The All-Star Snubs

Editor’s Note: The following was written by Andres (Dre) Alvarez and originally posted at Nerd Numbers. In this post, Dre reviews the players who should have – and who should not have – been named to the All-Star game (according to Wins Produced). In addition to Dre’s discussion of “All-Star Snubs”, I would like to let everyone know about the following:

Devin Dignam – at NBAEh? – has written a great discussion of the Rookie-Sophomore game. In addition, Devin has also offered a discussion of Championships and the NBA Draft. Finally, it should be noted tha Devin has created a “Required Readings” page that is well worth reading (it is required, after all).
Arturo Galletti – from Arturo’s Amazing Stats – has been offering posts that allow on to visualize NBA player evaluations.
Ty Willihnganz – at Courtside Analyst – has a series of posts evaluating all-time great NBA players (like Magic, MJ, Kareem, Sir Charles, etc…). He also has a post looking at what happens when a star player departs a team.
At the Miami Heat Index, Mosi Platt also offers a post examining all-time great NBA players. More specifically, Mosi discusses the 50 greatest players in NBA history.
For those who want a discussion of something more recent, Ian Levy – at Hickory High – has a recurring feature called “Last Night’s Numbers”. This feature reviews the numbers from each game played in the NBA (and you can’t get more recent than this).

All of this writing has been offered within the past 48 hours. So hopefully a quick tour around the Wages of Wins Network is becoming a part of your daily routine. If you miss a day or so, you will miss a great story (or two or three or four).

One last note before we get to Dre’s coverage of the All-Star selections. As Dre notes, my actual job is reducing how much time I have for blogging. Currently I am scheduled to write (by myself and with a number of co-authors), at least ten papers. These all need to be finished by the end of June. In addition, I am also teaching more than 200 undergraduates and graduate students this semester. I don’t plan on giving up writing for this blog and Huffington Post, but I suspect the amount I write on-line has to decline. I do want to thank everyone for making this forum a part of their day. And hope on days I don’t post you spend a bit more time looking around The Wages of Wins Network.

I don’t know if Dr. Berri has said this in any of his posts but he is actually quite busy with real professor work and as such has let some of us bench players fill in on some his posts. That topic seems rather valid today as the NBA announced the coaches’ picks for the reserve All-Stars. The starters are picked by the fans and they may use whichever metric they choose. The coaches are not allowed to vote for their own players and as such we’d expect they’d vote for the best talents rather than their favorites. So did the coaches pick the top talents in the league or were they swayed by other factors? Let’s take a look at how the final rosters shook out and if the coaches or fans left anybody deserving at home.

Player	Team	Pos	MP	WP48	Wins	Rank	RB36	AST36	PT36
Chris Paul	NOH	PG	1791	0.398	14.9	2	4.1	9.8	17.0
Kobe Bryant	LAL	SG	1643	0.223	7.6	15	5.3	5.2	27.5
Kevin Durant	OKC	SF	1753	0.225	8.2	10	6.2	2.6	26.4
Carmelo Anthony	DEN	F	1472	0.126	3.9	45	8.0	3.1	24.5
Tim Duncan	SAS	FC	1407	0.287	8.4	8	11.5	3.7	16.7
Pau Gasol	LAL	FC	1838	0.263	10.1	7	10.2	3.5	17.7
Manu Ginobili	SAS	GF	1511	0.253	8.0	12	4.4	5.5	21.5
Blake Griffin	LAC	FC	1806	0.273	10.3	3	12.2	3.5	22.0
Dirk Nowitzki	DAL	PF	1346	0.194	5.4	25	7.2	2.6	24.2
Russell Westbrook	OKC	PG	1732	0.198	7.2	16	5.1	8.5	22.3
Deron Williams	UTA	PG	1732	0.226	8.2	11	3.5	9.0	21.0
Total/Average			1639	0.242	8.4	14	7.1	5.2	21.9

Table 1: Western Conference All-Stars. Players in bold are “suspect” picks. Rebounds, assists and points listed per 36 minutes.

