Nearly fifty years ago the Warriors were located in Philadelphia. Had this franchise remained in the East, the Warriors would have enjoyed home-court advantage in the first round of the playoffs this past season. In the West, though, this team’s 48 victories left them out of the post-season. Consequently, if we define success and failure in terms of post-season outcomes, the Warriors were the best failures in the NBA since the Phoenix Suns of 1971-72 (the Suns won 49 games that season and missed the eight team playoff field).
If we ignore the playoffs, though, we can see that the Warriors were a pretty good team. The Warriors ranked 12th in the NBA in wins and the team’s efficiency differential (offensive efficiency minus defensive efficiency) of 2.16 ranked 13th in the Association. In sum, if “good” is above average, the Warriors were “good” in 2007-08.
The Top Players in Golden State
And this leads us to wonder, why were the Warriors “good” last season? For an answer we turn to Table One.
Table One: The Golden State Warriors in 2007-08
Table One reports the Wins Produced of each player on this team, in both 2006-07 and 2007-08. Given what these players did two seasons ago, one would have expected Golden State to win about 35 games in 2007-08. The team’s efficiency differential this past season, though, was consistent with about 47 victories. When we look at the performance of the individual players, we see much of this improvement can be tied to the play of Monta Ellis and Andris Biedrins.
The play of Ellis is especially interesting. In 2006-07, Ellis was named the Most Improved Player in the NBA by the sports media. As I noted at the time (see Not Monta Ellis), Wins Produced told a very different story. Surprisingly (at least to me) Wins Produced did indicate that Ellis was the Most Improved Player in 2007-08 (see Now Monta Ellis and Monta Ellis This Time and Not LaMarcus Aldridge). In fact, when we look at the following player rankings (posted originally last April), we see that Ellis was the third most productive shooting guard in the NBA last season (behind Kobe Bryant and Manu Ginobili).
When we look over these rankings we see two other players from Golden State. Last season Andris Biedrins ranked 5th among all centers. And Baron Davis ranked 7th among point guards. When we look back at Table One, we see that the trio of Ellis-Davis-Biedrins produced 37.6 of the team’s 46.8 Wins Produced. Beyond this group the team really didn’t get much from anyone else.
Unfortunately for Golden State, the 2008-09 season will begin with only one member of this trio. Davis signed with the Clippers this past summer. And although Ellis did re-sign with the Warriors, he suffered an injury this past summer (apparently doing something he wasn’t supposed to be doing). As a result, only Biedrins will take the floor on opening night.
Trading Davis for Maggette
The Warriors did make some changes to their roster in an effort to at least maintain their team production of wins. The most significant was the addition of Corey Maggette. Last season Maggette played for the Clippers, so essentially Golden State traded Davis to LA for Maggette.
When we look at this “trade”, we see that in terms of production the Warriors came out a little bit behind. Here are the career numbers [Wins Produced and Wins Produced per 48 minutes {Wp48}] for each player
Corey Maggette
2007-08: 6.9, 0.132 WP48
2006-07: 8.4, 0.176 WP48
2005-06: 2.6, 0.130 WP48
2004-05: 7.7, 0.152 WP48
2003-04: 9.0, 0.164 WP48
2002-03: 3.6, 0.087 WP48
2001-02: 1.6, 0.046 WP48
2000-01: 5.7, 0.200 WP48
1999-oo: 3.8, 0.134 WP48
Career: 49.2, 0.138
Baron Davis
2007-08: 11.8, 0.177 WP48
2006-07: 9.7, 0.209 WP48
2005-06: 6.5, 0.158WP48
2004-05: 6.4, 0.195 WP48
2003-04: 7.7, 0.137 WP48
2002-03: 4.0, 0.101 WP48
2001-02: 9.7, 0.141 WP48
2000-01: 11.2, 0.169 WP48
1999-oo: 1.7, 0.055 WP48
Career: 68.8, 0.153 WP48
As the above numbers show, Davis has offered more than Maggette. This can be seen in both Wins Produced and WP48. If we look at salaries, though, we see that Maggette is cheaper. So in the end, this deal might be quite even.
Looking at 2008-09
Even if Maggette could come in and provide the same number of wins that Davis produces, the injury to Ellis still leaves this team looking for an additional source of victories early in the season.
