Although the summer is not quite over, Chris Mannix of CNNSI.com has decided to rank the teams in the NBA right now. And coming in at the #8 spot is…. the Washington Wizards?
Last season the Wizards won 19 games and finished with a -7.9 efficiency differential (offensive efficiency minus defensive efficiency). This differential ranked 28th in the league and 0nly the Sacramento Kings and LA Clippers finished lower.
Looking at the top of the 2008-09 differential rankings we see that the 8th best team was the Denver Nuggets, a team that posted a 3.5 differential and won 54 games. The Wizards franchise has not fielded a team that was this good since 1978-79. At that time the team was named the Bullets and Washington was competing in the NBA Finals. Since that season, the best finish by a Washington team came in 2005-06 when the Wizards posted a 1.9 differential and won 42 games.* In sum, it looks like Mannix will only be correct if this is the best Washington team in 30 years.
Again, last year this team was very bad. And when we look at the players on this team, we can see why.
Table One: The Washington Wizards in 2008-09
An average player will post a 0.100 WP48 [Wins Produced per 48 minutes]. Last season the Wizards employed only four players – Antawn Jamison, Caron Butler, Dominic McGuire, and Gilbert Arenas – who were above average. And Arenas only played 63 minutes. Of the thirteen remaining players employed, eight were producing in the negative range. Consequently, we should not be surprised this team played so poorly.
What would it take for this team to play as well as the Nuggets last season? Again, the Nuggets won 54 games and posted a 3.5 differential. This means the Wizards would have to win 35 more games and improve its differential by 11.4. To put this improvement in perspective, here is the list of all teams since 1973-74 who have improved by 30 wins.
- Boston (2007-08): 42 more wins, 14.5 increase in differential
- San Antonio (1997-98): 36 more wins, 13.1 increase in differential
- San Antonio (1989-90): 35 more wins, 10.1 increase in differential
- Phoenix (2004-05): 33 more wins, 11.1 increase in differential
- Boston (1979-80): 32 more wins, 11.9 increase in differential
What do these five teams have in common? All of these teams added a very productive player. The Celtics added Larry Bird in 1979 and Kevin Garnett in 2007. The Spurs added David Robinson in 1989 and Tim Duncan in 1997. And the Suns added Steve Nash in 2004.
The Wizards traded away their lottery pick for Mike Miller and Randy Foye. Yes, Gilbert Arenas is coming back from injury. But even with a healthy Gilbert Arenas in the past, this team didn’t win 50 games. So what is Mannix thinking?
Well, maybe he looked at everyone who was expected to play for Washington in 2009-10. Here is the expected first and second string for Washington (WP48 = Wins Produced per 48 minutes):
First String
PG – Gilbert Arenas [2006-07]: 39.8 minutes per game, 10.7 Wins Produced, 0.174 WP48
SG – Mike Miller: 32.3 minutes per game, 13.9 Wins Produced, 0.282 WP48
SF – Caron Butler: 38.6 minutes per game, 8.1 Wins Produced, 0.150 WP48
PF – Antawn Jamison: 38.2 minutes per game, 10.1 Wins Produced, 0.157 WP48
C – Brendan Haywood [2007-08]: 27.9 minutes per game, 5.4 Wins Produced, 0.116 WP48
Second String
PG – Randy Foye: 35.6 minutes per game, 1.3 Wins Produced, 0.026 WP48
SG – Nick Young: 22.4 minutes per game, -1.2 Wins Produced, -0.032 WP48
SF – Dominic McGuire: 26.2 minutes per game, 7.5 Wins Produced, 0.174 WP48
PF – Andray Blatche: 24.0 minutes per game, -1.5 Wins Produced, -0.042 WP48
C – JaVale McGee: 15.2 minutes per game, 0.6 Wins Produced, 0.025 WP48
Adding together the Wins Produced of these players we get 54.8 wins. The minutes of these players, though, summed to 22,456. For these minutes to be played, the Wizards would have to average about 274 minutes per game. Assuming the Wizards are not playing a large number of overtime games, someone is going to have to play fewer minutes, and therefore, 54 wins is not quite right (assuming these players offer the same level of production). Nevertheless, it does look like these ten players can win between 45 and 50 games. And with a bit of improvement here and there, maybe this edition of the Wizards can indeed post numbers similar to what the Nuggets did last year.
Of course, we are assuming
- Arenas returns to what he was two years ago.