Player	Team	Pos	MP	WP48	Wins	Rank	RB36	AST36	PT36
Derrick Rose	CHI	PG	1783	0.199	7.4	11	4.3	7.8	23.3
Dwyane Wade	MIA	G	1615	0.314	10.6	4	6.6	4.1	25.2
LeBron James	MIA	F	1757	0.326	11.9	2	6.7	6.9	24.5
Amare Stoudemire	NYK	FC	1784	0.103	3.8	31	8.5	2.6	25.3
Dwight Howard	ORL	C	1704	0.370	13.2	1	13.5	1.2	22.2
Ray Allen	BOS	SG	1720	0.213	7.6	10	3.7	3.1	17.4
Chris Bosh	MIA	FC	1553	0.153	4.9	22	8.3	1.9	18.8
Kevin Garnett	BOS	PF	1210	0.341	8.6	8	10.4	2.6	17.4
Al Horford	ATL	FC	1644	0.309	10.6	3	10.2	3.7	16.6
Joe Johnson	ATL	GF	1457	0.143	4.3	27	4.2	5.4	20.3
Paul Pierce	BOS	GF	1652	0.245	8.4	9	5.3	3.4	20.0
Rajon Rondo	BOS	PG	1379	0.305	8.8	7	4.3	12.1	10.3
Total/Average			1456	0.235	7.7	10	6.8	3.9	18.2

Table 2: Eastern Conference All-Stars. Players in bold are “suspect” picks. Rebounds, assists and points listed per 36 minutes.

The players in bold did not make the top twenty in wins produced for their conference. We see that both out East and out West the fans picked one player that had no business starting in Amare and Carmelo. When it came to rectifying the fans mistakes the coaches actually did an alright job. With the exception of Dirk, Bosh and Johnson all of the replacements were sound. Out East they may have cheated a little and just picked all of the starting Celtics minus Shaq. Dirk is not even that terrible of a pick as he is a bit rusty and was in the top twenty before falling to injury. Chris Bosh is a tad iffy as a call but again he is not far removed from the top twenty. The real question mark however is Joe Johnson. Joe Johnson is playing above average, which is exactly the kind of positive language you want to use for a max contract guard. In returning to the Yay Points! Hypothesis we see his only major distinguishing characteristic is that he hit over 20.0 points per 36 minutes. It seems coaches can be fooled by points as well.

Having seen the All-Stars themselves we should take a look at the top players in each conference and see which players missed the cut and see if we can figure out why.

Player	Team	Pos	MP	WP48	Wins	RRB36	AST36	PTS36
Kevin Love	MIN	FC	1762	0.448	16.4	15.2	2.5	21.0
Chris Paul	NOH	PG	1791	0.398	14.9	4.1	9.8	17.0
Blake Griffin	LAC	FC	1806	0.273	10.3	12.2	3.5	22.0
Zach Randolph	MEM	FC	1675	0.295	10.3	13.0	1.5	19.8
Lamar Odom	LAL	F	1652	0.294	10.1	10.3	3.1	16.5
Steve Nash	PHO	PG	1488	0.326	10.1	4.0	12.1	18.3
Pau Gasol	LAL	FC	1838	0.263	10.1	10.2	3.5	17.7
Tim Duncan	SAS	FC	1407	0.287	8.4	11.5	3.7	16.7
Marcus Camby	POR	C	1119	0.354	8.2	14.2	2.9	7.5
Kevin Durant	OKC	SF	1753	0.225	8.2	6.2	2.6	26.4
Deron Williams	UTA	PG	1732	0.226	8.2	3.5	9.0	21.0
Manu Ginobili	SAS	GF	1511	0.253	8.0	4.4	5.5	21.5
Tyson Chandler	DAL	C	1260	0.295	7.7	11.9	0.5	13.1
Jason Kidd	DAL	PG	1609	0.228	7.7	5.1	9.0	8.7
Kobe Bryant	LAL	SG	1643	0.223	7.6	5.3	5.2	27.5
Russell Westbrook	OKC	PG	1732	0.198	7.2	5.1	8.5	22.3
Emeka Okafor	NOH	C	1586	0.216	7.1	11.3	0.6	12.3
Nene Hilario	DEN	FC	1347	0.239	6.7	8.6	2.5	17.8
Andre Miller	POR	PG	1551	0.187	6.0	4.2	8.0	14.5

Table 3: Top 20 Players out West. Possible Snubs in Bold. Rebounds, Assists and Points in Per 36 Minutes.

