The Eastern Conference has been reviewed (Atlantic, Central, and Southeast Division). Now it’s time to move on to the Western Conference. And so we head to the Northwest and the home of the Jazz, Nuggets, Blazers, Sonics, and T-Wolves.
Wins and Efficiency Differential
Following the format of the previous reviews, I will start by noting each team’s won-loss record and efficiency differential (offensive efficiency – defensive efficiency).
Utah: 14-12, 5.1
Denver: 14-10, 3.6
Portland: 13-12, -2.0
Seattle: 7-18, -7.0
Minnesota: 3-20, -8.4
The top team in this division – in terms of record – is the Denver Nuggets. But the top team in terms of efficiency differential is the Utah Jazz. The Jazz, though, have not won as often as their differential suggests. Their differential says this team will win at least 15 more games than the Blazers by the end of the season. But in mid-December, the Blazers are only ½ games behind Utah.
Why is Utah underperforming? And why are the Blazers out performing their differential?
Those are great questions. And I don’t know the answers. But I can tell you which players are responsible for the differential we observe. So let’s turn to that story.
Wins Produced and WP48 for the Northwest Division
This story begins with Table One, where the Wins Produced for each team is reported.
Table One: A Team-by-Team Review of the Northwest Division
As I have done with each previous division review, here are a few brief notes on each team.
Utah Jazz
The Jazz are led by four players – Carlos Boozer, Andrei Kirilenko, Deron Williams, and Ronnie Brewer. This quartet is on pace to produce 57 wins. Unfortunately, the rest of this team is on pace to produce less than nothing. The drop-off with respect to Mehmut Okur is most surprising. Relative to last year, Okur is rebounding less and having problems hitting shots. If Okur (or Paul Millsap or Matt Harpring) regains the form we saw last year, the Jazz might be able to contend with the Best in the West. If not, a repeat of last year – when this team was not quite as good as the very best in the Conference — may be the story when the season is over.
The Denver Nuggets
The good news in Denver is the improved play of Marcus Camby and Allen Iverson. Each of the past four seasons Camby has surpassed the 0.300 mark in WP48 [Wins Produced per 48 minutes]. Average is 0.100, so going past the 0.300 mark is amazing. Well, this year Camby’s WP48 stands at 0.461. Not sure what word to use for this level. Needless to say, Camby ranks among the league leaders in WP48 and Wins Produced.
Camby is not the only player who has improved. In 2004-05, Iverson posted a career high 0.152 WP48 for the 76ers. This year his mark stands at 0.160, so the Answer is playing as well as he ever has.
The same story cannot be said for Carmelo Anthony. Anthony’s WP48 stands at 0.048, his lowest mark since 2004-05. One issue is Melo’s shots from beyond the arc. Last year he was below average but still launched 2.3 bombs per game. This year he is still below average but now taking three per contest. In addition, he’s not shooting as well within the arc. For Anthony this decline in shooting efficiency is a real problem. Other than scoring, Anthony doesn’t do anything really well. And if he’s not scoring efficiently, then Melo is just not helping much.
Portland Trail Blazers
When Greg Oden was lost for the season it was expected Portland would struggle. But after 25 games this team is above the 0.500 mark. How is this possible?
As noted, efficiency differential says that this record is a bit of an illusion. But even if this team was winning as much as their differential suggests, Portland would still be exceeding expectations. At least, I didn’t expect this team to win 36 games this year.
When we look at Wins Produced we see two players leading this team’s charge back to respectability. And those players are Brandon Roy and Joel Przybilla. Roy was the Rookie of the Year last year, so his play is not a surprise. Przybilla, though, is one of those players who doesn’t look good when we consider traditional scoring based metrics (points per game, NBA Efficiency, Player Efficiency Rating). But when we consider his above average rates of rebounds, blocked shots, and shooting efficiency, we can understand why his WP48 is so high.
I would add that Przybilla posted WP48 marks of 0.243 and 0.186 in 2004-05 and 2005-06. So he has been above average before (in other words, Henry Abbott was correct to note his contribution).
Seattle Super Sonics
Kevin Durant – the supposed savior of this franchise (wherever it might eventually be located) – has already been the subject of numerous posts.
July 12, 2007: Looking Back at the NBA Draft, Part Two
July 7, 2007: Disappointing Durant
July 21, 2007: Durant Disappoints Again
October 31, 2007: Will Kevin Durant Be the Best Rookie?
November 16, 2007: Choosing the Best Rookie in November
November 27, 2007: Evaluating Future Stars in Baseball and Basketball
November 28, 2007: The Top Rookies, Again
November 29, 2007: Re-Hashing Durant, Melo, and Stack
What can we say now? When we look at the numbers after 25 games, we can say that Durant is getting better. Over all, though, he’s well below average.
I think such a statement very much needs to be clarified. People seem to be confusing these two questions:
1. How productive has a player been?
2. How productive will a player be in the future?
With respect to the first question, Durant has been quite bad. With respect to the second, he may indeed morph into a superstar. Certainly many people see that potential. But if we only look at what he has done so far, we would hold off on the word superstar.
As for the rest of the Sonics, this team has three above average big men – Kurt Thomas, Nick Collison, and Chris Wilcox – and then nothing else. Thomas and Collison have not played the entire season. If these two do play consistently the rest of the season, one can expect the Sonics to finish a bit beyond the 22 wins currently projected.
