Is Russell Westbrook a superstar?

[DSMok1]

Can this player be the best player on a championship team?[...]
Wade has shot 10% better and scored 10% more.

Isn't this huge enough gap to say that Westbrook doesn't qualify?

Per my link from above:

Code: Select all

Rk              Player From   To  Tm  Lg    G   GS   MP  PER  TS% eFG% ORB% DRB% TRB% AST% STL% BLK% TOV% USG% ORtg DRtg  OWS  DWS   WS WS/48  FG%  3P%  FT%
1          Dwyane Wade 2006 2008 MIA NBA  177  174 6777 26.2 .571 .495  3.6 11.6  7.7 36.4  2.7  1.8 14.7 33.3  111  105 17.9  8.6 26.5  .188 .486 .241 .784
2    Russell Westbrook 2011 2013 OKC NBA  176  176 6203 23.1 .532 .462  5.1 10.0  7.7 37.8  2.7  0.7 14.9 31.9  109  106 14.2  6.7 20.9  .162 .441 .326 .828

TS% is a 4% difference, Usage a 1.4% difference, and Wade has quite a few more blocks. Westbrook has 1.4% better AST%.

Yes, Wade has been better, but not by a ton. And there are no other players in history with a similar profile of extremely high assist and usage numbers.

Loosening up the list a bit more, LeBron, Rose, Grant Hill, and Marbury are somewhat similar: http://bkref.com/tiny/01dnq

[Mike G]

[...]
Wade has shot 10% better and scored 10% more.

TS% is a 4% difference, Usage a 1.4% difference, and Wade has quite a few more blocks. Westbrook has 1.4% better AST%.
Yes, Wade has been better, but not by a ton. And there are no other players in history with a similar profile of extremely high assist and usage numbers.

The difference between 55% and 51% is 4% -- not huge.
Without selecting for Wade's off-year, it's more like .566/.514 = 1.101. Wade got 10% more points by shooting just as often and making 10% more than Westbrook made.

Loosening up the list a bit more, LeBron, Rose, Grant Hill, and Marbury are somewhat similar: http://bkref.com/tiny/01dnq

Mostly superstars who carry their team, and Marbs. Well, he did have some elite statistics.
Westbrook is therefore the first clear #2 player to ever do this role?

[DSMok1]

Loosening up the list a bit more, LeBron, Rose, Grant Hill, and Marbury are somewhat similar: http://bkref.com/tiny/01dnq

Mostly superstars who carry their team, and Marbs. Well, he did have some elite statistics.
Westbrook is therefore the first clear #2 player to ever do this role?

He's more like a #1A player.

I'm working on some positionality research, and KD and Westbrook don't play all that differently from one another, except Westbrook passes more to KD than vise-versa, and KD's shooting percentage is higher (almost half of his baskets were assisted last year, while Westbrook was around 25%). The way OKC plays, I would say Westbrook is more the driver of the offense than KD is!

[Mike G]

Abruptly changing the subject; why are Chris Paul's FG 30% assisted at home and only 16.7% on the road?
http://www.basketball-reference.com/pla ... ting/2013/

He also shoots a lot better at home: .560 eFG to .478 away; plays more minutes on the road.

[talkingpractice]

perhaps because since 1977-78, looking at all PGs from the ages of just 21-23, he has averaged 29.3 scoring opportunities per 48 minutes - (FGA + FTA/2)/48 - most among PGs in this age range...

Isn't this in large part to him being so much better than almost everyone else in that sample? As written, it implies that the relevant sample here is n = alot, but its really more like n = 10 or something.

he is currently shooting just 41.7% on 2pters - and close to 4/5 of all of his FGAs have been 2pt attempts. that's an awful lot of missed shots rebounded by the defense, i.e. zero point team possessions that he is personally responsible for...

But what about some of the particulars that point the other way, such as

(a) How much of the credit to the high FG% of the rest of the team should rightfully go to RWB?
(b) His FGAs are mostly unassisted.

The usage/assist thing comes into play a lot for me here. In our testing, we can't empirically/analytically find practically any value to an assist unless you take into account the usage interaction. He's a rock star in this regard, and that's a good part of why his IPV is so high. He scores this highly on offense in just about every plus/minus or statistical plus/minus metric, I think.

