The Chicago Bulls finished the 2008-09 season by taking the Boston Celtics – the defending NBA champions — to seven games in the first round of the playoffs. Such a performance likely gave fans of the Bulls hope for the future.
Since the Chicago-Boston series ended, though, little has happened. In the draft Chicago added James Johnson and Taj Gibson. Both Johnson and Gibson appear to play power forward, or the same position as Tyrus Thomas (and perhaps Joakim Noah). So it’s unclear how much either rookie will play (or if they are an upgrade over what the Bulls currently have).
Gordon vs. Pargo
The only other move the Bulls made was the signing of Jannero Pargo, a move made at the same time Ben Gordon signed with the Detroit Pistons. Both Gordon and Pargo are undersized shooting guards who like to launch shots from beyond the arc. There are, though, substantial differences. On the positive side, Pargo is far cheaper. Unfortunately, the Bulls are getting what they pay for. In terms of on-court productivity, Pargo is very much a downgrade.
As has been noted in the past in this forum (and also in The Wages of Wins), Gordon – relative to an average shooting guard – is not very good. In fact, Gordon has never posted a WP48 [Wins Produced per 48 minutes] mark that was above average. WP48, though, compares a player to the average at his position. If Gordon could be compared to Pargo, he would be a superstar.
Table One reports the productivity of each player the Chicago Bulls employed last season. As one can see, Gordon’s WP48 was 0.097 last season. Pargo didn’t play in the NBA last season, but his career WP48 is -0.018. So Gordon was 0.115 better than Pargo’s career average; or if Pargo was an average player [WP48 of 0.100], Gordon would have a 0.215 WP48 [i.e. Gordon would be a star].
Table One: The Chicago Bulls in 2008-09
To put these numbers in perspective, Gordon produced 6.1 wins for the Bulls last year. Pargo’s career mark indicates he would have produced -1.1 wins in Gordon’s minutes. So replacing Gordon with Pargo would have cost the Bulls 7.2 wins and their opening round match-up with the Celtics.
Looking for Hope
The comparison between Gordon and Pargo suggests the Bulls are going to decline in 2009-10. When we consider the team’s current depth chart (taken from ESPN.com), though, there is reason for optimism (2008-09 WP48 reported)
First String
PG: Derrick Rose, 0.084 WP48
SG: John Salmons, 0.098 WP48 (for entire 2008-09 season)
SF: Luol Deng, 0.121 WP48
PF: Joakim Noah, 0.208 WP48
C: Brad Miller, 0.122 WP48 (for entire 2008-09 season)
Second String
PG: Kirk Hinrich, 0.104 WP48
SG: Jannero Pargo, –0.043 WP48 in 2007-08
PF: Tyrus Thomas, 0.120 WP48
Looking at these numbers we see that five players who will be part of the rotation next year were above average in 2008-09. And Salmons and Rose were very close to average. So Pargo is the only weak link on the team. It appears, though, that Salmons is the primary replacement for Gordon; so Pargo’s negative impact will be mitigated.
Looking at just these numbers – and assuming the remaining roster is not a detriment to the team (the team will employ more than the eight players listed above) – this team is capable of winning at least half their games. It’s possible, though, for the Bulls to do better.
For example, Luol Deng posted the following numbers prior to this season.
2004-05: 5.0 Wins Produced, 0.148 WP48
2005-06: 10.7 Wins Produced, 0.198 WP48
2006-07: 14.7 Wins Produced, 0.230 WP48
2007-08: 5.9 Wins Produced, 0.133 WP48
The first three years of Deng’s career we see the standard pattern in young players. Each year Deng got better. In 2006-07, though, Deng missed 2o games and his per-minute performance declined. Deng also missed games last year. If Deng could be healthy and productive, though, the Bulls could expect about five additional wins.
And then there is the case of Derrick Rose. The media and coaches stated that Rose was the top rookie last season. The Wins Produced story, though, suggests Rose was only average. If Rose continues as an average player (and Deng doesn’t return to form), the Bulls will probably be close to an average team. But what if Rose follows the career path of LeBron James and Kevin Durant? Both LeBron and Durant struggled their first year and then developed into above average performers. If Rose follows the same path, then the prospects of the Bulls improve dramatically.
To illustrate, Rose produced about five wins last season as an essentially average point guard and the team won 41 games. Here is what happens if Rose plays better:
WP48 = 0.150, Wins Produced are about 9.0, Bulls win about 45 games
WP48 = 0.200, Wins Produced are about 12.0, Bulls win about 48 games
WP48 = 0.250, Wins Produced are about 15.0, Bulls win about 51 games
WP48 = 0.300, Wins Produced are about 18.0, Bulls win about 54 games
In sum, if Rose becomes the player people think he was last year, the Bulls can be above average. If Deng returns to form, the team is even better. And if the Bulls sign Dwyane Wade in 2010 (a rumor I have seen), the Chicago Bulls could be a title contender in 2011.
