 4. And they can't all be playing SG.
http://www.espn.com/nba/statistics/rpm/ ... position/2
Those 3 total 6.1 defensive pts/100 better than NBA avg
Maybe when Kawhi isn't playing, Pop goes with at least one of these guys?

Does any data source show a team's DRtg after a given player has taken or missed a shot?
Could it be that Kawhi is now taking shots that put him in bad defensive position? They could still be a net positive for the Spurs, but only on offense.
I see he's taking more unassisted FGA, more and better in the 16-24' range, and fewer from the corner and at the rim.
http://www.basketball-reference.com/pla ... l_shooting

Kawhi is having an all-time great offensive season. If he is also DPOY candidate, he should be consensus MVP.
http://bkref.com/tiny/jiqmN

From nba.com

Leonard 2016-17

Defensive rating 103.8 8.9 points worse than last season.
"Defensive rating" tied with his rookie season for worst of his career.

overall dfg% 46.9% (fifth best on team) 2%pts worse than last season

Diffference between actual dFG% and weighted opponent average fg% -0.8 (8th best on team) Drop of 6 spots in team ranking

3pt defg%actual vs expected ).2% (10th best on team) Drop of 3 spots in team ranking

Defg% actual vs expected from within 6 feet of rim +3 (10th best on team) change of -10% pts / drop of 6 spots in team ranking

OPP PTS OFF TO Per 100 Possessions 13.6 Up slightly from last season

OPP PTS 2ND CHANCE Per 100 Possessions 12.6 (6th best on team) Up almost 25% from last season / drop of 5 spots in team ranking

OPP PTS PAINT Per 100 Possessions 42.1 (6th best on team) Up slightly from last season


Leonard 2015-16

Defensive rating 94.9

overall dfg% 44.9% (3rd best on team)

Diffference between actual dFG% and weighted opponent average fg% -5.6 (2nd best on team)

3pt defg%actual vs expected -1.3 (tied for 7th best on team)

Defg% actual vs expected from within 6 feet of rim -7% (4th best on team)

OPP PTS OFF TO Per 100 Possessions 12.8 (Best on team)

OPP PTS 2ND CHANCE Per 100 Possessions 10.4 (best on team)

OPP PTS PAINT Per 100 Possessions 41.0 (4th best on team)

Steal % same this season as last season. Block % down about 25%. Defensive rebound rate down almost 20%.
DBPM, by far worst of his career.


Defense by play type available for those who want it.

Every player defensive stat & tracking stat I checked here got worse for Leonard for this season compared to last season. Many considerably.

So after that, there is no conflict between the actual player data and everything else (the raw on / off data, the raw on data, RPM, BPM and WinShares per 48, to name just the most available ones) that say Leonard's defensive impact has slipped considerably.

Every player defensive stat & tracking stat I checked here got worse for Leonard for this season compared to last season. Many considerably. So after that, there is no conflict between the actual player data and everything else (the raw on / off data, the raw on data, RPM, BPM and WinShares per 48, to name just the most available ones) that say Leonard's defensive impact has slipped considerably.

well then let's dig a little deeper...

take out your calculator and try this - divide the DFGA he's faced, now 522 (stats.nba.com), by his minutes played, 1950, and what do you get? try 522/1950*40min = 10.7 FGA/40min faced on defense...

now look at say - oh i don't know - how about his teammate david lee, who's DFG% and DFG% DIFF are currently both better/lower than that of leonard. that's 530/1186*40min = 17.9 FGA/40min faced on defense...

hhmm... why do you think that might be? why would players on offense want to attempt far more shots (17.9 FGA/40min) against david lee than against say kawhi leonard (10.7 FGA/40min)? especially considering the fact that lee allows a lower FG%?...

why does kawhi leonard on defense face so few FGAs per 40 minutes? care to take a guess?...

you think it's because he only guards players that rarely shoot? perhaps they just don't like to shoot?...