Also, if RWB isn't indeed as good as some of us think, then the counter-argument almost surely has to be that Durant is even way better than we all think right now. The credit for all their ptdiff has to go somewhere, and there's nothing else that could realistically be making up for all of the value that we currently credit to RWB. It's all offensive value, so it doesn't make sense to think that we could credit it to Perkins or Sefolosha, and not much sense to credit much to Ibaka. So if RWB isn't a stud, then the implication is that Durant is performing at insane levels (way more than Lebron) on offense. I think that this argument is less plausible than the "RWB aint no KD, but he is really darned good too" argument.

I get the feeling that qualitative sorts of things don't make it into these threads much. But the intangibles all favor him. The guy has never missed a game. He plays hard as heck at all times. He gets better constantly. He could play 44 mpg in the Western Conference Finals, and I wouldn't be blown away if he did. He really gives a damn and wants to win. His assists are going up up up, etc. He's even a key player on Team USA. He doesn't have the other attributes that usually go along with being a "selfish" type of player.

Is there a conversion from IPV ( * minutes or possessions) to individual wins added?

In terms of converting IPV We do basically what Dsmok said, but we use -3.3 for replacement level. The MVP column in our player values table takes into account team pace and total MP to get a warp-like measure, but we keep it in points rather than converting over to wins.

I guess I don't much trust that Dave Berri thing? That's probably due to the source more than due to the method used.

[v-zero]

In terms of converting IPV We do basically what Dsmok said, but we use -3.3 for replacement level. The MVP column in our player values table takes into account team pace and total MP to get a warp-like measure, but we keep it in points rather than converting over to wins.

Is IPV relative to zero or replacement level (-3.3) ?

[talkingpractice]

In terms of converting IPV We do basically what Dsmok said, but we use -3.3 for replacement level. The MVP column in our player values table takes into account team pace and total MP to get a warp-like measure, but we keep it in points rather than converting over to wins.

Is IPV relative to zero or replacement level (-3.3) ?

It's relative to 0... same scale as the other p/m and spm metrics.

[Dr Positivity]

I have him around the same level as Blake Griffin, LMA, Curry, Noah, etc. All-star caliber but not a superstar.

I don't think Westbrook is selfish, I think he just has a weak feel for the game and spatial awareness, which is one of D Wade's strengths

[bchaikin]

Isn't this in large part to him being so much better than almost everyone else in that sample? As written, it implies that the relevant sample here is n = alot, but its really more like n = 10 or something.

re-read the statement:

perhaps because since 1977-78, looking at - ALL - PGs from the ages of just 21-23, he has averaged 29.3 scoring opportunities per 48 minutes - (FGA + FTA/2)/48 - most among PGs in this age range...

since 1977-78, 356 different PGs played at least 50 total minutes in the league (westbrook 6266 minutes) from the ages of 21-23, and among those 356 different PGs in that age range, westbrook has averaged taking the most scoring opportunities per minute...

not quite sure where this:

its really more like n = 10

is coming from...

among those same 356 PGs, westbrook's 45.8% 2pt FG% is actually less than the average 2pt FG% of this group of PGs ages 21-23, which is 46.1%...

right now in 12-13 his 2pt FG% is just 42.1%, while the league average 2pt FG% is 47.5%, over 5% better. he has taken more 2pt FGAs than all but 7 other players in the league, yet among the 24 different players to have taken the most 2pt FGAs, he has the lowest/worst 2pt FG%...

that means alot of missed 2pt FGAs followed by defensive rebounds, each of which is a zero point team possession due solely to westbrook...

i am not saying he isn't a good player - he is - but at some point in statements like this:

So, is he 1st team all-NBA, 2nd team, or merely an all-star?

i would think offensive efficiency (and defensive ability) would be considered, rather than just per game stats like pts/g, reb/g, and ast/g...

also oklahoma city has already shown - in 2011-12 - that they do not need alot of assists to have a great offensive team, as they were 1st in offensive efficiency last season but dead last in the league in ast/fgm...

they are currently 1st in offensive efficiency again this season, yet are worse than the league average in ast/fgm. so any proposed argument that westbrook's assists are necessary for their excellent offense flies in the face of what they did just a season ago...

then the counter-argument almost surely has to be that Durant is even way better than we all think right now

durant is currently scoring 28.5 pts/g on a 63.0% ScFG% (2s, 3s, and FTs). his offensive efficiency is at 2.79 pts/0ptposs (points scored per zero point team possession personally responsible for). the last player to play at least 500 minutes in a season and score that much shooting that well was charles barkley back in 1987-88 some 25 years ago. the last player to score that much with that high of offensive efficiency was michael jordan back in 1990-91, some 22 years ago...

on top of this he also leads the thunder in defensive rebounds, and http://www.82games.com shows opponent counterpart data that infers his defense has been very good to excellent...