Minnesota Timberwolves
This team was the subject of the following post a few days ago.
T-Wolves Minus Garnett Equals the Worst NBA Team
There isn’t much to add to this column, other than to note that after 23 games Al Jefferson is the only above average player on the roster. That’s it.
Ranking All Players
Before I close this post, here are all 70 players ranked in terms of Projected Wins Produced.
Table Two: The Players of the Northwest Division
Again one should note how simple a projection Table Two presents. This analysis simply takes the minutes played so far and projects to 82 games. So the projection for Nene – or anyone else who hasn’t played all season — is incorrect. Still, for most players it gives us a sense where they are going.
What stands out in Table Two is the dominance of the Utah Jazz. Four of the top six players in the division are in Utah. They currently have the top point guard, shooting guard, small forward, and power forward. Unfortunately, they also have two of the least productive centers. And as noted, as long as that continues Utah – despite an amazing starting line-up – may fall short of contending with San Antonio and Phoenix.
– DJ
Our research on the NBA was summarized HERE.
The Technical Notes at wagesofwins.com provides substantially more information on the published research behind Wins Produced and Win Score
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Introducing PAWSmin — and a Defense of Box Score Statistics
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.
dustin
December 19, 2007
Dave, I am calculating the mean and median wp48 for each division you have reviewed so far, and the average wp48 is not .100 . What am I doing wrong? I’m pretty sure my math is correct since I’m using excel (<3 computers)
richmindseed
December 19, 2007
dustin,
dave’ll correct me if i’m wrong, but there’s no reason the mean in a given division has to be .100. after all, it’s possible that the atlantic division has all the really good players, and the pacific division has all the really bad players (not saying that’s the case, just giving an example). as such, wait to calculate until all the season previews are in…if you’re still seeing an average that’s off, we have a potential issue.
additionally, it’s probably the case that dave’s using the averages from last year (or the last X years) to calculate WP48. since it’s possible that players are playing differently this year, some error is expectable. for instance, if FG% in the league as a whole was up 10% this year, we’d see an average WP48 across the league that was higher than .100. Again, not saying this is actually happening, but some variation in average values from one season to next is somewhat expected, and might contribute to the error you’re seeing. a simple t-test would tell us what range of error to expect in a given value. this could actually answer those questions about “years the league is particularly strong,” since you could point to years that the average FG%
was X SD’s above the mean, say…
Thinking about it, I’m not actually 100% sure that second factor exists – after all, we’d then have years where more/less wins were produced by NBA players than there were wins in the NBA. is there a correction for this built into the model that i’m not seeing, or is that simply an accepted consequence to using long-run averages? if there IS no correction, then that provides another way to answer those questions – just look at years when the total WP across the league was particularly high, and those are probably years the league was really strong…
The Franchise
December 19, 2007
The median player rating will also generally be below .100 because it’s an average rating on a per minute basis, and better players tend to be on the floor a disproportionately high amount.
mrparker
December 19, 2007
richmindsheed,
I’m pretty sure berri uses averages over years of data so that shouldn’t be an issue.
franchise,
dont some really bad players see a bunch of minutes while some really good players see less. this might even out…though I’m not sure
dustin
December 19, 2007
Let me rephrase that. No division reviewed so far has a mean/median wp48 of .1 . All 4 divisions reviewed so far have a wp48 of less than .1 . I know the southwest division is good, but I don’t think it is good enough to raise the aggregate mean/median of the 4 other divisions from .04 to .1
dberri
December 19, 2007
I checked on this and Dustin is correct. But that is because the winning percentage in each division is less than 0.500. The Atlantic is close to 0.500 and the weighted WP48 (weighted by minutes played) is o.099.
When we look at the other two divisions this will change.
About the average performance I am using… I am using last years averages.
Jason
December 19, 2007
Since better players tend to play more, the median WP48 will almost always be below .100.
As for the average, as noted above there’s no reason why the average in any division will be .100, though the average production across the league will be .100, weighted for time on the court. I do not know how you are calculating your average. Are you simply adding up every wp48 and dividing by the number of players or are you weighting it by playing time? If you do the former, a couple of guys who have only played a few minutes and have very low WP48 will drag the ‘average’ down, though this isn’t the right way to compute the average in terms of on-court productivity across the league.
dustin
December 19, 2007
Ah ok I need to do a weighted average. Thanks all.
On another side note, what percentage of games does a team play in division/conference? Does this affect a player’s wp48 if they are playing an above average opponent more often than not?
Joseph
December 19, 2007
As always, this is a good post.
mrparker
December 19, 2007
dj,
Do u always lose last years. I was under the impression that you were using the averages in a chart I found on wages of wins.
nice post as always
Panda Bear
December 19, 2007
Celtics lost 3rd game tonight. At least this will discourage all the “best team that ever lived” talk.
Panda Bear
December 19, 2007
Wow, Knicks won by 18 against the Cavs. Isiah’s motivational speech worked!!!
Owen
December 19, 2007
David Lee had a monster monster first half, winscore of 13.5 in the first…
Going to be to keep him on the bench. Isaiah could swing public opinion back in his favor quite a bit by giving him the starting job….
Panda Bear
December 19, 2007
Isaiah would have to not only give Lee the starting job but clone him a couple times and turn Marbury into Nash or Kidd just for the Knicks to get back to .500.