- Haywood returns to what he was a year ago.
- Miller maintains the productivity we saw last year in Minnesota
So we are assuming quite a bit and therefore, there is no guarantee that the Wizards will approach 50 wins.
That being said, it does look like what Mannix argued – as unbelievable as it appeared at first glance – is at least possible. Without adding a future Hall-of-Fame player, the Wizards might improve by 30 games in the standings. And Washington fans might see a team that is better than any edition of this franchise since Jimmy Carter was president.
*- the Wizards did win 45 games in 2004-05 but the team’s differential was only -0.3.
– 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 
Anon
August 27, 2009
According to the chart, Arenas played 63 minutes, not 288 (which would be impressive in 2 games).
Anyways, good post. I hope Arenas is finally healthy because he’s really fun to watch.
dberri
August 27, 2009
Thanks. I fixed the post.
Man of Steele
August 27, 2009
Dr. Berri,
In keeping with your theme, perhaps the Wizards will be gaining an extremely productive player, all told. Adding Mike Miller and getting Arenas would essentially be the same as adding a player like Garnett or Duncan, right?
dberri
August 27, 2009
Man of Steele,
I think I was trying to say something like that, but it never quite came out. Thanks.
Rob O'Malley
August 27, 2009
Also Butler Jamison and Haywood could return to their 07-08 numbers.
Butler = .217
Jamison = .180
Haywood = .153
:-D
Also McGee could improve on his rookie season. These are additional ways in which Washington can improve. :-D Big Wizards fan here.
However, I would rather have had Ricky Rubios rights than Miller and Foye. Additionally, I would like to have won the lottery and Blake Griffin. :-P Because unless we pull the best possible scenario season we’re really mediocre, again.
Truth About It
August 28, 2009
Nice work … although I kinda wish you would have at least mentioned Fabrico Oberto’s name.
Sure he will be a role player, but also a player that Roger Mason Jr. calls a “sleeper” signing.
Oberto can do all those little things that can lead to wins … but then again, most of those little things don’t show up in the stat book.
Nick
August 28, 2009
Quick suggestion for a future post.
Perhaps you can investigate the Grizzlies w/ Randolph.
Every thing I’ve been reading has stated that adding Zach Randolph makes them even worse.
However, I know you had an Iverson post to Memphis before, but adding Iverson and Randolph, helps push that starting line-up to a mediocre level no? (Rudy Gay drags it down)
TRad
August 28, 2009
Actually Spurs added Duncan _and_ Robinson in 1997. The season before DRob has played only 147 minutes. And he was not very bad in 1997/98.
Oren
August 28, 2009
When you consider that it’s likely that the second team players with the exception of Foye are the ones whose minutes are going to be cut, and that those are the players with the lowest scores, fifty wins is quite likely according to your model.
P-Dawg
August 28, 2009
This is another in a series of arresting, compelling posts, but I’m still waiting for the final 2009-2010 NBA team analysis. Also, the folks here in Boston seem to believe that, assuming Kevin Garnett stays healthy, another title is all but assured with the addition of Rasheed Wallace, the resigning of Glen Davis and the addition of Marquis Daniels and Sheldon Williams. I say that, other than Rondo, the good players are too old and the young players aren’t too good. Will you weigh in and counter-balance some of the inane sports writing we get here in the Hub?
AndNone
August 28, 2009
Awesome post .. thanks for the optimism
Bri-Bri
August 28, 2009
I have to ask: in what alternate reality is Dominic McGuire, he of the .483 TS%, 21.4 TO%, and the Bowenesque 6.2 Pts per 36 minutes, an “above-average” player? I know you’re no huge fan of PER, but his 9.7 career rating is not just below average, it’s below replacement level. How is it even possible that WP sees him as above average? Other than good blocks and d-rebs, everything about his statistical profile is really horrible.
KSean
August 28, 2009
@Bri-Bri
D-Mac’s impact on the court doesn’t show up in stats. But if you watch his game closely, and if you look at Shane Battier as the mold for a “stat-less wonder,” you’ll realize that most people who follow the Wizards hold him in high regard. Now, understand that he is young, raw, and makes mistakes, but he is consistent, technically sound, and has a strong impact on the overall defense of the team whenever he is on the court. No, I don’t have numbers right now, but maybe someone on here wants to back me up.
