20 Second Timeout is the place to find the best analysis and commentary about the NBA.

Tuesday, December 08, 2009

Allen Iverson Ends Short-Lived Retirement

"I know I can play and I'm going to prove that."--Allen Iverson, 12/3/09 press conference.

Allen Iverson ended perhaps the shortest retirement in sports history by scoring 11 points, dishing off six assists and grabbing five rebounds in his second coming as a Philadelphia 76er but the 76ers lost to the Denver Nuggets 93-83. Chauncey Billups led the Nuggets with 31 points, eight rebounds and eight assists, helping to overcome a sluggish 14 point, 5-21 shooting performance from Carmelo Anthony, who failed to score 20 points for the first time this season. Andre Iguodala topped the Sixers with 31 points but that was not nearly enough to overcome Denver's ability to draw fouls and make free throws. Billups made all 11 of his free throws as the Nuggets shot 24-26 from the free throw line, compared to just 8-9 free throw shooting by the Sixers.

Iverson's relatively modest statistics were actually pretty good for someone who has not played in an NBA game in a month and it is worth noting that he had the second best plus/minus number for the 76ers (-1, trailing only Iguodala's +1); in other words, the Nuggets did most of their damage when Iverson was not on the court. The Sixers built a 10 point lead in the first half, were up 44-41 at halftime and were ahead 65-61 when Iverson went to the bench with 37.7 seconds remaining in the third quarter. When Iverson returned at the 9:08 mark in the fourth quarter the Nuggets led 72-65 and the Sixers were not able to gain any ground the rest of the way.

The game's outcome is hardly surprising; the Sixers were 5-15 prior to Iverson's arrival and had lost nine straight games. In fact, the Sixers have yet to win more than 41 games in a season since jettisoning Iverson early in the 2006-07 campaign; with Iverson onboard, they won at least 43 games in five of the seven full seasons prior to that and posted a .561 winning percentage (equivalent to 46 wins in an 82 game season) during the lockout shortened 1999 season, Iverson's third year in the league.

The real story is not the result of one game but the tremendous battering that Iverson's image has taken in the past year or so: in the 2007-08 season, then-Nugget Iverson ranked first in the NBA in minutes (41.8 mpg), third in the NBA in scoring (26.4 ppg), eighth in the NBA in steals (2.0 spg) and ninth in the NBA in assists (7.1 apg) for a team that won 50 games in the tough Western Conference--but since that time he has been traded by Denver and essentially banished by Detroit before fleeing Memphis early this season after playing just three games. In the blink of an eye--and without suffering a serious injury or any tangible loss of his physical skills--Iverson went from being one of the most prolific players in the league to persona non grata, someone who even the most woeful teams in the league expressed no public interest in signing. Are we really supposed to believe that Iverson could not help the Nets, Knicks or some of the league's other cellar dwellers? A lot of teams are putting all of their eggs in the proverbial LeBron James basket and it will be really interesting to see how their fans react when James does not sign with any of those teams and their fans realize that they have been sold a bill of goods.

Supposedly, Iverson's unwillingness to come off of the bench scared away most of his potential suitors but--as Iverson mentioned in his recent interview with John Thompson--prior to Iverson's arrival in Detroit there had never been a question about whether or not Iverson should be a starter. Then the Pistons inexplicably decided that Rodney Stuckey simply has to be a starter, so they sent Iverson to the bench (Stuckey has started all 20 games for the Pistons this year and near the season's quarter pole the Pistons are on pace to post their worst record since 2000-01). Iverson told Thompson that he would be willing to come off of the bench if someone beat him out for the starting spot by outplaying him and if "I know in my heart that that player is better than me. I wouldn't have (a) problem if (a coach told me) 'This guy right here outplayed you. He's a better player than you. He's the starter'...Why would I have a problem with that? Just like guys got to accept when I start over them--but they know when they look in the mirror and ask themselves 'Are you better than Allen Iverson? Should you be starting over Allen Iverson?' They know the answer."

