Better Experts Needed: More on an article in the Philadelphia City Paper

Posted on April 19, 2008 by


A few days ago the following article by E. James Beale appeared in The Philadelphia City Paper (and as you will see, I really like this article):

Better Experts Needed: How the Sixers turned it around … sort of

The article begins with the following observation:

Back in October nearly every major media outlet believed the 76ers had a chance to be historically bad. The team was coming off of another losing season and had traded away the face of its franchise, Allen Iverson, for what appeared to be a hodgepodge of throwaway assets – an aging point-guard named Andre Miller and two late first-round draft picks. ESPN asked 10 of its NBA experts to predict how teams would fare; nine picked the Sixers to finish last in their division, and seven said they would be the worst in the entire Eastern Conference.

If the predictions were harsh, the explanations were even worse: Jon Barry called the team “just brutal.” Chris Broussard said “the Sixers will be bad. Real bad.” Chad Ford couldn’t see Philly “making a serious run at 30 wins, let alone the playoffs,” and John Hollinger, ESPN’s stat-geek – the one supposedly immune to trendy picks – pegged the Sixers for only 21 wins, good for last place in the conference.

It wasn’t just ESPN, either. Sports Illustrated wrote: “it will take a superlative effort from [all the players] for Philly to avoid spending the season at the bottom of the division.” Its annual preview placed the Sixers 15th out of 15 teams in the Eastern Conference. So did Yahoo’s sports section. And CBS Sportsline. And Slam magazine. Even NBA TV, which has a vested interest in promoting its product, was not kind to our Sixers. “Can they be a winning team? Not yet,” said Frank Isola. Peter Vecsey was more blunt: “This team is going nowhere.”

Obviously all these forecasts were incorrect.  The Sixers finished the regular season with a positive efficiency differential (offensive efficiency minus defensive efficiency) and a record of 40-42.  With an opening round match-up with the Detroit Pistons, Philadelphia’s season isn’t likely to last much longer.  But given the view of this team before the season started, for Philadelphia to make the playoffs is a major achievement.

Or is it?

Here is more from the Beale article:

David Berri is a professor of economics at California State University-Bakersfield, and one of the authors of The Wages of Wins, a book with an accompanying blog. He posits that the success of basketball teams can be accurately measured in the stats of individual players. He claims to have come up with an algorithm that gauges the value of NBA players based on “Win Scores,” the predicted number of wins a certain player is responsible for over the course of an NBA season. Berri says the Sixers’ record is no shock, and that the so-called experts should have seen this coming. After all, he did.

In a recent e-mail, Berri traced the Sixers’ resurgence to something that at first glance seems counterintuitive: last year’s trade of The Answer. “The key issue when it comes to the 76ers is the impact of Allen Iverson. People believe that Iverson is one of the greatest players to play the game. And when such a player departs a team, you should be much worse. When we measure productivity, though, we find that Andre Miller … is actually more productive.” The trade – widely panned as a talent-dump – may have actually made the Sixers better.

How much better? Well, according to Berri, pretty much exactly as “better” as the Sixers have been this season. In a blog post following last winter’s AI trade, Berri wrote that “the 76ers can expect to win about 30 more games [in 2006-07]. This gives the team a final record of 35-47.”

The Sixers’ final record last year: 35-47.

“When we look at all the numbers generated last year by the current [76ers] players … and we value those statistics in terms of wins – we see that fans of the 76ers should have expected this team to win about 41 games,” Berri says. In other words, this playoff-bound Sixer team is not a mirage or a fluke, but rather almost exactly as good as it was after they traded AI last year. They’ve played better since their slow start not because of philosophy or chemistry, but because of the law of averages.

If Berri is right – and on the subject of the Sixers, he has been now, twice – it may mean that traditional basketball experts are missing something when it comes to analysis. After all, they’re getting beaten by an econ professor with a blog.

Of course, statistics don’t capture the entire picture. They couldn’t predict that Thad Young would dedicate himself to defense. (By his own admission, he “played no D in college.”) Nor could they have guessed that the Sixers players who have produced the most wins – Andre Miller, Andre Iguodala, and Samuel Dalembert – would miss a combined zero games this year.

But while effort and luck – and maturity and chemistry – may matter, ultimately, they can’t compare to player quality, and that’s best measured by on-court statistics. When it comes to wins and losses, it seems, everyone should leave the predicting to the professor.

Obviously I like this article.  But I think should clarify the argument being made.

NBA observers – as is often argued in this forum – tend to over-emphasize scoring.  Consequently, when a major scorer like Allen Iverson departs a team, it’s expected by many NBA observers that the team should get worse.  And this expectation is hard to let go.  The Sixers – without Iverson – finished the 2006-07 with a 30-28 mark.  This record over 58 games should have convinced the “experts” that the Sixers were capable of being an average team – which is all they are – without The Answer.  Despite this evidence, though, the experts “knew” that losing Iverson must make the Sixers much worse.  And that “knowledge” drove the forecasts.

When you look at the pas productivity of the players the Sixers employed this year, you did not expect this team to be quite as bad as people in the media suggested.  All that being said, I don’t want to give the impression that we can simply look at past performance and predict the future record of each team.  In other words, as I have said before, numbers are not a crystal ball. 

The lesson we learn from the Philadelphia story is that the explanation for a team’s record should begin with the past performance of the players.  And for the Sixers, the story pretty much ends at that point.

But for other teams – Chicago and Dallas immediately come to mind– past performance doesn’t match perfectly with the future.  At least, it doesn’t for every player on the team. When this happens, we have to look at the individual players to see specifically which player’s production has changed.  And then we have to look at the individual numbers to see where player performance is different.

The numbers can tell us why a team is successful (or not).  Often those numbers tell us that a team is successful (or not) because it has good (or bad) players.  When the past performance of the players isn’t the explanation, those very same numbers are going to tell us which players are doing something different (and what specifically are the differences).  In sum, The Answer is – much more often than not – in the numbers.  And you can often find that answer, if you look beyond the numbers in the scoring column.

– DJ

For more on the Sixers, see the following posts:

The Data Ain’t Lying

The Benefit of Low Expectations in Philadelphia

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.