Bill Simmons is Wrong! (But also…) On Russell and Chamberlain’s Supporting Casts
I just can’t let this go. My distaste for Bill Simmons’ smug pseudo-argumentation has led me on a four-day journey down a rabbit hole of advanced statistics and I feel compelled to share my report of the trip.
At issue? The debate over whether Bill Russell or Wilt Chamberlain had the better supporting casts. For those who don’t follow basketball (those who do can skip this paragraph), Russell and Chamberlain were unarguably two of the best centers ever to play the game. During the ten years they were both in the NBA, their teams played against each other 147 times, including eight playoff series (two NBA finals series). Pretty much since their first face-off, on November 7, 1959, fans and writers have debated which of the two is the greater player. Among those who partake, the consensus seems to come down in favor of Russell, primarily on the grounds that his teams won more of their head to head games and playoff series (his Celtics teams beat Chamberlain’s teams won 87 of their games and 7 of their 8 playoff matchups). Though Chamberlain amassed superior individual statistics in nearly every category, the argument goes, basketball is a team game in which success is ultimately measured by team success. Over the years, the debate has grown beyond a question of basketball skills to involve playing styles and the moral virtues (and deficiencies) these styles supposedly embody. In this way, Russell appears as intelligent, principled, unselfish, dedicated and hardworking and Chamberlain as short sighted, opportunistic, selfish, lazy, and naturally talented.
Bill Simmons is only the most recent—though also one of the most popular and influential—writers to take this position. In Chapter Two of his 2010 New York Times bestselling The Book of Basketball, Simmons claims to lay to rest what he calls the “insane” position that Wilt was superior by debunking what he sees as the six myths upon which this position is founded. Here, I’m only interested in the first of the so-called “myths” that Simmons takes on: that Russell had a better supporting cast. Simmons purports to debunk this one by examining their respective supporting casts, year by year. He assesses the strengths of their teammates primarily through (selective) references to how many of their teammates appeared in All-Star games or on the annual All-NBA teams, or would eventually be inducted in the Hall of Fame status. His conclusion, which he frames as generous to partisans of Chamberlain, is that Russell had the superior supporting casts 5 times (1960-1964), Wilt had the upper hand 4 times (1966-1969), and they were even in 1965.
Setting aside the selective way that he uses these indicators, I believe that the indicators themselves are probably not the best for assessing supporting cast strength. The annual All-NBA teams are voted at the end of each NBA season by a panel of US and Canadian sportswriters, starting lineups for the NBA All-Star game are determined by fan balloting and reserves by a poll of the coaches in each conference. Both of these then involve subjective judgments without any clearly or explicitly defined criteria and, moreover, don’t give us a clear sense of whether the player is being honored for achievement in this season or whether past accomplishments are being considered as well (whether or not they should be). As for the NBA Hall of Fame, the selection process for its inductees takes place only well-after the conclusion of a career (and so gives no evidence of a player’s contributions in a given season) and involves an even smaller judging panel and criteria that are even more shrouded in mystery and, for some (including Simmons, who devotes an entire chapter to reinventing what he sees as a flawed Hall of Fame), controversy. Finally and strikingly given the anti-individual slant of Russell partisans like Simmons—all there of these measures tend to magnify the impact of exceptional individuals. What would be better would be a measure built upon something less subjective, with clearer criteria, and that offered a picture of the contribution of the whole supporting cast—not just the All-Stars and Hall of Famers—to the team’s victories. After all, it’s a team game, right?
Since the debate itself largely hinges on winning, it would be better to measure the contributions of the supporting cast members to the wins accumulated by Russell’s and Chamberlain’s teams. Fortunately, there is a metric that approximates exactly this. Win Shares per 48 Minutes (WS/48) is designed to measure an individual player’s contribution to team wins per every 48 minutes played. Though I’m far from adept at statistical thinking, I got some advice from Neil Paine of Basketball-Reference.com, who has thought a bit about this himself before proceeding (though I should say that any errors in approach or execution here are my own). Here’s what I wound up doing:
- I included only those teammates who played enough minutes in each season to qualify for the statistic according to Basketball-Reference.com.
- I tabulated (1) their age; (2) Career Regular Season WS/48; (3) Career Playoff WS/48; (4) WS/48 for each regular season and (5) each playoff run that they played with Russell or Chamberlain.