Any intelligent, informed basketball observer--for instance, Jeff Van Gundy--understands that the Pistons bungled the Iverson situation; I would go further and say that the Pistons actually damaged Iverson's reputation around the league (not that this was their intention or goal per se). Uninformed basketball observers--such as "stat guru" Dave Berri*--leaped at the opportunity to pile on to Iverson; last season had barely gotten underway and the Pistons had not yet re-signed Antonio McDyess (who was traded to Denver in the Iverson deal but then let go by the Nuggets) but Berri immediately declared that the Pistons were heading into a downward spiral that was entirely Iverson's fault. The Pistons did in fact have a subpar season but there were many factors involved, including a coaching change that did not work, injuries to key players, McDyess' absence for a month--and the bizarre way that the Pistons seemed to do everything possible to minimize Iverson's effectiveness, as opposed to playing an uptempo, wide open style that would have taken advantage of Iverson's ability to create open shots for himself and others; after all, Pistons President Joe Dumars publicly stated that he traded Billups and McDyess for Iverson precisely because Iverson could add a dimension to Detroit's offense that Billups did not provide. Why acquire Iverson and make that statement if you have no intention of taking advantage of Iverson's skill set strengths? Was Dumars' grand plan really simply to acquire salary cap space to spend on Ben Gordon and Charlie Villanueva? Don't expect to see that Stuckey-Gordon-Villanueva "nucleus" in the NBA Finals any time soon.

Whether or not you like Iverson, he has earned respect because of his productivity and because of how hard he plays. The reason that Berri's articles about Iverson are so detestable and the reason that I respond so forcefully to them is that it seems like Berri--a nobody in his primary field, someone whose work assuredly will not stand the test of time--is trying to make a name for himself as an NBA "stat guru" at Iverson's expense, bamboozling the general public with mathematical gimmickry that the average person is not equipped to debunk; most people do not have much formal mathematical training, so they are reluctant to challenge an economist when he spouts off about what the "numbers" supposedly say, though intelligent people understand that Economics is Not a Science, Nor is Basketball Statistical Analysis and that, as I wrote earlier this year, "Basketball statistical analysts do not yet have all of the necessary data to completely 'model' the sport, nor do they fully understand how to use the data that they have."

Watch Iverson play for the rest of this season, look back at old footage of Iverson--like this nearly nine minute (!) compilation of dunks by the diminutive guard--consider Iverson's high rankings in pro basketball history in categories like scoring average, steals and minutes played and then decide for yourself whether my perspective or Berri's perspective best represents what you honestly think and feel about Iverson's place in basketball history: do you believe "no sensible person can deny that he is not only one of the greatest 'little men' (six feet and under) in pro basketball history but one of the greatest players of all-time, period" or do you believe "Iverson--across his entire career--has been slightly below average"?

*--One of the great mysteries of the universe (or least the basketball universe) is why ESPN's Henry Abbott has seemingly made it his personal mission to heavily promote Berri's half baked basketball ideology; even other "stat gurus" are extremely skeptical about Berri's methodologies, yet Abbott incessantly links to Berri as if Berri has discovered the Holy Grail of basketball analysis. ESPN employs its own "stat guru"--John Hollinger, who has sometimes crossed swords with Berri--and even though Abbott also links to Hollinger it seems like Abbott does so out of duty more than belief, whereas Abbott seems determined to lift Berri--a minor figure in his chosen field (Berri is an associate economics professor at that bastion of economic theory, Southern Utah University)--into prominence in the basketball world.

Labels: , , , , , , ,

posted by David Friedman @ 4:16 AM



At Tuesday, December 08, 2009 10:40:00 AM, Blogger Ziller said...

Your continued ridiculous angst toward those who use advanced metrics aside, I think you're off re: Henry's alleged favoritism of Berri.

Henry links to numerous critiques of Berri, and if anything more frequently discusses adjusted plus-minus (which Berri detests). I don't want to speak for Henry or fight his battle, but you're just not being accurate.

At Tuesday, December 08, 2009 11:59:00 AM, Anonymous J said...

I did not catch this game, and I have expressed my views about Iverson and particularly the Iverson-Billups/McDyess trade before, but in light of my comments in recent posts about Iverson's aversion to practice, which you challenged, I thought these two pieces were interesting:

"Allen Iverson wasn't required to attend every practice during his stint with Denver . . ."

I note the article does coddle AI a bit, claiming that "On the flip side, Karl had the right to make certain practices mandatory. If the Nuggets were in the midst of three or four days off and Karl wanted the full team together, Iverson had to attend. As long as Iverson didn't embarrass the organization with a non-excused no-show, the arrangement remained intact.

