Prior to this past season Troy Murphy had played at least 2,000 minutes in four seasons in his NBA career.  And in these four seasons his Wins Produced and WP48 [Wins Produced per 48 minutes] were as follows:

2002-03: 9.8 Wins Produced, 0.187 WP48

2004-05: 7.9 Wins Produced, 0.159 WP48

2005-06: 9.6 Wins Produced, 0.183 WP48

2007-08: 6.6 Wins Produced, 0.151 WP48

Murphy’s average WP48 across these four seasons is 0.171, indicating that Murphy has been consistently a good, but not great, NBA player. 

Unlike what we see in football and baseball, basketball players tend to be very consistent.  Performance can change due to injury and age, but typically, what you have seen in the past is very similar to what you will see in the future.  Consequently, given what Murphy had done in the past, we should have expected Murphy to produce about nine wins in the 2,482 minutes he played in 2008-09.  Instead, Murphy produced 19.1 wins. 

Again, such a leap is unusual.  If we look at Basketball-Reference, we can why this leap happened.  Last season Murphy – relative to what he did in the past – grabbed more rebounds and hit more of his shots.  Of these two, his leap in shooting efficiency is the most dramatic.  Murphy’s career adjusted field goal percentage is 0.485.  Last year, though, his mark was 0.580.  From beyond the arc he converted on 45% of his shots (on a career high number of attempts).  Yes, Murphy in 2008-09 was a very rare big man who could both hit shots from beyond the arc and rebound at a very high rate.    

Despite Murphy’s performance, though, the Pacers missed the playoffs.  And when we look at Table One, we can see the problem.  Beyond Murphy, the Pacers only employed one other player who played at least 1,000 minutes and posted a WP48 mark that was above average. That one player was Danny Granger, and his mark of 0.119 was only slightly above average. 

Table One: The Indiana Pacers in 2008-09

Last spring, Granger was the subject of three different posts. 

Danny Granger is the Most Improved?

There is Not Much Difference Between Danny Granger and Kobe Bryant?

Bob Newhart, Danny Granger, and Group Therapy in Indiana

Of these three, the Bob Newhart post was my favorite.  Any of these, though, tells the same story: Although Granger is above average, he is not really as productive as is commonly believed.

Nevertheless, Granger was the second most productive player on the Pacers last year.  And when we look at the potential depth chart for 2009-10 (taken from ESPN.com and Yahoo.com), it seems possible that Granger will be the second most productive player again.

Potential First String

PG: T.J. Ford [4.3 Wins Produced, 0.092 WP48]

SG: Brandon Rush [0.8 Wins produced, 0.022 WP48]

SF: Danny Granger [6.0 Wins Produced, 0.119 WP48]

PF: Troy Murphy [19.1 Wins Produced, 0.369 WP48]

C: Jeff Foster [3.4 Wins Produced, 0.089 WP48]

Potential Second String

PG: Travis Diener [2.9 Wins Produced, 0.196 WP48]

SG: Dahntay Jones [1.7 Wins Produced, 0.057 WP48]

SF: Mike Dunleavy [10.6 Wins Produced, 0.173 WP48 in 2007-08]

PF: Tyler Hansbrough [Rookie]

C: Roy Hibbert [-1.6 Wins Produced, -0.078 WP48]

Of the ten players listed above, eight played for the Pacers last year. Across the past three seasons, Indiana has won 35, 36, and 36 games.  In other words, this team has been remarkably consistent.  And now they are bringing back essentially the same team.  Does this mean the Pacers should expect — and be happy — to win 36 games again?

One player who might make a difference is Mike Dunleavy.  Two years ago Dunleavy produced 10.6 wins and posted a 0.173 WP48.  Last years, though, Dunleavy was hurt and his production dropped to 0.1 wins.  If Dunleavy comes back this year – and returns to what we saw two years ago – the Pacers have a chance to make the playoffs in 2010.