Bri-Bri
August 28, 2009
A) But Prof. Berri’s Wins Produced only uses box score stats, so if his impact doesn’t show up in stats that’s kinda beside the point. I’m wondering how, using only his terrible box score stats, he shows up as > avg.
B) If he was doing something Battier-like that didn’t show up, wouldn’t it show up in the adjusted plus minus? But McGuire’s 1-year and 2-year scores are below average and very bad, respectively:
http://basketballvalue.com/teamplayers.php?team=WAS&year=2008-2009
But like I say, that’s really beside the point. The point is how a guy with awful stats can show up as >avg.
dberri
August 28, 2009
Bri-Bri,
Apparently he doesn’t have “awful” stats. More specifically…. yes, McGuire can’t shoot. But he doesn’t shoot much. So the negative impact of his poor shooting is reduced by the fact he doesn’t shoot much (by the way, people who don’t shoot will tend to have very low PER values).
What he does do is rebound and blocked shots. And those stats are valuable.
Can you have a team of McGuires? No, but you can’t have a team of all shooters either.
Tom Mandel
August 29, 2009
Dave — you have JaVale McGee’s numbers wrong.
Your last indication of an average NBA center’s numbers was 10.8 — using your fudge factor for a center, and your conversion formula, that produces a .104 WP rate — just about average and close enough for jazz (or economics).
Draftexpress has him with a 9.1 WS40 minutes. I.e. 10.9 in 48 minutes. Doing the math myself, I get 10.88. Obviously, that’s very slightly above average.
Dividing, we get .227. Using your fudge factor and conversion formula, young JaVale winds up at .107.
You have him at .025 — what gives?
Tom Mandel
August 29, 2009
Follow up note on JaVale McGee. Per 48 minutes he contributes:
11.2 net possessions (vs. 10.8 for an average NBA center)
-1.2 via blocks/assists/pf (vs. -.65 for an average NBA center)
.8 by scoring (20.5 points on 16.3 FGAs and 6.8 FTAs) (vs. .65 for an average NBA center [(17.7 points on 14.3 FGAs and 5.5 FTAs])
dberri
August 29, 2009
Tom,
I am using the average values from 2008-09. Centers last year were unusually good, so McGee looks worse than you might expect.
I am also doing the complete WP48 calculation, and that can make some difference also.
Tom Mandel
August 29, 2009
Note that I’ve culled your average productivity numbers for every position from a variety of tables you’ve posted comparing sets of players to averages.
It would be *great* if you’d post a table of averages for each position — and link to it in the header of the blog!
Tom Mandel
August 29, 2009
Dave — interesting that centers had a good year last year. What position got worse?
Makes it all the more desirable to have *annual* averages for every position.
dberri
August 29, 2009
For this past year the WS48 averages are as follows:
C: 12.54
PF: 10.30
SF: 7.87
SG: 6.33
PG: 6.83
Hope that helps.
MrParker
August 29, 2009
Dberri,
I was wondering if you ever had a chance to look at basketball-statistics.com data set on the effects of 3 point shooting from the point guard position. That data points to pgs who shoot 3 pointers extremely well having a tremendous effect on their teams offenses. A similar effect was seen from sg’s.
If the Wizards employ a backcourt of Arenas (career 35%)and Miller(career 40%) signs point them being able to replicate their offensive rating of 110.7(3rd in the league) from 2006-2007 when they employed a backcourt of Arenas(35% that season) and Stevenson(40% that season).
Furthermore by having Arenas replace Stevenson at pg(better career wp48) and Miller replacing Arenas at sg(better career wp48) the Wizards should see an improvement in effeciency differential over that year.
I believe the Cleveland Cavaliers are a good example of the difference a good 3p% backcourt can have on a teams offense. In 07-08 playoffs they played the champs the closest of almost any team employing a backcourt of D. West(37.8% career) and D. Gibson(career 41.2%). In the regular season(-.4 eff diff) Devin Brown, Damon Jones, Sasha Pavlovic, and Larry Hughes saw significant minutes. Of those only Damon Jones was an effective shooter(above 35%).
Flash forward to the next season and with a backcourt of D. West(37.8 %career) and (38.6% career) the Cavaliers jumped to 4th in offense after being 19th the previous season.
The Wizards could see a similar offensive improvement. Furthermore those players who are bringing the offensive change seem to be defensive improvements as well.