According to sources, Iverson never abused the pact and was a model citizen during his nearly two years in Denver. Having a veteran coach able to understand and massage egos, plus a front office that laid down the ground rules at the start, made the deal work."

I have my doubts, and I cannot imagine that allowing AI to miss practices at his whim was a positive influence on trouble-prone younger players like JR Smith and Carmelo Anthony.

Also, this column about last night's game claims that "Iverson[] steered his Rolls Royce into the player’s parking lot with little over an hour to spare until the opening tip."

You certainly have more knowledge about this than I do, but arriving to the arena slightly more than an hour before tipoff seems late to me. Is it not?

At Tuesday, December 08, 2009 2:07:00 PM, Blogger David Friedman said...


The truth is that Iverson is not the only player who does not fully participate in every practice; many players who log heavy minutes are accorded some type of relief from practice, particularly players who are past the age of 30. As I have mentioned before, Bill Russell sometimes drank tea on the sidelines and watched the Celtics practice; Red Auerbach understood the importance of not wearing out a star player who gave so much effort every single game. The important thing is for the star player, the coach and the team management to all be on the same page--and, according to the article you cited, that was the case with Iverson in Denver.

It is amazing how many people try to revise history and act like Iverson failed in Denver even though the Nuggets had their best regular season in two decades with Iverson putting up elite level scoring, assist and steal numbers while also posting the second best field goal percentage of his career.

You are correct that arriving an hour before tip off would be considered late. I don't know why Iverson was late but he probably will be fined for that (whether or not the team publicly says so).

Dennis Rodman was late for countless practices when he was with the Bulls, but Phil Jackson simply fined Rodman without making a public issue out of it and Rodman was a very productive member of three Chicago championship teams. Similarly, Iverson may be "high maintenance" in some ways but he undeniably plays hard and has a unique skill set.

At Tuesday, December 08, 2009 2:28:00 PM, Blogger David Friedman said...


"Angst" is not the correct way to describe my attitude regarding "advanced metrics." As I have repeatedly said, I am very much in favor of the idea of trying to accurately quantify what happens in a basketball game--I just disagree with those (like Berri) who believe that they have found the Holy Grail to do so. I also find it very odd that "stat gurus" like Berri appear to not be the slightest bit concerned about the subjective nature of many of the boxscore statistics that they plug into their formulas; I have done many posts about how assists are a subjective statistic and I am sure that similar problems exist in tabulating steals and blocked shots (though of course those numbers do not play as significant a factor in player ratings as assists do).

There are two problems with Berri:

1) His "system" is nonsense, as has been pointed out by Dan Rosenbaum and other figures in the Association for Professional Basketball Research (APBR) stat community.

2) Unlike Rosenbaum, Dean Oliver and other people who are doing meaningful work with basketball statistics, Berri appears to be incapable of admitting that he or his system can ever be wrong about anything.

I cannot claim to have read every single True Hoop post--and, frankly, I rarely visit the site at all anymore--but considering points 1 and 2 mentioned above I feel safe in saying that Abbott has given Berri way more positive publicity than Berri has earned and this is bad for two reasons: not only does it advance Berri's work despite its lack of merit but it also holds back the work of others, because instead of promoting Berri's fictions Abbott could have used that bandwidth more productively--or, to use an economics term that Berri surely knows, there is an opportunity cost to Abbott's readers every time Abbott links to nonsense.

At Wednesday, December 09, 2009 12:04:00 AM, Blogger Unknown said...

hi long time reader first timer commenter. first off i really enjoy reading this blog, and david, although i do not agree with everything you say i find your nuanced reasoning refreshing and easy to take.

i completely agree that iverson is being unfairly attacked here, this man is a athletic marvel. i still remember that game 1 of the finals series against the lakers...absolutely amazing someone of his size having such an enormous impact on any game. to use a soccer analogy, hes like a maradona figure whos skill is so great that it completely overshadows any percieved physical disadvantages.

anyways, my main gripe to really comment are those stat guys. as an economics major, i completely understand that they would try to quantify everything. BUT those stat guys should know better than anyone else that stats CAN lie and it is pretty impossible to model something like basketball in a perfect manner. economics is not a science..it should be used as a tool to reinforce what we are seeing, not to be used a magic eight ball. it is flat out wrong to say otherwise. it is so wrong that these stat guys can't understand that just iverson was on the roster during detriot's spiral that will be due to other factors. thank you for a well written blog and hope to see more in articles very soon.