Player	Team	Pos	WP48	Wins	RRB36	AST36	PTS36
Dwight Howard	ORL	C	0.370	13.2	13.5	1.2	22.2
LeBron James	MIA	F	0.326	11.9	6.7	6.9	24.5
Al Horford	ATL	FC	0.309	10.6	10.2	3.7	16.6
Dwyane Wade	MIA	G	0.314	10.6	6.6	4.1	25.2
Kris Humphries	NJN	PF	0.350	9.3	13.0	1.6	12.5
Landry Fields	NYK	GF	0.283	9.1	8.0	2.1	11.2
Rajon Rondo	BOS	PG	0.305	8.8	4.3	12.1	10.3
Kevin Garnett	BOS	PF	0.341	8.6	10.4	2.6	17.4
Paul Pierce	BOS	GF	0.245	8.4	5.3	3.4	20.0
Ray Allen	BOS	SG	0.213	7.6	3.7	3.1	17.4
Derrick Rose	CHI	PG	0.199	7.4	4.3	7.8	23.3
Josh Smith	ATL	F	0.207	7.4	9.2	3.6	16.7
Andre Iguodala	PHI	GF	0.249	7.0	5.6	5.4	13.9
Andrew Bogut	MIL	C	0.230	6.7	11.7	1.9	13.4
Mike Dunleavy	IND	GF	0.219	6.2	6.0	2.4	14.2
Jose Calderon	TOR	PG	0.201	5.5	3.5	10.1	12.2
Joakim Noah	CHI	FC	0.286	5.3	11.5	2.6	13.7
Elton Brand	PHI	FC	0.150	5.1	8.9	1.5	15.5
Carlos Boozer	CHI	FC	0.250	5.0	11.5	2.5	22.1
Amir Johnson	TOR	PF	0.186	5.0	9.2	1.6	14.2

Table 4: Top 20 Players out East. Possible Snubs in Bold. Rebounds, Assists and Points in Per 36 Minutes.

I was careful to say possible snubs in the table captions. A snub is a player that should have been picked instead of more deserving players. I’ll let you guys pick at the top twenty lists to agree or disagree with who should have gotten in. I would like to list a few players I think were snubbed.

Western Conference Snubs

Kevin Love – Head and shoulders the best player in the league right now. He is scoring over 20 points a game on a TS% of 59.3%. He’s pulling down over 15 boards a game. I have no clue what the heck else the coaches want in a player. Epic failure by both the fans, who picked Melo over him and the coaches, who apparently are crazy. Editor’s Note: Since Dre wrote this, Love was named to the team by David Stern as the replacement for Yao Ming.
Zach Randolph – Another player snubbed for players on better teams. Zach Randolph is putting up amazing numbers and has Arturo Galletti (@ArturoGalletti) saying crazy things on twitter like that the Grizzlies may actually be contenders.
Lamar Odom – This may inadvertently be the fault of the fans over the coaches. Lamar Odom could arguably have fit in as a small forward next to an amazing front court of Love, Griffin and Randolph. However, the fans picked two small forwards as their starters. Lamar is still a top four big out West though and definitely deserved a nod. He keeps up his streak of being underappreciated.
Steve Nash – He just managed to go toe to toe with Love and Howard for the month of January. Chris Paul is the undisputed top point guard but Nash is hands down the second best guard out West. The confusing thing is his scoring and assist numbers are great. The fans get a reprieve as Chris Paul was the obvious starter at PG but the coaches fail here.

In the case of Love, Randolph and Nash the players seem to suffer from not having a strong enough team. In short, the coaches don’t realize how much they need to credit these players. The same is actually true in Odom’s case. The fans and coaches agree Kobe is important and probably Gasol as well, but they may not realize how much Odom contributes on that team.

Eastern Conference Snubs

Kris Humphries – A classic story of an underrated player good at getting boards. Part of his problem is that he is not getting enough minutes. Unfortunately both coaches and fans are swayed by per game numbers and Humphries is below 10 on both points and rebounds. Looking at his per 36 numbers shows how well he’s actually been playing.