Unfortunately, Dunleavy hasn’t yet come back and is currently listed as out “indefinitely.”  So even if Murphy maintains the productivity we saw last year (and I think that counts as an “if”), the consistency of the Pacers will likely continue.  In other words, without Dunleavy this team is probably headed for another 30-40 win season and another trip to the lottery.  And if Murphy returns to form…. well, the Pacers will probably get a much better lottery pick.

- 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

Wins Produced vs. Win Score

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.

Just saw the following in the comments and thought I would make pass it along to everyone (by the way, I won’t have time to play in a fantasy league this year):

I’ve just created an ESPN NBA.COM Fantasy League based on Win Score. You can check out the scoring settings for the league at http://games.espn.go.com/fba/tools/leaguesettings?leagueId=133748. I’m looking for other members to join the league before the season starts. If you’re interested, send your email address to reservoirgod at hotmail dot com.

Here’s something I wish I would have seen before posting yesterday’s comment on the Toronto Raptors.

Much has been made of NBA teams employing more and more statistical analysis in their decision-making process. It’s important to remember, though, that such analysis isn’t used by all teams (and even teams where this analysis happens, it’s not clear that the stats trump the other information teams employ).  To see this point, consider the following comment from Malcolm Gladwell. (HT: RaptorsHQ)

I remember once having a conversation with a top executive with the Toronto Raptors. I asked her about the stats revolution in basketball and she just kind of shrugged and said, “It’s interesting, and we look at those things, but you have to understand that for our purposes, it’s all [about] character.” The thing that separates players is that some have a work ethic, some don’t; some are coachable, some aren’t; some party all night, some go to bed early. From her standpoint, it’s all those intangibles.

In reading this quote we have to remember… the Raptors have never won more than 47 games in a season, and have only advanced past the first round of the playoffs once.  Perhaps (just perhaps) there is more to player analysis than focusing on the “intangibles.”

One last observation… once upon a time I noted that the owner of the Raptors tried to get his general manager in basketball to read The Wages of Wins.  Maybe he will have more luck assigning our second book (am I being too snarky?  Sorry about that).

- DJ

The WoW Journal Comments Policy

The Toronto Raptors appear to be on a mission.  Somehow, someway, this team needs to convince Chris Bosh to stay in Canada.  The issue for the Raptors isn’t money.  To convince Bosh to stay, Toronto must convince Bosh that he will someday play for a winner.  So far, though, the facts presented to Bosh haven’t been too convincing.

Bosh came to Toronto in 2003.   Since his arrival the Raptors have only posted a record beyond the 0.500 mark once.  In 2006-07, the Raptors won 47 games.  The next season the Raptors only won 41 games.  And then last year – as Table One indicates – Toronto won just 33 games.

Table One: Toronto Raptors in 2008-09

After such a poor season it’s not surprising the Raptors have decided to make changes. 

Rebuilding the Raptors?

Actually “some” is an understatement.  When we look at the players who were with the Raptors twelve months ago, only three – Bosh, Jose Calderon, and Andrea Bargnani – are still playing with Toronto.

When we look back at Table One we see nine players who produced in the negative range.  Most of these players are now gone (Bargnani and Patrick O’Bryant – a midseason addition last year – remain).  Three more below average players (Joey Graham, Pops Mensah-Bonsu, and Will Solomon) have also departed.  Unfortunately, the re-shaping of Toronto’s roster means that Anthony Parker [6.3 Wins Produced, 0.114 WP48], Jamario Moon [6.0 Wins Produced, 0.210 WP48], Shawn Marion [3.7 Wins Produced, 0.184 WP48], and Kris Humphries [0.7 Wins Produced, o.118 WP48] have also left town (Moon departed in the trade that brought Marion to Toronto). 

It’s interesting where these players will play in 2009-10.  Moon and Parker will be part of the rotation in Cleveland, a team that clearly has aspirations to win the NBA title in 2010.  Another team that hopes to go far in the playoffs is Dallas, and they added Marion and Humphries.  In sum, the four above average players that were allowed to depart Toronto went to teams that have been quite good in the past and plan to be quite good in the future. 