At Thursday, December 10, 2009 8:28:00 PM, Blogger Bhel Atlantic said...


I certainly take your point about how basketball statistics can be subjectively determined (what is an assist? who got the rebound? was that a steal or a fumble? etc.) However, we could posit that (just to make up some numbers) the official scorer fails to catch 10% of all "true" assists, but also adds some number of "fake" assists, say 20% of true assists, to the recorded assist total. If this were true, wouldn't ALL players be subject to the same data problems? Or, do you think the official scorers are systematically biased to inflate assists MORE for certain players / types of players?

At Friday, December 11, 2009 4:55:00 AM, Blogger David Friedman said...

Bhel Atlantic:

That is an intelligent question. Regarding assists, my answer is that when I tracked assists for Chris Paul and other point guards over the course of several games I charted every legit assist that I saw first and then went back to see what the official scorer did. I never found one example of a "missed" assist but I found numerous examples of bogus assists. I seriously doubt that there is a problem of assists being "missed." Regarding other statistics such as steals and blocked shots there certainly could be situations in which the wrong player is credited, particularly in a case when one player tips a ball and another player recovers it. Those mistakes/discrepancies are important in one sense but in terms of player ratings I don't think that they are statistically significant, though they could of course affect defensive ratings (as opposed to overall ratings in which defensive statistics would carry less weight).

As for all players being affected equally by scoring errors I emphatically do not believe that to be the case. Players like Steve Nash, Chris Paul and LeBron James tend to have the ball in their hands from the start of a possession until their team scores--they either take the shot themselves or make the final pass before a shot is attempted, whether or not that pass fits the definition of an assist. The overcrediting of assists favors such players at the expense of players who do not handle the ball so frequently; as I have pointed out many times, Kobe Bryant creates many shot opportunities by drawing double teams but because of the spacing in the Triangle he often makes the pass that leads to the pass that is credited with the assist. Bryant still generally leads the Lakers in assists but he does not put up the gaudy assist totals that Paul, James and Nash do even though he does just as much to create scoring opportunities for his teammates.

Just to be perfectly clear, I am not saying that there is some kind of intentional conspiracy to increase the assist totals of certain players. I think that the scorekeepers simply are applying a looser definition for assists than the rulebook provides and that this looser definition tends to benefit a handful of players who monopolize the ball during their teams' offensive possessions. My problem with this is twofold: (1) it renders assist records meaningless because players from previous eras compiled their totals by a stricter standard; (2) it has an impact on these ubiquitous "advanced" player ratings. Look at how highly James, Paul and Nash generally rank in these systems even though the consensus opinion among knowledgeable NBA observers (as confirmed in the recent Sporting News poll of NBA executives and distinguished players) remains that Bryant is the league's best player. James is exceptional and I believe that he was the most productive player in the NBA last year (and thus narrowly more deserving of the MVP than Bryant, though James won the media-driven MVP voting in a landslide) but Bryant is still the most complete all-around player in the game. Paul and Nash are great point guards but their inflated assist totals give an imprecise representation of their true impact on games.

At Friday, December 11, 2009 11:36:00 AM, Blogger Bhel Atlantic said...


Thank you for your detailed response.

You say that "The overcrediting of assists favors such players [Nash/Paul] at the expense of players who do not handle the ball so frequently ... Kobe Bryant ... makes the pass that leads to the pass that is credited with the assist. "

Well, if that's the case, then Bryant does *not* deserve any assists, under the textbook definition of "assist". Perhaps a new statistic called "hockey assist" should be created to accommodate the situation you describe. (In fact, perhaps teams are already tracking such a statistic internally.)

In your view, if Nash or Paul makes a pass that leads immediately to a basket (possibly with some extent of juking or dribbling by the scorer), is he MORE LIKELY to be credited with an assist, compared to a scenario where Bryant or Shane Battier makes the *same* pass? If so, then inter-player comparisons are biased. But if not (i.e. if all players have the same likelihood of being credited with an assist, and even if that likelihood is somewhat inflated by over-generous official scorers) then I don't see the problem.

At Friday, December 11, 2009 4:29:00 PM, Blogger David Friedman said...