Of course, all these “ifs” might not happen. But contrary to what I said last March, it’s possible (not a guarantee, just a possibility) that Barack Obama can legitimately invite the Bulls to the White House before his first term ends. And the signing of Pargo is not enough to derail this hope.
– 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 
romalley
September 9, 2009
I would be interesting to see Derrick Rose’s second half numbers. Kevin Durant’s numbers improved pretty significantly as his rookie season went on and he continued that trend through the following year. I know Michael Beasleys improved in the second half of his season too and hopefully he can continue to get better. I’m curious to look at Rose’s numbers.
Ronen
September 10, 2009
Dear Professor Berri,
I am a professor of computer science at a university in Israel. (This information is irrelevant as I am not an expert in economoics, still I hope it is more likely that you read and respond to this post given this information…).
I have read the “Wages of wins” and am following this blog for quite a while. Overall, I think you are doing very good work although I have reservations regarding the justifications you have of your claim that the sum of past win-scores of individual players is a good way to predict the overall win-score of a team.
In a fascinating series of posts you are now trying to predict the wins (or win-score or wins-produced) of teams using the following methodology:
1. Examine the roster and estimate the number of minutes each player will play.
2. Use the production of a player in the past season to predict his productivity in the next season.
3. For players where the past season seems to be a bad indicator, speculate on possible values for production.
I think it may be very easy to test how this method works on past data. (As the method is so simple, I think no one can raise the objection that this method was tailored to work well on past data). Have you perfomed such tests?
I suggest the following specific experiment to try and predict the win-scores of teams in the past season based on production of players in the season before that. Specifically, I suggest to:
1. Use the actual minutes played by players last season. That way, the need to forecast the number of minutes played is removed.
2. Use the win-score (without position adjustments) of the players two years ago.
3. In cases where there is not enough data from two years ago (say, too few minutes played), use the actual win-score from last season. That way, the need to forecast the productivity of a player that does not have sufficient statistics is removed.
It will be very interesting to see the results of such an experiment. Personally, I expect that your system will do very well and this will serve as justification that the way you are using win-score (or wins-produced) in your current analysis is justified.
Note that this is different than saying that “win-scores of players are attributes of the players” as this experiment really tests the actual method you are using.
As an academic I know how much we hate being “suggested” to do things, so even if you can’t find the time (or the time of a student) to do this, I think that if you just agree that this is a good idea someone else following this blog may try to do it.
Best wishes,
Ronen
Austin
September 10, 2009
Well, I can’t speak for WP48, but looking at Rose’s pre and post- all star break splits:
http://www.basketball-reference.com/players/r/rosede01/splits/2009/
He improved somewhat at rebounding. Nothing else (from my brief glance) seemed to change much.
Joe
September 10, 2009
Austin/rom,
His est WP48 for the months are as follows…(assuming I didn’t mess it up)
October 0.0630125
November 0.153024038
December 0.035414736
January 0.048146711
February 0.139978225
March 0.125678851
April 0.245694146
total 0.10962332
So he did seem to get better as the season went along. He was “above” his season average est WP48 each of the last 3(well, 2 1/2) months. Remember that April(6 games) and October(2 games) are not full NBA months.
If I were a Bulls fan(trust me, I’m not) I’d just mentally throw away his December and January probably. I might even assume his April was the player he will be soon enough.
Tball
September 10, 2009
Romalley,
Dig through the archives on this site for articles on Durant about 18 months ago, possibly in relation to the ROY discussion. I recall Berri indicated that Durant did not post better WP48 numbers as his rookie season progressed, despite media indication his play improved during the season. Read the article(s) to get the explanation.
Tim
September 10, 2009
Thanks for turning to the Bulls.
I think there are other players besides Deng and Rose who could play better than last year. Noah and Thomas are still young enough to improve, and maybe Miller and Salmons will benefit from a whole year with the team. Even the coaching may improve in Del Negro’s second year. And last year’s playoff performance may have encouraged the players to dedicate themselves during the off season.
On the other hand, if no one improves the Bulls are in big trouble, and they have a killer fall schedule. This season is full of “ifs,” very unpredictable despite the fact that few players have changed since the end of last season.
As a fan, I’m hopeful that all the upsides will pan out and the team will win 55 games, but I won’t be surprised if it all falls apart and the team wins 30 games, or if the ups and downs cancel out and they win 42 games and barely make the playoffs.
Next year will be interesting, as the Bulls have positioned themselves well for free agency, and also may sign Omer Asik. But they would have lots more options if this year’s team proves that it can win without Wade or Bosh or other free agent signings.
todd2
September 10, 2009
The Eastern Conference looks as if it will be stronger his year…could be difficult for the Bulls to maintain/improve.
D Fitzpatrick
October 1, 2009
Hi, Just reading wayne winston comments on gordon on his blog. He is the mavericks stat guru. He calls gordon one of the top 40 players in the league, and worth the money the pistons gave him. Just wondering what you think of the big contradiction here. The only thing I know is I am not smart enough to figure it out. Thanks