Of course, losing “good” players isn’t a problem if you add even better players.  To see if this happened, let’s consider Toronto’s potential depth chart (taken from ESPN.com and Yahoo.com):

Potential First String

PG: Jose Calderon [11.5 Wins Produced, 0.236 WP48]

SG: Marco Belinelli [0.1 Wins Produced, 0.007 WP48]

SF: Hedo Turkoglu [6.7 Wins Produced, 0.115 WP48]

PF: Andrea Bargnani [-0.3 Wins Produced, -0.006 WP48]

C: Chris Bosh [9.9 Wins Produced, 0.162 WP48]

Potential Second String

PG: Jarrett Jack [4.4 Wins Produced, 0.079 WP48]

SG: DeMar DeRozan [Rookie]

SF: Antoine Wright [-2.4 Wins Produced, -0.076 WP48]

PF: Reggie Evans [3.5 Wins Produced, 0.147 WP48]

C: Rasho Nesterovic [-1.3 Wins Produced, -0.051 WP48]

From this list it appears that Anthony Parker’s minute will be going to Marco Belinelli and/or DeMar DeRozan (and maybe Antoine Wright).  Parker was above average last year while Belinelli (in limited minutes) was well below average.  DeRozan is a rookie taken with the 8^th pick in the 2009 NBA draft.  His draft position suggests he could be above-average, although his college numbers tell a very different story.  Of the 47 players taken out of college last year, DeRozan ranked 39^th in PAWS40 [Position Adjusted Win Score per 40 minutes].  Such numbers suggest that the move from Parker to Belinelli-DeRozan is not a step forward for Toronto.

And then there is the Hedo Turkoglu acquisition. Here is what I said about this move last July:

In a move that dashed some hope in Portland, Toronto “stole” Turkoglu with a contract that will pay the former Magic player more than $10 million per season (and one wonders if the NBA Finals impacted that contract).  For his career, Turkoglu has produced 41.4 wins with a 0.105 WP48.  Average WP48 is 0.100.  So Turkoglu has been essentially average across his career.  And now he is 30 years of age (not a good age for an NBA player) and will turn 31 before the next season ends (an even worse age for an NBA player). 

Again, the numbers suggest Toronto has taken a step backwards.  Both Moon and Marion have been far above average in their respective careers.  Going from above average to average should not be considered progress.  Especially given the money paid to Turkoglu.

These were not the only moves the Raptors made.  At power forward Toronto has added Reggie Evans and Amir Johnson [2.7 Wins Produced, 0.147 WP48].  Both of these players are above average (although each play behind Bargnani).  And at point guard, Jarrett Jack was added to the roster.  Jack was close to average last year.  If Jack simply takes the minutes of Roko Ukic and Will Solomon, the Raptors are clearly better.  If Jack takes some minutes from Jose Calderon – the most productive player on the Raptors last year – then Toronto is worse off.

Toronto did not stop with the acquisitions of Evans, Johnson, and Jack.  As noted, in reshaping the roster a number of players who produced in the negative range last year were jettisoned.  Unfortunately it appears that Toronto can’t live without such players. Antoine Wright, Rasho Nesterovic, Marcus Banks [-0.075 WP48 for entire season], Quincy Douby [-0.015 WP48 for entire season], and Patrick O’Bryant [-0.056 WP48 for entire season] have all been added or retained from last year’s roster.  And all of these players produced in the negative range last year.

When we put the entire picture together, it doesn’t look too good for fans of Toronto.  The Raptors will still employ the services of Calderon and Bosh, and these players produced about 21 wins last year.  And it’s possible that Turkoglu, Jack, Evans, and Johnson can contribute another 15 wins.  After these players, though, who else is going to produce significant quantity of wins?  And if no one else produces much, how will Bosh believe that the Raptors are building a contender?

More on Bosh and Bargnani

One might ask if it’s necessary for the Raptors to appease Bosh. Across Bosh’s career he has produced 51.7 wins and posted a 0.152 WP48 [Wins Produced per 48 minutes].  If we exclude his first two seasons we see in the past four seasons he has posted a 0.179 WP48; with a career high of 0.210 in 2006-07.  In sum, Bosh is a good player.  But he is not LeBron James or Dwight Howard.   In other words, Toronto is going to need quite a bit more if it’s ever going to contend for a title.