In the draft the Warriors added Anthony Randolph in the lottery and Richard Hendrix in the second round. If we look at Win Score (see my last post), we see that one could argue that the order of these draft choices should have been reversed. Regardless of this point, both are on the roster. Judging by last year’s college performance, Hendrix is the most likely to make a contribution. Of course, he is also a second round draft choice and unlikely to play. In sum, it seems likely that the 2008 draft choices are not going to help much in 2008-09.
When we look back at Table One, we do see evidence that the 2007 lottery choice could help. Brandon Wright posted a 0.157 WP48 last year in very limited minutes. Such a mark is above average (average is 0.100) and so it is possible that Wright could help. Again, though, he has to play to make a contribution and we don’t have evidence yet that Don Nelson is going to put Wright on the floor.
In the end, looking at what these players did last year – and given how Nelson doesn’t like to play young players – it seems likely the Warriors are going to slip in the West. And that means, it seems quite likely Golden State will once again miss the playoffs. So in that sense, it seems likely that basketball in Golden State will look about the same come next April.
– DJ
The WoW Journal Comments Policy
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.
Wages of Wins Journal 
Vince Gagliano
October 2, 2008
You sure about that, Dave?
For one, Denver is *definitely* going to slip without Camby, although Nick Fazekas was signed in free agency. Dallas and Phoenix will bring in new head coaches, which, theorically, might not directly impact production, but the revamped chemistry and schemas involved might damage win production some. Yao and McGrady might slow down in Houston, although that might be balanced out a bit by Artest. San Antonio lost Brent Barry to said Rockets.
And even the potential spikes in production in LA and New Orleans help out the Warriors. More wins for them mean less for everyone else, which means more open season on the playoffs. To an extreme, we see this in the East races all the time.
And besides, Ellis’s loss does hurt, but it’s not like it was Biedrins out on the moped. If he were to get more minutes to compensate for Monta’s loss, Brendan Wright could somehow showcase his ability with more playing time, AND this team gets some lucky breaks with some other West teams, we might be looking at a fifth or sixth seed in the playoff race.
Maybe, just maybe, it would make up the half-dozen or so wins that this team would lose with Monta’s abscence.
William
October 2, 2008
Even with his new contract, Biedrins remains underpaid. He’ll be making “only” $9 million per year, less than Nene, Troy Murphy, and Bobby Simmons. He’s a classic example of why NBA GMs hurt their franchises by ignoring Dave Berri’s research. Instead of clearing room for Chris Bosh and Carmelo Anthony, I would have been setting up a $12 million/year offer for Biedrins.
Mountain
October 3, 2008
This would just be finetuning – if it worked and was valid- but would any set of slight weight adjustments for position help make team wins based on WP summed for all positions / all players even more accurate across the league? Just a thought. This would assume there are value trends by position not fully picked up by boxscore stats, such as defense or passing value beyond assists.
Mountain
October 3, 2008
By WP, a team of the 10th best at every position – J Nelson- A Parker= J Moon= R Wallace- B Miller would win 41 games themselves? How many wins does an average bench contribute?
A team of the 5th best at every position-Billups-Roy-Dunleavy-Marion-Biedrins would win 65 themselves plus what the bench provided?
A team of the 15th best at every position Harris-Wells-Maggette-Hayes-Przybilla
would win 32 themselves plus whatever bench provided?
Center is the deepest position in the league in terms of 10+ win guys and SG least?
Mountain
October 3, 2008
Players above 10th best and how far above 10th best appears very important for team wins by this metric and I can accept that and that is generally in line with conventional wisdom.
Mountain
October 3, 2008
Some blend of box-score and adjusted +/- seems preferable to me over purely one or the other.
Or the average position adjustment I threw out, expecting that player specific adjusted +/- isn’t going to be added to WP.
Dave Elkin
October 8, 2008
So Stephen Jackson didn’t help the Warriors win any games last year? Despite the turnaround after the 0-7 start? In fact his production was so bad it cost the team wins? And yet he got so much burn and at crunch time too…
This says to me something is wrong with your metric. Perhaps it does not deal well with players who play multiple positions during a game?
Actually, what I don’t believe is that you don’t address this at all in your article.