if you watched kawhi leonard on any kind of regular basis you would know why...

on a nightly basis leonard routinely guards the opposition's best offensive player (sans C), and he literally smothers his opponent with ball-deny defense and top shot defense, such that they cannot get shots off...

also - his ability to limit his opponents' FGAs has improved annually, from 12.7 FGA/40min faced in 13-14 to only 11.6 FGA/40min in 14-15 to just 11.5 FGA/40min in 15-16 to now a career best/low 10.7 FGA/40min faced on defense in 16-17...

i.e. so far this 16-17 season he's been better at denying his opponents FGAs - you know - some of the best offensive players in the league - on a regular basis at a career best/low rate...

and if you take the time to look at the FGA/40min faced on defense by all players in the league, you will find his rate of just 10.7 FGA/40min is one of the absolute best/lowest, pretty incredible considering who he guards on a nightly basis...

you think reducing the FGAs of some of the best offensive players in the league on a nightly basis might be one reason for the spurs being the best defensive team in the league right now?...

so what's more important defensively - keeping your opponent's FG% down or stopping him from getting shots off in the first place? especially if you are entrusted to guard the opposition's best offensive player on a regular basis...

and he does this while also finding the time to have one of the highest steal rates among SFs, block some shots, and be a good defensive rebounder for a SF...

add to that the fact that...

Kawhi is having an all-time great offensive season. If he is also DPOY candidate, he should be consensus MVP.

agreed - and if the spurs finish with the best record (gsw without durant makes that possible) he likely will be...

FGA faced is important. Good point. Read some articles this afternoon making same point. Not a foreign observation to me, but I can't always remember everything at once.

Excellent reason to look at global impact as measured by RPM and to some extent by team defensive data.

His defense has declined in many ways directly and perhaps his impact has been reduced also to some degree by opponents going away from him.

The story fleshes out step by step.

It's a hard to ask question, on one hand Kawhi's defensive +/- is outrageously poor, OTOH I just have a hard time believing that Kawhi taking the other team's best offensive option out of his game like a cornerback taking a star receiver out, is somehow hurting the Spurs, you would think eliminating the biggest threat is only a positive. If the other Spurs like Parker and Pau are getting exposed by this go away from Kawhi strategy, even when the star is being covered, it just means they're that bad on defense. If Kawhi wasn't on the Spurs and Pau, Parker, LMA had to face all they are now and the best offensive player too, wouldn't they just get killed more? The simplest explanation to me is that Kawhi plays with worse defenders than the bench unit has, combined with probably his D getting a little worse with offensive workload up

Calculated the minutes weighted team DRPM for Spurs using the values listed earlier and compared to the actual difference between the team defensive rating and league average after 60 games. They were very close. Within 0.6% of each other. Team just slightly better than what DRPM represents as a whole for them. Lots of positive DRPMs add up.

Calculated the minutes weighted team DRPM for Spurs using the values listed earlier and compared to the actual difference between the team defensive rating and league average after 60 games. They were very close. Within 0.6% of each other. Team just slightly better than what DRPM represents as a whole for them. Lots of positive DRPMs add up.

i don't buy it...

4 of 5 spurs starters rank between just 20th to 38th in DRPM at their respective positions, and the team's 6th man patty mills - who also happens to be 4th on the team in minutes played - ranks 48th among PGs in DRPM...

that's 5 players among the 6 with the most in minutes played on the team, that have played 1/2 of the team's total minutes played, 5 with DRPMs nowhere near the best at their positions, yet that and their reserves (and green) somehow adds up to the best defensive team in the league??...

so now dewayne dedmon (21 starts, 17 min/g), david lee (9 starts, 19 min/g), kyle anderson (12 starts, 13 min/g), jonathon simmons (3 starts, 18 min/g), and manu ginobili (no starts, 19 min/g), who have likely spent far more of their time playing against other team's reserves than have gasol, aldridge, leonard, parker, and mills - make this spurs team the best defensive team in the league?...

i can assure you that when kawhi leonard - all-NBA 1st team last year and last year's DPOY - has been on the floor this season the spurs opposition had their starters on the floor most of that time, and when he sat, teams more often than not played their reserves...

but because of what the spurs reserves did against other teams reserves in the 1/3+ of the spurs floor time leonard sat, DRPM lists leonard just 26th in DRPM among all SFs?...

i don't think so...

