Aging Billups and Telling Stories

Posted on March 22, 2009 by


On Sunday, while most basketball fans were tuned to CBS to watch the NCAA Tournament, ESPN telecast a game between the Miami Heat and Detroit Pistons.  As a fan of Detroit, I decided to look away from the tournament and look in on the Pistons basketball.  The picture was almost good.  Specifically the Pistons — without the services of Allen Iverson, Rasheed Wallace, and Richard Hamilton – managed to lead most of the way before faltering at the end. 

Although the game was entertaining, the comments of Mark Jackson and Jeff Van Gundy were more interesting.  At different points in the broadcast the conversation turned to why the Pistons have faltered.  And once again the viewers were able to see “expert” broadcasters try and reconcile these three “facts.”

  • Allen Iverson is one of the greatest players to ever play the game.
  • Chauncey Billups is a very good player, but not one of the greatest players to ever play the game.
  • The Pistons with Allen Iverson are clearly worse than the Pistons with Billups.

The simplest way to reconcile these three statements is to acknowledge that the first is not true.  At least, if we focus on productivity, Iverson is not one of the most productive NBA players in history. 

But if you can’t make that leap, then you have to resort to words like “team chemistry”.  Unfortunately, since “team chemistry” can’t be quantified (unlike player productivity) we can never properly evaluate the merits of the “chemistry” argument.

As I have noted throughout the season – most recently last month – one does not need “chemistry” to explain the Pistons demise.  No, much of the decline (although not all) is tied to the fact that Iverson is not as productive as Billups.

Another Reason Why Trading Billups was a Good Idea

Again, I have said this before.  What I wished to add to the subject is something I observed in looking at the numbers Billups is posting in Denver this season. 

Table One: The Denver Nuggets after 70 games in 2008-09

As Table One notes, Billups has produced more wins than any other player on Denver’s roster this season.  But if we compare his production to what he did last year in Detroit, he clearly has declined.  Last year he posted a 0.304 WP48. This year in Denver his mark is only 0.188.  Yes, he is still above average (average is 0.100). But he clearly is offering less.

When we look at the individual numbers – posted in Table Two – we can see where he has declined.  This season Billups has seen declines with respect to both his shooting efficiency and assists. 

Table Two: Evaluating Chauncey Billups

The next question we should ask is why Billups is doing less.  And one issue I would emphasize is his age.  Billups is 32 and will turn 33 in September.   This means he is rapidly approaching the age where playing basketball in the NBA is not possible.  To illustrate, across the past 30 years, 95% of player seasons were played by players who were younger than 35 years of age.   In sum, the clock is approaching midnight for Mr. Big Shot and when it hits 12, he won’t be of much use to an NBA team.

All of this suggests that the Iverson-Billups trade was a good move by Joe Dumars (Detroit’s GM).  Yes, fans of Detroit are suffering this year.  But as noted previously, Iverson’s contract expires and this gives Detroit hope for next year.  Plus, Dumars got rid of a player that will be approaching 35 years of age in 2010-11 while collecting $13 million. 

In sum, it looks like Dumars has done the same thing to Billups he did to Ben Wallace.  He let a player depart whose production was destined to slip.  And this is something every prudent general manager should be doing.  As the late Cotton Fitzsimmons once said (and I can’t find the quote but I think he said something like this): “Please don’t let my great players retire on me.” 

Forecasting Wins

The predictability of age brings me to the issue of prediction.  If we look back at Table One we see two projections of the Nuggets. The first assumes that what the players did in the past (typically last year) is what they will do in the future.  The second looks at how many wins a team should get given what the players are doing this season.  When we look at these two projections we see that despite the decline we see in the production of Billups, Denver’s Wins Produced this season is consistent with what we would have expected if we believed that Denver’s players would keep doing exactly what they did in the most recent past.

The approach presented in Table One is the standard approach I take in evaluating a team.  But I think it has led to some confusion. 

To see this, consider what J.A. Adande said at this past week. In a wonderful article on Malcolm Gladwell’s latest book – Outliers (an excellent book as I have noted in the past) – Adande made the following observation.

The Wages of Wins data suggested the individual components of the 2007-08 Boston Celtics were good for 52 victories based on their production the previous season. The Celtics wound up winning 66 games and the NBA championship. Clearly, something was up that couldn’t be explained by the numbers.

The table Adande linked to was part of a column I wrote on the Boston Celtics last summer.  In this column I noted the following:

As Table Two indicates, given what these players did in 2006-07, the Celtics should have expected about 52 wins last year.  In other words, the team should have expected about a 28 game improvement in the standings. 

One should note, though, that 2006-07 was a down year for Paul Pierce, Kendrick Perkins, KG, and R. Allen.  If we look at the 2005-06 numbers for these four players – numbers that appear more consistent with each player’s career marks -Boston would have expected about 65 wins last year.  When the season was over, the Celtics – led by KG and Pierce (and let’s not forget Rajon Rondo) – won 66 games.  The 42 additional wins was the largest regular season turnaround in NBA history.  And when the post-season ended, the Celtics had won the franchise’s first title since the days of Larry Bird.

As one can see, if we consider more than just what Boston’s players did in 2006-07, then the 2007-08 season is easy to explain. 

Obviously Adande ignored the words and focused strictly on the numbers I presented.  And this is understandable.  It certainly looks like I am arguing that what Boston would do in 2007-08 is strictly a function of their players did in 2006-07. 

Certainly it’s the case the most important factor determining player performance in this season is what the player did in the past.  And that is a story I think needs to be told.  NBA players – relative to what we see in other sports – are very consistent. 

That being said, past performance does not explain all of the future.  A proper forecast would note the issue of age, the productivity of teammates, injury, and in some instances, coaching.  None of these issues are raised in my simple tables.

They are, though, often raised in my posts.  What I typically do, as I did with my discussion of the Billups and the Nuggets, is first present the table.  Then I look for changes in performance and make that the focus of the discussion.  For most players, there is little change. But for a few, we do see differences.  And it’s these differences that make for an interesting story.

In sum, the tables are not designed to be a formal forecast.  They are designed to start a conversation about a team.  Part of that discussion emphasizes player consistency.  But part of that discussion emphasizes other causal factors like age, injury, etc…. (but probably not “chemistry”).

– DJ

The WoW Journal Comments Policy

Our research on the NBA was summarized HERE.

The Technical Notes at 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.