Is that something more Bargnani?  Bargnani was the first player taken in the 2006 NBA draft.  Across his first three seasons he has produced -5.4 wins and posted a -0.043 WP48.  Obviously these are very poor numbers.  Fans of Bargnani, though, have argued that he played much better in the second half of the 2008-09 season.  If we compare Bargnani in the first 41 games of 2008-09 to the second 41 games (i.e. split the season exactly in half), then it doesn’t appear Bargnani got any better.  In the first half his Wins Score per 48 minutes [WS48] was 8.1.  In the second half he posted a 7.6 WS48.  To put that in perspective, an average power forward posts a 10.4 WS48; so Bargnani was very below average in both the first and second half of the 2008-09 season (for a power forward).

Basketball-Reference.com does offer month-by-month performance data, and this view might make Bargnani fans a bit happier.  Here is Bargnani’s WS48 by month last season:

• October: 14.4 WS48 in 50 minutes
• November: 9.1 WS48 in 401 minutes
• December: 2.0 WS48 in 392 minutes
• January: 9.4 WS48 in 569 minutes
• February: 7.2 WS48 in 441 minutes
• March: 11.5 WS48 in 399 minutes
• April: 5.0 WS48 in 201

If a person focuses only on March, it looks like Bargnani is an above average power forward.  To keep this belief, though, you have to ignore what happened in November, December, January, February, April, 2007-08, and 2006-07.  So you have to ignore quite a bit to keep believing in Bargnani. But if one tries really hard, a positive view can be kept.

Apparently, the Raptors were able to accomplish this feat.  In the off-season, Bargnani was signed to a new $50 million dollar contract.  If Bargnani can replicate what he did last March, he might live up to most of his new paycheck.  Unfortunately one suspects that November, December, January, February, April, 2007-08, and 2006-07 were not a fluke.  It seems likely that Bargnani is never going to live up to his new contract.

Given the money paid, though, Toronto appears to believe Bargnani can be part of a winning team.  And that will be true if Toronto acquires a few more players like Calderon and Bosh.  Until that happens, though, Toronto is not likely to contend for a title.  And keeping Bosh seems less and less likely.

- 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

Wins Produced vs. Win Score

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.

The Pareto Principle states that 80% of outcomes can be linked to 20% of people.  Although it’s not clear (at least to me) how much this simple rule – originally noted by the famed economist Vilfredo Pareto – applies in the general economy; it does appear that the Pareto Principle works in the NBA.  At least, it works in general.  About 80% of wins are produced by 20% of NBA players.

Reviewing 2008-09 in Minnesota

When we apply this rule to an individual team, though, Pareto can come up short.  For example, consider the Minnesota Timberwolves.   Last season the T-Wolves employed the following three players:

• Mike Miller (13.9 Wins Produced)
• Kevin Love (8.9 Wins Produced)
• Al Jefferson (6.1 Wins Produced)

In general, 20% of a team’s roster equals three players.  And these three players combined to produce 28.8 wins for Minnesota.  Had the Pareto Principle applied, the T-Wolves would have won about 36 games in 2008-09 (28.8 is about 80% of 36).  But as Table One reveals, the Wins Produced of this team was only 27.6 (and the team actually won just 24 games).  Yes, after Minnesota’s top three, the remainder of the roster combined to produce in the negative range.

Table One: The Minnesota Timberwolves in 2008-09

Certainly the injury to Al Jefferson hurt Minnesota.  But when a team employs 14 players who combine to produce -1.3 wins; that’s going to hurt also.

Looking Forward to 2009-10

This past summer the T-Wolves made changes.  Compared to the start of the 2008-09 season, Minnesota has made changes in the front office, head coach and coaching staff, and throughout the roster.  In fact, Minnesota only has five players – Jefferson, Love, Ryan Gomes, Corey Brewer, and Brian Cardinal – who were with the team last year.  So it’s safe to assume, the T-Wolves of 2009-10 will be very different from what was seen in 2008-09.  Then again, maybe not.