His defense has declined in many ways directly

his defense hasn't declined at all - take off your RPM blinders and actually watch him play, or look at actual data, whether that synergy, FGA/40min faced, or others...

Kawhi's defensive +/- is outrageously poor

his defense is outrageously good, the metric is poor...

high steal rate, good defensive rebounder, blocks some shots, holds opposing SFs (or whomever he guards) - typically the opposing team's best offensive player - to very few FGAs, and leads the best defensive team in the league in minutes played...

that Kawhi taking the other team's best offensive option out of his game like a cornerback taking a star receiver out, is somehow hurting the Spurs

of course it doesn't hurt the spurs - only DRPM says it does...

If the other Spurs like Parker and Pau are getting exposed by this go away from Kawhi strategy, even when the star is being covered, it just means they're that bad on defense.

who says they're bad on defense? this team is currently the league's best defensive team - they may not be as good defensively as kawhi, but they certainly are not bad...

if you want to see bad defense, watch the lakers, nuggets, or nets...

this is alot like the "RPM can't evaluate anthony davis" conundrum of a few years ago - in 13-14 and 14-15 davis had very similar stats, some almost identical, yet one season RPM ranks him something like 100th in the league but the next season top 5...

kawhi leonard keeps getting better and better, may be league MVP this season, could have been last season (was 2nd in mvp voting)...

again - how many mistakes does DRPM have to make before one realizes it does not measure defense? it currently ranks spurs PF lamarcus aldridge 29th in DRPM among PFs at 1.11, but denver nuggets C nikola jokic (they list him at PF) 15th at 1.87...

yet the publicly available sportvu data at stats.nba.com shows jokic allows a much higher/worse FG% than does aldridge, and also faces 19.8 FGA/40min vs 14.9 FGA/40min for aldridge. what do RPM advocates say about that?...

spurs best in the league in defense, nuggets 2nd worst, sportvu data shows better defense by aldridge (synergy too), but DRPM is right?...

good luck with that...

Bob, you either knew the player tracking data for Leonard season and season and failed to mention all the declines or you didn't check season to season change.

If you don't buy the team DRPM aggregation, you either didn't do it or just don't want to buy it.

It is pretty easy to miss a piece of analysis or to repeat the parts that make the most sense to you. I've done it. There is quite of bit to chew on in this thread about players and methods. Not easy, not fully resolved. But probably better to wrestle with it, than not.

bchaikin wrote: again - how many mistakes does DRPM have to make before one realizes it does not measure defense? it currently ranks spurs PF lamarcus aldridge 29th in DRPM among PFs at 1.11, but denver nuggets C nikola jokic (they list him at PF) 15th at 1.87...

yet the publicly available sportvu data at stats.nba.com shows jokic allows a much higher/worse FG% than does aldridge, and also faces 19.8 FGA/40min vs 14.9 FGA/40min for aldridge. what do RPM advocates say about that?...

spurs best in the league in defense, nuggets 2nd worst, sportvu data shows better defense by aldridge (synergy too), but DRPM is right?...

good luck with that...

Forget DRPM for a second, the Spurs are 8.5 points worse defensively with Kawhi on the floor. When asking what is going on with Kawhi's defensive numbers, it's not hidden in the gears of an advanced stat.