[Mike G]

I'll propose that a significant amount of Westbrook's production is a result of having Durant on his team, and vise versa.
If Westy weren't there, Durant might get double-teamed a lot more.

The Thunder's Ast/FG ratio last year was lowest in the league, about .014 lower than the next-lowest Kings.
But their away Ast/FG ratio was by far the lowest, at .465. This is .044 lower than 2nd-lowest Sac. League avg was .566
They "marked up" their assists at home by some 13%, 5th most egregious in the league.
Worst were the Clipps at 118% of their road ratio.

How did they do so well with so few assists? Again guessing, their top 3 guys were very good at creating shots for themselves, when they had Harden.
This year, they're much closer to the NBA avg (.579) Away Ast/FG, at .558

[DSMok1]

How did they do so well with so few assists? Again guessing, their top 3 guys were very good at creating shots for themselves, when they had Harden.
This year, they're much closer to the NBA avg (.579) Away Ast/FG, at .558

Last year, on the pick and roll (a very common play) they almost never passed it to the roll man. A lot of assists come from the roll/pop man in the pick and roll. The big 3 all almost exclusively took their own shot last year. This year, they are making a conscious effort to get it to Ibaka.

[Crow]

This season with and without each other:

http://www.nba.com/advancedstats/player ... 3;season=r

Little difference for RW's shooting with and without KD; but for KD, he is shooting the 3 ball better with Russ this season while the team +/- is notably better without Russ.

But, to look further:

KD shoot better from both 2 and 3 pt range last playoffs with Russ. Team better with KD and Russ than just KD. Russ shot the 2 pt shot better without KD but didnt hit a single 3 pointer without him. OKC won big the Russ without KD minutes, but there were very few such minutes.

KD was again better shooting in 2010-11 with Russ (not on assists); Russ slightly worse shooting with KD than without.

In 2011 playoffs Russ shot significantly better with KD compared to without. KD was mixed- better from 2 pt with Russ, better from 3 pt without. In 2010-11 regular season it was little difference with and without Russ for KD but clearly better with KD for Russ.

In 5 time periods, KD with Russ better than KD without Russ on team +/- by a score of 3-2. Russ without KD also wins 3-2 over Russ with KD. At the individual level KD was mostly better over the periods with Russ than without. Russ used to be better with KD but he was been worse with KD since the 2011 playoffs. Random or a real shift that will continue? Probably the former but could / should look closer and follow it going forward.

[talkingpractice]

Isn't this in large part to him being so much better than almost everyone else in that sample? As written, it implies that the relevant sample here is n = alot, but its really more like n = 10 or something.

re-read the statement:

perhaps because since 1977-78, looking at - ALL - PGs from the ages of just 21-23, he has averaged 29.3 scoring opportunities per 48 minutes - (FGA + FTA/2)/48 - most among PGs in this age range...

since 1977-78, 356 different PGs played at least 50 total minutes in the league (westbrook 6266 minutes) from the ages of 21-23, and among those 356 different PGs in that age range, westbrook has averaged taking the most scoring opportunities per minute...

not quite sure where this:

its really more like n = 10

is coming from...

Oh, sorry... maybe I phrased this poorly. My claim was simply that imo the relevant sample size here is more like n=10 (the guys who are good enough to have even conceivably taken that many shots), as opposed to n=350 or whatever. I think that most of these n=350 guys 'dont really count', if we're honestly comparing/discussing how much usg that RWB eats up compared to other pgs, since almost all of those guys (n=340 of them or so?) aren't 'honest candidates' to have used that % of their teams possessions (they're just not good enough).. But that's just me, and I'm wrong a lot, so who knows.

I can say that RWB has an IPV of 4.47 (to Durant's 6.18), and that our track record is sweet .