I am not suggesting that Bryant "deserves" more assists or should be given assists for passes that do not fit the strict definition of assists; I am saying that scorekeepers should stick to the rulebook requirements and "stat gurus" should acknowledge that the boxscore data that they use contains errors that influence what their formulas say.

I have not tracked assists for enough players to answer the second part of your question with any degree of certainty. Have you read my various posts about charting assists? Do you really think that it is not a problem if Chris Paul passes to David West, drifts to the corner, watches West put on the Kevin McHale low post clinic and then Paul gets credit for an assist after West scores? I have charted many such plays involving West and I have good reason to think that a substantial portion of his assists are bogus. Keep in mind that Paul has supposedly set all kinds of records regarding his assist totals in regular season and postseason play. How should Oscar Robertson, Isiah Thomas and others whose records Paul has broken feel about this? How should basketball fans feel about this?

At Friday, December 11, 2009 5:52:00 PM, Blogger Bhel Atlantic said...

David: Thank you for your reply. I agree that "assist inflation", where CP3 is credited with an assist after he passes to West and West does the McHale clinic, is problematic. I've read your prior posts on the topic. However, do you believe that "assist inflation" has worsened over time, so that Isiah, Oscar, etc. did not receive such a benefit from the official scorers? (e.g. Isiah passes to Mark Aguirre, who jukes and jives then scores, and Isiah gets credited with an assist) If assist inflation has always been a problem to the same extent, then we can make ordinal (though perhaps not cardinal**) comparisons of CP3 against historical peers.

**Here's what I mean about ordinal vs. cardinal. Let's say Paul's "true" assists per season is 1000 and Stockton's "true" assists per season was 900. If we posit that scorekeepers always inflate the true number by 20%, then Paul's recorded assists per season is 1200 and Stockton's recorded assists per season is 1080. Over five seasons, Paul's true assists is 5000 and his recorded assists is 6000; Stockton's true assists is 4500 and his recorded assists is 5400. The relative ranking between Paul and Stockton is the same whether we look at "true" numbers or recorded numbers; the absolute gap in true numbers is 500 and the absolute gap in recorded numbers is 600. I think we care more about relative rankings, don't we? Paul wins either way. So in a sense I don't care much that the data are inflated. However, if we posit that the assist inflation by official scorers has worsened (increased) over time, then that's a problem, as it would mess up the ordinal inter-player rankings.

Similarly, do you believe that Nash/Paul benefit from assist inflation more than other 2009 PGs do (normalizing for a given amount of passes thrown)?

Thank you.

At Saturday, December 12, 2009 4:55:00 AM, Blogger David Friedman said...


I definitely believe that assist inflation has worsened over time. I say that based not only on my own observations but also based on what I learned from interviews with retired legends like Rick Barry and Oscar Robertson. You can also verify this for yourself by looking at the ratio of assists to field goals made leaguewide over the years; the NBA game is based more on isolation than it was in years past and yet the ratio of assists to field goals made has increased.

I think that assist inflation began some time in the 1980s but has gotten worse in recent years; that is my subjective impression and is not based on a scientifically rigorous study, because the only formal look that I have made into this subject consists of the various games in which I tracked assists for Chris Paul and other point guards.

I don't have the time or resources to do a comprehensive examination of this issue but what I found in the games that I charted certainly should have sparked some interest from the league and "stat gurus" but apparently the rest of the world is quite content to sweep the whole matter under the rug.

As I explained in my previous comments, I believe that players who monopolize the ball--guys like Paul and Nash--do benefit disproportionately from the way that assists are currently tabulated. Can I prove that? No, but the games that I charted regarding Paul certainly suggest that my subjective impression is correct or, at the very least, should be investigated more deeply.

By the way, the NBA has taken away two triple doubles from LeBron James--one for a bogus rebound and one for a bogus assist. The league claims that it reviews every game for scorekeeping errors and makes corrections when necessary but to the best of my knowledge none of the bogus assists for Paul that I uncovered have been corrected. My work on this subject was linked to by Henry Abbott (before True Hoop went completely downhill), the Atlantic's Matthew Yglesias and others and I have spoken about this subject to a reporter from the Wall Street Journal, so the information is certainly out there for the NBA and the "stat gurus" to take action.


Post a Comment

<< Home