Landry Fields – Dr. Berri said Landry Fields for the hall of fame on the podcast (he may have been being sarcastic) I love Landry Fields and he belongs in this game. While he did just get his points per game to above 10 his stat line doesn’t have the classic 15.0+ points or 10+ assists that coached tend to hope for in guards. He is an amazing talent that may be partially responsible for Amare starting. Editor’s Note: I also said Landry Fields will never be in the Hall of Fame. The “never” part of this statement was disputed.

Josh Smith or Andre Iguodala – I list this one as a tie. If the coaches wanted a guard-forward, then Andre Iguodala was a much better choice than Joe Johnson. If the coaches wanted a Hawk then Josh Smith was a much better option. Sure, they don’t hit 20 points per game but it turns out that the game involves other things and these two players are actually quite good at them.

Out East it looks like the snubs were more forgivable than out West. David Stern will be able to name a replacement for Yao and unfortunately I don’t think he’ll be able to right all wrongs with just one pick. The clear choice is of course Kevin Love, who sadly isn’t getting much love from the fans or the coaches. Out East all I can say is that Joe Johnson gets to be part of yet another bad choice made by NBA decision makers.

-Dre

23 Responses to "The All-Star Snubs"

stephanieg says:

February 5, 2011 at 9:11 am

IMO the ASG should return to its roots and be focused on entertainment, not pats on the back for being on a good team. That leads to Mo Williams being in it or Joe Johnson being one of the worst ASG performers of all time. The weird thing to me is that Josh Smith and Nash are built for exhibition games and still didn’t get it. And Nash is like the only two time winning MVP to not get in while still healthy and playing well.
Greyberger says:

February 5, 2011 at 11:19 am

Yeah the ASG voting process is a travesty. But what can we say about that, when the electoral college still exists? Anyway, if Landry Fields’ rebounding was really worth nine wins through forty-nine games they wouldn’t be one of the worst rebounding teams in the league.

Landry Fields is a nice player for the Knicks. He’s not the second best two guard in the league. That’s so preposterous the laugh test might be relevant.

Fields v. Wade, Fields v. Ginobili. Kris Humphries up against Nowitski, Boozer, West, Duncan, Garnett, Randolph, Aldridge, Millsap and Bosh. Would you really make these trades, or suggest that a GM consider them?
kevin says:

February 5, 2011 at 12:04 pm

Dre,

Can you reformat the tables? They overlap the links column on the right and are hard to read.

Thanks for the thread.
dberri says:

February 5, 2011 at 12:14 pm

Sorry about the tables. They look better at Dre’s site.
Nerd Numbers says:

February 5, 2011 at 1:25 pm

Greyburger,
The laugh test doesn’t pass the laugh test :)
Your trades
Field over Wade – Tough one. Wade’s contract gets pricey in a few years when he leaves his prime. He’ll still have trade value so I keep Wade.
Fields over Ginobli – Fields. Ginobli is older and Fields contract is a steal for two years and his extension will still be underpriced
Humphies wins over Nowitzki (age+price), Boozer (injury + anger concerns), Duncan (age), Garnett (age), Randolph (rough one but price + off the court = no), Aldridge (What’s he doing here?) and Bosh (price and not as good).
Millsap is a rough one as he is very affordable and has been playing well. Still go with Humphries

Nothing personal but the “would a GM do that?” doesn’t really make me back down from my thoughts as some GMs have made some very bad moves.
marparker says:

February 6, 2011 at 12:18 am

Its illogical to both try to improve gm decisio making yet ud.e their previous decisions as a basis for argument.
mystic says:

February 6, 2011 at 8:29 am

marparker,

right. I doubt that the people here understand that they applying their metric to players, which were already choosen by GMs to participate in the NBA.
The only way to prove their claims is something which never will happen. Thus they can pretty much say what they want and can feel they are right.

They are calling that science, but in reality it is just the fake science of an economy guy who felt in love with the regression analysis without having to show causality. Berri never showed that the distribution of the wins among players has any connection to the reality. Their prediction of the Warriors in this season is a prime example of the failure of that metric. Even with the current minutes distribution the Warriors would be predicted to win 32 games so far by WP48. They are at 22 right now.