- I also came up with a rough measure of how well they performed in (6) a given regular season or (7) given playoff run by dividing their regular season and playoff numbers for that season by their overall career regular season and playoff average and multiplying by 100 (yielding what I’ll call, for lack of a better phrase, an “actual performance index”).
- Lastly, I noted whether or not each player’s (8) regular season or (9) playoff WS/48 number ranked in the All-Time Top 250 for the category (and where), just for the fun of seeing who put up monster performances.
- I then totaled and averaged these figures for each team in each season.
- Separately, I gathered the same statistics and performed the same calculations for Russell and for Chamberlain in each season.
Let me remind you of Simmons’ conclusions, Russell had the superior supporting casts from 1960-1964, they were even in 1965, and Chamberlain had the superior supporting cast from 1966-1969. Since Russell’s teams won championships in all of those years except one (1967), Simmons concludes that Russell, on this score anyway, was the superior player.
I can’t say who was the superior player, but based on what I found, Simmons seems drastically to have missed the mark in his assessment of supporting cast strength.
That’s right. According to the figures below, and contrary to Simmons’ conclusion, Russell had the better supporting cast for the regular season and the playoffs in each of the ten years both players were in the league.
Let me share some of the key data (if you’d like to see all of it, just let me know). Green shading indicates that Russell’s supporting cast outperformed Chamberlain’s in the category. Red (or in the case of 1969 purple) shading indicates that Chamberlain’s supporting cast got the upper hand. The number indicates the actual difference in the relevant total WS/48 value of the two supporting casts. For the final two categories, I’ve simply listed the ranking on the Regular Season or Playoff All-Time Top 250 WS/48 for that season or playoff run (no team under consideration ever had more than one individual supporting cast member place in the top 250). In other words, the more green you see in a given year, the better Russell’s supporting cast, and the numbers in the right two columns tell you when one of their teammates made a historically strong contribution to regular season or playoff wins, respectively.
So, yeah, it’s lookin’ pretty green…Celtic green. What’s obviously striking here is that in every single season—including in 1967, when Chamberlain’s team won the championship and including 1969, when (playing alongside legends Elgin Baylor and Jerry West) Chamberlain’s Lakers were favored to win it—Russell’s cast of characters brought the higher Career Regular Season and Career Playoff WS/48 and contributed more to the regular season and playoff wins of their team than their counterparts on Chamberlain’s team every single year. I think that Boston’s domination of the NBA from 1956 through 1969 (11 of 13 NBA titles, including eight straight from 1958-1966) is very impressive. But, once I saw these numbers, I started to wonder not how they managed that feat, but how it was that Chamberlain’s teams contended in so many of these seasons, often taking Boston to seven games before succumbing; let alone how they won a championship in 1967.
To suggest a response to that, here’s the same table, only now I’ve plugged in Russell and Chamberlain’s individual numbers: again shading green for Russell and red (or Purple) for Chamberlain (I’ve left their career numbers out because they obviously would remain fixed throughout the ten year span—for the record, though both are among the league’s All-Time leaders, Chamberlain’s Career Regular Season and Playoff WS/48 were somewhat higher than Russell’s).
Lots of red (and a little purple) now mix into the sea of green. In nine out of ten regular seasons, Chamberlain surpassed Russell’s WS/48 mark. And, in seven out of the nine seasons in which both players were in the playoffs Chamberlain did the same. Chamberlain also offered five playoff performances worthy of the Top 250 all time, compared to two by Russell, and nine regular season Top 250 years, compared to three by Russell.
Don’t get me wrong, the numbers don’t show—and I’m not saying—that Russell was carried to victory like a sick dog by his superior teams despite heroic efforts by Chamberlain to overcome his shitty teammates. They just suggest that far more often than not Russell got greater contributions from his supporting cast than Chamberlain, and that Chamberlain, as most already concede, outperformed Russell individually, though clearly not by enough (except in 1967) to make to make up for the difference in their supporting casts. You can pinpoint this a bit just by looking at the substantial difference between Russell’s supporting cast’s advantage over their counterparts and Chamberlain’s advantage over Russell.