So the question is, how is the Spurs so much worse defensively when Kawhi plays? I have a hard time wrapping my ahead around it still. Again the most logical answer to me is that the 4 guys Kawhi plays with in the starting lineup are a lot worse defensively than the 5 guys off the bench

From Kevin Pelton's article today on Kawhi Leonard's MVP case: http://insider.espn.com/nba/insider/sto ... p-case-nba ($ ESPN Insider)

"This effect is enough to explain nearly the entire difference in San Antonio's defensive rating with and without Leonard. His on-court defensive rating would be 6.7 points per 100 possessions better if opponents shot as poorly on 3s with him on the court as they do with him on the bench (28.1 percent).

There's strong evidence that teams have little control over the percentage opponents shoot from 3-point range, and there's even less reason to believe individual defenders can affect it. Odds are this is a fluke and one that explains a large percentage of Leonard's decline from a plus-3.9 defensive RPM in 2015-16 (and plus-4.6 in 2014-15) to plus-0.8 this season."

"This effect is enough to explain nearly the entire difference in San Antonio's defensive rating with and without Leonard. His on-court defensive rating would be 6.7 points per 100 possessions better if opponents shot as poorly on 3s with him on the court as they do with him on the bench (28.1 percent)... There's strong evidence that teams have little control over the percentage opponents shoot from 3-point range, and there's even less reason to believe individual defenders can affect it. Odds are this is a fluke and one that explains a large percentage of Leonard's decline from a plus-3.9 defensive RPM in 2015-16 (and plus-4.6 in 2014-15) to plus-0.8 this season."

and even in defending kawhi leonard by pointing out this deficiency in DRPM as a fluke - by actually looking at specific real stats - the author is inadvertently and unintentionally doing a disservice to leonard by saying Leonard's decline...

the decline is nothing more than in a calculation in a time slot specifically unique to leonard looking at multiple 5 man lineups on both the spurs defense and opponents offense. but it says nothing specific about what kawhi leonard himself did during that specific time slot and makes assumptions about all players involved, kind of like how DRtg rates each player on defense on the same team the same normalized to minutes played outside of steals, blocks, and defensive rebounding, which is of course ridiculous. sans steals, blocks, and def rebs leonard is an outstanding defender, far better than other spurs starters (except perhaps danny green, who is also excellent)...

also lost in the thoughts of advocates for DRPM is the fact that at the time this discussion of leonard started the spurs were 50-13 with a +8.3 average per game point differential. that's an average per game point differential for both wins and losses. the average point differential of just the wins was obviously much higher - and some portion of those wins were blow outs, meaning some portion of the time leonard was not on the floor (highly likely during a blowout) the stats are really of little consequence...

now at that time leonard had played 1923 minutes out of 3046 possible minutes - he was on the floor for 63% of the time. he was not on the floor for 1123 minutes, or 37% of the time. it's tough enough trying to compare sample population sizes of the same size but here one is close to twice the other, meaning the data of the smaller is more easily skewed...

yet you are now looking at a sample size of 1123 minutes where some portion of that was strictly garbage time - how relevant are those stats? if they are winning 37 games by an average per game point differentiL of some 10+ pts/g, are we looking at maybe some 100-200 minutes of garbage time (over 37 wins)? which then means a sample size of 1923 minutes versus now what some 900-1000 minutes, again much more susceptible to being skewed...

so there is no Leonard decline on defense - nothing actually specific to kawhi leonard. if you want to determine how leonard does on defense, look at actual stats specific to him and/or watch him play. the actual numbers i see say the spurs are the league's top defensive team, leonard holds the opponents he guards to very few FGAs, and from watching him play i know he routinely guards the toughest assignment on the opposition, from PGs to PFs...

also why aren't RPM proponents discussing the anthony davis 1314/1415 difference, or the lamarcus aldridge/nikola jokic numbers for this season? the answer is because they won't tell you anything specific about these players, but will simply give you possible scenarios to attempt to validate RPM numbers that have no basis in fact. anytime you question an RPM number, rather than a discussion explaining the merit of that number, you are criticized for not understanding the metric...