What Berri is not accounting for is the fact that most players will get the same minutes in the same role as they got the season before. Thus, it will give a reasonable correlation to the future success when we are using WP48 from the previous season. But that is also true for the minutes per games as Rosenbaum&Lewin (2007) shows. That saves his metric as long as they are not any great changes like we can see that on the Warriors right now.

Without having to prove anything at all on an individual player level you can claim all you want, you can always find an excuse why it isn’t working that well. They can claim that they choose Humphries over those players, because they don’t have to make that decision in real life, neither can they prove that this decision will lead to more success in real life.

It is quite telling that the Mavericks traded Humphries for salary purposes and picked Nowitzki. With Nowitzki playing this season they are now 33-8, they are 2-7 without him. When Nowitzki is on the court they are outscoring their opponents by +12.4 per 100 possessions, but they get outscored by 9.6 points per 100 possessions when he is not playing. WP48 isn’t telling anyone that the offense of the Mavericks completely falters when he isn’t on the court. They don’t know that Kidd has to take contested jump shots, if he doesn’t want to turnover the ball, when Nowitzki isn’t on the court instead of his open set shots. Anyone saw the game winner by Kidd in the game against the Celtics? Watch closely how the defense collapsed on Nowitzki as he had the ball (after the bad pass to his knees by Terry). Watch how the Celtics are going back on defense and giving Kidd the open shot, because they actually want him to take that shot. They are already boxing out, before Kidd even took the shot. Why? Because Kidd is a weak shooter who doesn’t draw any attention from the defender. That is also the reason why his teams never did exceptional well on offense. That’s something which this metric is not accounting for. But the reality is doing.

They can refuse those +/- and On/Off numbers as much as they like, but that is the reality. That’s what really happens on the court. And their metric is doing a poor job at predicting success of certain lineups. Should we really assume that a lineup Barea, Terry, Marion, Nowitzki, Haywood always just faced the worst lineup of the opponents, because WP48 suggest they are playing 0.365 ball, while in reality this lineup with the most minutes of the Mavericks is outscoring their opponents heavily during their 267 minutes they played together. On the other side WP48 suggests that a lineup with Ridnour, Brewer, Beasley, Love, Milicic has together 0.552 wins per 48 minutes. Well, in reality they are getting outscored by 7.2 per 100 possessions. Did they really faced always lineups with a higher WP48?

But scientifically those examples aren’t worth much, because the sample size is rather small and the +/- numbers are not “consistent” enough. Thus those people really don’t need to fear anyone who will show the worthlessness of their metric on the individual player basis via a peer-reviewed paper. You can use a coin flip to decide whether you pick center A or B from your team or WP48, it is basically the same when it comes to the reality. Most teams have already only a limited selection of players anyway for each position. That helps to make bold statements, because Berri and his lemmings never had to make a decision which player should be signed. That was already done for them. Now, you can claim whatever you want, because that will not be tested in the reality anyway. And when it is tested on a reasonable example with changed minutes distribution like the Warriors in this season, the metric fails badly. The excuses are cute, thus they will not learn from that failure.
dj2007 says:

February 6, 2011 at 10:48 am

how do you explain the fact that kevin love by himself has produced nearly 18 wins so far, but his team has only 11??? Seems like a major discrepancy, and it’s not like other players on the t wolves are producing large numbers of negative wins….
Nerd Numbers says:

February 6, 2011 at 12:25 pm

dj2007,
Excellent question. First my numbers do put Love at about 17 wins. Love and Ridnour are the only two Wolves above average at this point earning about 20.6 wins. The rest of the Wolves are below average and racking up -5.4 wins.

Next I did a quick check on close games for the Wolves (OT or end score less than 5 points in difference) The Wolves have had 14 such games and won only 2 of them. Had they won 2-3 more their record would be closer to their point differential (which is what Wins Produced is modeled on)
dj2007 says:

February 6, 2011 at 1:51 pm

So basically, they’ve been unlucky….
Buddahfan says:

February 6, 2011 at 6:41 pm

Not proofed for typos – Sorry about that
—————————————————–
Amir Johnson continues to put up very good advanced stats even after his promotion through Evans injury to the starting PF spot on the hapless Raptors

Johnson is now averaging 26.1 mpg but close to 29.5 mpg as a starter.