What we don’t know from these numbers alone is the cause of the difference in support in cast contributions? Did Russell’s “unselfishness” make his teammates better than they otherwise were (as conventional wisdom frequently holds)? Did Chamberlain’s “selfishness” make his worse? The fact that Russell’s supporting cast had substantially higher Career WS/48 numbers might suggest that they were simply better players, regardless of who they played with. But we actually don’t have enough information to make that claim.
What we are really asking with these questions is the following: setting everything else aside, what effect did one player (Russell or Chamberlain) have on another (any one of their teammates)? This turns out to be an enormously complicated question, requiring rather minute tracking of game play. Economics professor Dan Rosenblum tried to figure this out for just two NBA seasons and James Piette, Lisa Pham, and Sathyanarayan Anand sought, if I understand their work, to root out the limitations in Rosenblum’s model in this paper. When you look at the complexity of the operation, you can understand why, if for no other reason than the creakiness of the stats gears in my mind, I hesitate to take it on (though I’d be delighted if someone else did!). In any event, my numbers by themselves don’t tell us. In fact, I’m no stat guru, but I think that we lack the raw data we’d need to figure this out where Chamberlain and Russell were concerned.
Hopefully, what the numbers do show nonetheless usefully complicates the pseudo-argumentation and so-called conventional wisdom purveyed by Bill Simmons and other adherents to the “Myth of the Rivalry.” It can be fun to engage in such debates, I know, and to bring passion and partisanship to them. If I didn’t think so I wouldn’t have spent the last four days poring over websites and trying to remember how to use Excel. What’s more, though it’s never inclined me to say he was better than Russell, I’ll acknowledge that Wilt holds a powerful grip on my heart, one that runs through my recently deceased father and back into my early childhood and that it’s important to express these things. But if one is going to play the sophist and craft arguments to support the views of one’s heart, let these arguments hold up to minimal rational scrutiny, or, if not that, then at least be beautiful. Otherwise, it’s just sports guys shouting, trying to make up in arrogance and bluster what they lack in evidence and reasoning ability.
I must say, however, that I would prefer—even more than sprinkling a bit more reason or elegance into the fray—to understand what the hell we are doing as a culture—the culture of basketball, I mean—by aggressively arguing over an unresolvable question for more than half a century. Or, to put it another way, to take seriously the rhetorical question that Chris Flink posed here:
@YagoColas “You see a man with 31k points. Another has 11 rings. You think: I have to choose. What the hell is wrong with you?”
— Chris Collision (@cfCollision) November 16, 2013
Really: what is wrong with you? With us? To begin to understand that, we’ve got to step out of the debate entirely. Then, we’ve got to carefully examine how the whole debate is structured in thought and language (I made a first, rough and hasty stab at this a few years ago), learn the facts of the history of basketball and society, and reflect on how the topic of debate and the way that people approach it fits into the issues of its time and place, thereby shedding some light on what sort of social and cultural work its doing.
Having done some of this work, I believe (as I will argue in third chapter of my forthcoming book Ball Don’t Lie!) the debate expresses powerful anxieties and desires stirred within what I call the White Basketball Unconscious in America by the the rapid racial integration and sharp rise to preeminence of African-American men at basketball’s highest and most visible level, beginning in the decade ushered in by these two players. In the face of the integration of basketball and the undeniable fact that the league’s best players were disproportionately African-American, the debate appeared to accept these realities (by confining the debate over who was best to two African-American players) while simultaneously denying them (by reducing Russell and Chamberlain to embodiments of athletic and moral attributes racialized as “white” and “black,” respectively). In so doing, and by declaring Russell the superior player, it served to protect the imagined “whiteness” of a static, idealized form of the game.
So that’s some disturbing shit right there that should make us cautious about casually getting into that particular argument. But even setting aside the racist implications, the way the debate encases certain styles of play and tactical approaches in moral shatter-proof glass strikes me as utterly contrary to the spirit of a game whose essence, if it has one, has been continual, dynamic variation from the moment the first game was played up through the present day. Both Russell and Chamberlain embody that variation, and each with a level of unsurpassed excellence. In that sense, finally, the debate is not just outrageous but profoundly saddening to me in that it reduces the enormous athletic and human richness and complexity of these two figures, on and off the court, to an infantile morality puppet-show. We, the game, and they deserve more thoughtful treatment.
And don’t worry, after this adventure in stat-land, I promise to go back home to incisive cultural analysis and never come back!