Tom Mandel
August 29, 2009
Dave — off the top of my head (not yet having calculated), it looks like your new averages will require a change in the formula (or fudge factors) for converting WS to WP48 — otherwise, average guys won’t come out around .100.
Put another way — is an average Center, your new average I mean, a .100 WP48 player?
dberri
August 29, 2009
Tom,
It shouldn’t be all that different. A player who is average will have a PAWSMin of zero. So his expected WP48 from the formula is about 0.100. That is the same result you get regardless of what averages you use.
Tom Mandel
August 29, 2009
Dave — that’s my point, i.e.:
Center — 12.54/48 = .261
Subtract your Center adjustment of .225 = .036
Use your conversion formula: .104+(1.621 x .036) = .162 — *way* better than “average.”
Where have I gone wrong?
Tom Mandel
August 29, 2009
Moreover, given that “wins” is a constant number from season to season, if Centers get “better” (produce more wins), then another position must get “worse” (produce fewer wins) — no?
Hence my question a few comments above: which position got worse?
dberri
August 29, 2009
Tom,
If a center is at 12.54 (or average), then it works out this way…
PAWSMIN = 0.104 + (player performance minus average at position)
The latter term is 0 (because the player is average). So you are left with 0.104.
Tom Mandel
August 30, 2009
Dave — I must be being dense. I would have divided 12.54 by 48 to get the PAWSMIN. Obviously, that doesn’t yield .104.
Would you mind terribly noting the process here in more detail — and telling me what I did *wrong* in my version just above your last comment?
dberri
August 30, 2009
Tom,
Here is the formula for going from PAWSMIN to WP48.
WP48 = 0.104 + 1.621*PAWSMIN
PAWSMIN is calculated by subtracting the average per-minute Win Score at a player’s position from a player’s per-minute Win Score.
So let’s say a center has a WS48 of 12.54. Per-minute, this mark is 0.261. As I noted, this is the average mark this year. So if you subtract the average per-minute Win Score at the center position from this player’s mark (i.e. you subtract 0.261 from 0.261), you get zero.
Going back to the formula…
WP48 = 0.104 + 0
Or…
WP48 = 0.104
In sum, regardless of the average value in a give year, the calculation is the same.
Remember, this is just an approximation. Actually calculating WP48 involves a bit more work.
Phil
August 30, 2009
Tom,
There is good reason that WP could fluctuate from season to season based upon the box score itself.
Percentage of shots that are credited with an assist, for instance, change each year.
Ditto percentage of points that come from three pointers; a team that shoots more three pointers generates more rebounding opportunites, but may still score at the same rate.
The proportion of turnovers that are credited as a steal also varies.
The subjectivity in scorekeeping and subtle changes in the way the game is played year-on-year make minor changes.
Italian Stallion
August 30, 2009
Dberri,
“So the negative impact of his poor shooting is reduced by the fact he doesn’t shoot much (by the way, people who don’t shoot will tend to have very low PER values).
What he does do is rebound and blocked shots. And those stats are valuable.
Can you have a team of McGuires? No, but you can’t have a team of all shooters either.”
I think this is the question many people still have.
If you do have a player like McGuire on the team, then we know the other 4 players on the court have to take more shots on average. At the same time, the defense can easily make defensive adjustments to take advantage of the fact that McGuire is not an average outside offensive threat.
So does all that lead to some negative impact on the other four players that is not being picked up in the box score or other statistical measurements of McGuire?
I think the answer almost HAS TO be YES.
Perhaps in the grand scheme of things it is not huge and also quite difficult to measure, but it almost has to be there in some lower level detailed cases.
I see too many players with woeful outside shooting skills that are otherwise terrific when measured on their other stats that have a tough time getting minutes.
I refuse to believe that all these great coaches, great players etc… that get to see these players every night in practice, play with them, see them in games etc… are simply totally misunderstanding their contribution. I think they are also seeing and feeling some of the negative things that the stats are not picking up.
There may be a place in the middle where reality lies.
cells
August 30, 2009
are there any metrics that are used for coaches? the wizards are banking that Flip Saunders is going to create more wins as well. how does coaching factor into this?
Tom Mandel
August 30, 2009
Dave — got it! The Position Adjustments change each year.
Michael
August 31, 2009
cells I believe with the exception of Phil Jackson coaches don’t tend to make a difference.
lace wigs
August 31, 2009
the wizards are finally coming into their own. take any team, pump millions of dollars into it and they eventually will win games. but top 10 this year? NAH.