Using Basketball-Reference.com which computes their Win Shares slightly different than WOW Journal I ran a scan of all players in the NBA averaging under 30 mpg sorted by Win Shares. Even though Johnson is averaging only 26.1 mpg for he season and playing for the hapless Raptors who only have 14 wins he is currently 3rd in the NBA in Win Shares among all players who are averaging under 30 mpg with 4.7 Win Shares

The two players ahead of him in Win Shares in the under 30 mpg group are Chandler 6.4 and Duncan 5.1 Win Shares.

However when you look at the the three players and their number of Win Shares as a percentage of all Win Shares on their respective teams you get this.

Total 2010-11 Minutes
————
Griffin 1,845
Love 1,835
Johnson 1,329 – Approximately 72% of Griffin and Love’s minutes

Raptors Total Team Win Shares – 16.8
Johnson Win Shares – 4.7
Percent of Team Total Win Shares – 28.0%

Mavericks Total Team Win Shares – 30.1
Chandler Win Shares – 6.4
Percent of Team Total Win Shares – 21.3%

Spurs Total Team Wins Shares – 37.1
Duncan Win Shares – 5.1
Percent of Team Total Win Shares – 13.7%

I also looked at Love and Griffin as they are also along with Johnson on bad teams.

T-Wolves Total Team Wins Shares – 16.3
Love Win Shares – 8.0
Percent of Team Total Win Shares – 49.1% – Is this a record high %?

Clippers Total Team Wins Shares – 20.6
Griffin Win Shares – 6.1
Percent of Team Total Win Shares – 29.6%

Love’s Win Share number and percentage of the total team is pretty amazing. Lets look at Lebron last season at Cleveland and this season at Miami.

2009-10
—————
Cavs Total Team Wins Shares – 60.5
James Win Shares – 18.5
Percent of Team Total Win Shares – 30.6%

2010-11
—————
Heat Total Team Wins Shares – 39.1
James Win Shares – 9.3
Percent of Team Total Win Shares – 23.8%

I find it interesting that though the 2009-10 Cavs had a slightly higher winning percentage .744 than the 2010-11 Heat do so far of .725 that James had a significantly greater percentage 30.6% of total Cavs total Win Shares in 2009-10 compared to just 23.8% of this season’s Heat total Win Shares even though the Cavs had a better winning percentage in 2009-10 than the Heat do so far this season.

I would have thought that when the best player on a team plays for a team with a higher winning percentage that their percentage of their team’s total Win Shares would go down compared to when that player is playing on a team with lower winning percentage.

For example if we look at Bosh 2009-10 vs 2010-11 we find

2009-10
—————
Raptors Total Team Wins Shares – 37.3
Bosh Win Shares – 9.6
Percent of Team Total Win Shares – 25.7%

2010-11
—————
Heat Total Team Wins Shares – 39.1
Bosh Win Shares – 5.8
Percent of Team Total Win Shares – 14.8%

Interestingly enough we see that James Win Share percentage of the 2009-10 Cavs total team Wins Shares was 30.6% while with the Heat his percentage of total team Wins Shares is 23.8% though the two teams have comparable winning percentages.

When you look at the Cavs record this season maybe what this is telling us is that James was a lot more important to the Cavs winning then he is to the Heat winning. However, it seems to me that we already knew that.

http://is.gd/tytzwi
Italian Stallion says:

February 6, 2011 at 7:53 pm

>Their prediction of the Warriors in this season is a prime example of the failure of that metric. Even with the current minutes distribution the Warriors would be predicted to win 32 games so far by WP48. They are at 22 right now.>

Last I checked The Warriors were above .500 when Lee played and 1-8 when he didn’t.

Considering the optimistic projection for GS was tied to the addition of Lee, perhaps that’s more than just a coincidence. In addition, even when Lee came back he was playing hurt and performing well below his normal level of production. Curry and Biedrins have also missed a number of games.