When we consider Wins Produced per 48 minutes [WP48] from the 2008-09 season, the top three players on the current roster are as follows:

• Kevin Love [0.208]
• Ramon Session [0.198]
• Al Jefferson [0.160]

Let’s imagine that this trio manages to play 36 minutes a night throughout the 82 game regular season.  And let’s imagine the productivity of these players doesn’t change.   Given our imagination, the T-Wolves would get 34.8 Wins Produced from these players.  Now if we also imagine that the Pareto Principle applies, then we can expect the T-Wolves to win about 44 games in 2009-10.  Perhaps this isn’t good enough to make the playoffs.  But it certainly is the best mark Minnesota has ever achieved without Kevin Garnett on the roster.

Unfortunately for fans of Minnesota, these “ifs” are not going to come true.  For starters, Kevin Love is hurt and he’s going to miss several weeks.  More importantly – despite all the changes to the roster – the problem that existed in Minnesota last year still remains.  Once you get past the top three players on the team, there are simply not many wins to be had.  To see this point, consider the team’s depth chart (taken from ESPN.com and Yahoo.com):

Potential First String

PG: Ramon Sessions [9.0 Wins Produced, 0.198 WP48]

SG: Corey Brewer [0.3 Wins Produced, 0.053 WP48]

SF: Sasha Pavlovic [0.7 Wins Produced, 0.034 WP48]

PF: Kevin Love [8.9 Wins Produced, 0.208 WP48]

C: Al Jefferson [6.1 Wins Produced, 0.160 WP48]

Potential Second String

PG: Jonny Flynn [Rookie]

SG: Wayne Ellington [Rookie]

SF: Ryan Gomes [0.6 Wins Produced, 0.012 WP48]

PF: Oleksiy Pecherov [-0.1 Wins Produced, -0.014 WP48]

C: Ryan Hollins [-0.5 Wins Produced, -0.051 WP48]

In addition to these ten players, the T-Wolves will also employ the following:

PG: Antonio Daniels [1.3 Wins Produced, 0.063 WP48]

SG: Damien Wilkins [-0.6 Wins Produced, -0.046 WP48]

PF: Brian Cardinal [-0.6 Wins Produced, -0.033 WP48]

C: Mark Blount [-0.8 Wins Produced, -0.176 WP48]

The career WP48 numbers of Corey Brewer [-0.004] and Sasha Pavlovic [-0.024] are both in the negative range (the same can be said of Pecherov, Hollins, and Blount).  Brewer is still young, so he might improve (and the same could be said for Pecherov and Hollins).   But the phrase is “might improve.”  At this point, the T-Wolves are employing three very good players – Sessions, Love, and Jefferson – and a large collection of veteran players who were below average in the past. 

Minnesota is adding Jonny Flynn, a lottery pick in 2009.  But his college numbers don’t suggest he is going to be a significant producer of wins in 2009-10.  As for Ellington, he was above average in college last season.  But one wonders how much the 28^th pick will actually play this next season.

The Same Story in Minnesota

Okay, let’s put the whole picture together.  Last year Minnesota had three above average performers and a collection of unproductive players.  After vast changes to the team’s roster, the T-Wolves still have three above average performers and a collection of unproductive players.  In sum, Minnesota – despite all the changes – looks like the same team we saw last year.  Except this time, Jefferson is healthy (well, maybe) and Love is going to miss significant time (last year it was the other way around).

All of this is bad news for the LA Clippers. According to DraftExpress, Minnesota’s first round pick in 2010 will go to the Clippers, but only if this pick falls out of the top 10. Given the T-Wolves current roster, though, this pick should remain in the top ten.  And if Minnesota fails to improve next summer, the T-Wolves pick shouldn’t move to LA in 2011 either.  In 2012, though, the pick is unprotected.  So by 2012 the level in happiness will have to increase in Minnesota, or a little bit of happiness will be moving to Clipperland.

- 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

Wins Produced vs. Win Score

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.