Injuries are part if the game and many teams have had key players hurt, but GS has had more than their share and doesn’t have the depth that many other teams have to absorb the loss of key players well.

IMHO, GS is an underrated team right now. Most fans are looking at their record and point differential without giving due consideration to circumstance. They are decent.
Italian Stallion says:

February 6, 2011 at 8:00 pm

DJ2007,

>So basically, they’ve been unlucky….<

My view on this probably won't be very popular on this forum, but I think winning and losing is partly a psychological phenomenon.

The ability to execute well in the last few minutes of a close game comes from confidence, experience under fire, mental strength, a good nervous system etc… It's not unusual for young inexperienced teams to find ways to lose games that they should have won. IMHO, Minnesota has probably snatched defeat from the jaws of victory more times than any team in the NBA. I don't think they have been unlucky even though luck is also a factor in basketball.
Italian Stallion says:

February 6, 2011 at 8:27 pm

mystic,

With all due respect, I’d be the first to agree that this metric isn’t perfect, but neither is PER, or adjusted +/-. Of the three, I have an easier time understanding occasional aberrational results from WP and PER than adjusted +/- .

Adjusted +/- produces so many preposterous results it’s hard to take seriously over a single year, let alone portions of single seasons. By the time multiple years accumulate and the data starts making sense, much of it can be become irrelevant because the player’s skill level has changed.

I’ve seen instances where two versions of adjusted +/- created by equally bright and informed people using advanced mathematics produced diametrically opposed results for the same player over the same period. No one that needs reasonable values for players (like a gambler for instance) could ever take adjusted +/- seriously.

I’d rather look at PER even though I think it overrates inefficient scoring and doesn’t recognize the large diminishing returns in scoring either.

I’d rather look at WP even though I think it doesn’t give enough credit or penalize some players at the extremes of scoring usage and recognizes no diminishing returns for rebounding even though IMO they exist.

I can make minor mental adjustments from there or use the model I developed on my own to approximate what I believe are the appropriate values on which I am willing to gamble real money on.
Italian Stallion says:

February 6, 2011 at 9:06 pm

For all you Landry Fields lovers, the Knicks rookie had 25 points and 10 rebounds in a key game against the 76ers today.

He was 10-13 from the field and 5-7 from beyond arc. That’s a ho hum TS% of about 95% give or take a little.

Foe those that believe in impact, several of his 3s were real daggers that either extended the lead at critical time or stopped a Sixer run in its tracks. Huge impact performance over and above the sensational numbers.
Adam C. Madison says:

February 7, 2011 at 7:13 am

Much of the problem with this site comes from the groupthink. It’s also painfully evident David Berri doesn’t get the chance to watch a ton of basketball.

So there is a lot of value from WP, but when it is your only foundation you run into problems.

I think it’s about 80% right, but the groupthink makes it appear like it’s 100% right.
Adam C. Madison says:

February 7, 2011 at 7:14 am

It also doesn’t help that Berri has a very passive-aggressive writing style.
WC says:

February 7, 2011 at 10:35 am

Adam,

I think D. Berri is very reluctant to make changes to a model that observers intuitively believe needs to be fixed without a firm mathematical basis for making the change .

The problem is that many things are extremely difficult to measure with high level studies of all players.

IMHO we are mostly talking about the extremes of scoring usage and diminishing returns in rebounds. But the roles of players tend to change slowly in accordance with the changes of their skill. When they change teams, they are typically being brought in to do more of what they already do.

That makes it difficult to isolate the value of things that “observers” feel could be tweaked to make the model better.

The other obvious problem is that supporters and detractors tend to dig into their positions and lose all sense of objectivity in the discussion.
reservoirgod says:

February 7, 2011 at 1:15 pm

Adam Morrison:

Dr. Berri said on the 1st podcast link he posted that he watches 1 game a night.
Richard Routh says:

February 7, 2011 at 5:15 pm

This truly is a very strange viewpoint that is being espoused here…I only know a little bit about this Win Shares statistic….but let me say this. I am a fan of all new statistics in basketball and I’m certainly not one of those people that is dismissive of stuff that goes against conventional wisdom. BUT, basketball is a game that is hard to quantify, and each different advanced statistic is also not perfect, so what they are REALLY useful for, is pointing stuff out for further examination. For instance, this points out that guys like Kris Humphries, despite their traditional stats and reputations, are playing quite well. And having watched a lot of Nets games this year it’s true, Humph has improved quite a bit and is playing well. (prepare for huge “having said that”)….

HAVING SAID THAT, you need to be reasonable about the whole thing and take some perspective! This stats says Humphries is better than Dirk, or Bosh, or Boozer or whoever else – that’s absolutely insane. It really is, and although I’m not going to doubt anyone’s game-watching credentials, it makes it feel like we’re watching two different games. Humphries is a limited offensive player who, if asked to take the load of a Dirk or someone, would totally embarass himself. I have Nets season tickets by the way so I see him play a lot. The guy above who said he would trade Dirk and Boozer and Aldridge and all those guys for Humphries – go be the GM of some team and see how many wins your guys get.

And I watch almost every Knick game too, and Fields is a hell of a player, but to even consider him and Wade (or even Ginobili) to be of equal value….it’s just crazy. And Fields is a big part of the Knicks success this year, but to say he’s more responsible than Amare…..my impression of the stat is this: it over-favors low-usage, high efficiency guys with big rate stats. It ignores the value of shot creation. And it doesn’t jive with what I’m seeing in the games. Fields’ game – the brilliance of his ability to fill in the gaps – is enabled by someone like Amare, who builds most of the structure which has gaps that need filling, if that analogy makes sense….Fields could not, at this point anyway, take on the role of an Amare or Wade and carry a team like that. How can that just be ignored?
Devin says:

February 7, 2011 at 11:34 pm

Richard Routh:

Wins Produced is descriptive – it tells you how productive a player has been when he plays. What it does not tell you is who is a “better player”.

Sure, if you tried to use Hump like Dirk, Hump wouldn’t be as productive – duh. But forcing Humphries into a role that doesn’t suit him would be stupid. Likewise, putting Nowitzki into Hump’s role may not be the best way or maximizing Dirk’s value. Now, there are some exceptional players who are good at almost everything – Jordan, Magic, Bird, Barkley, LeBron, Wade, and even Kobe to some extent – but these player are freaks of nature.

I’d want Humphries on my team because he’s very productive in his limited role, he’s pretty young, and he’s cheap. If I want my team to be good, somewhere on my roster I’d also need someone (or some ones) who takes more shots – but I’d still want the higher usage guy(s) to be as efficient as possible.
Italian Stallion says:

February 8, 2011 at 8:50 am

IMO Devin has it right.

Just as you can’t put 5 great PGs or 5 great Cs on the court together and expect to succeed, you can’t put 5 low usage scorers that rebound well or 5 scorers than can’t rebound on the court together and expect to succeed.

Once you get past the traditional positional assignment and the bias towards scoring and start thinking of players as scorers, play makers, rebounders, lock down defenders, and various combinations of the above etc… the values become clearer.

One can argue about the relative value of rebounds vs. high usage scoring, but both clearly have value and not all players can do both things well.

It should be clear that a great rebounder that doesn’t score much can add more value than a high usage scorer if the team already has an abundance of players than can score. The opposite is also true.
Italian Stallion says:

February 8, 2011 at 9:00 am

I want to add one other thing that I am almost certain confuses some people.

Cs and PFs tend to be more productive than PGs, SGs and SFs. However, when determining WP, players are not measured based on what they produce. They are measured based on what they produce relative to other players that play their position.

So very often the best player on the team (either a C or PF) is NOT producing the most wins. That will happen when the gap between him and the average PF/C is not as large as the gap between an inferior player and his position.

A good example of this is Amare and Fields for NY.

Fields is producing a lot more wins than Amare even though they are rated very similarly before positional adjustments because Fields is being more productive than the average SG by more than Amare is being compared to the average PF/C.

February 2011
M	T	W	T	F	S	S
« Jan				Mar »
	1	2	3	4	5	6
7	8	9	10	11	12	13
14	15	16	17	18	19	20
21	22	23	24	25	26	27
28

Share this:

Leave a Reply Cancel reply