G.O.A.T. – The Limitations of Rankings

by

Over at the Basket-Reference Blog, there’s been a running debate in the comments of a fairly innocuous post about team dominance.  The debate, which has been very civil, concerns the Wilt-old question of who the Greatest Of All Time might be and about what criteria should be used to determine something as nebulous as the ranking of Greatness.

Old guys on the right, short guys in the front, Money, I mean Michael, in the middle.

The basics of the discussion are what you might expect: how do you rank the value of players? Is it individual skills, team impact, awards, stats, or winning or some combination? I think most people are willing to agree that you have to consider all of those sorts of things or you risk over-estimating certain players. You can overvalue statistics and call a player who has not been successful in the playoffs, like David Robinson or Karl Malone, better than more stat-sacrificing and clutch performers like Larry Bird and Tim Duncan. If you only care about winning titles you’ll wind up considering role players like Robert Horry or Ron Harper more valuable than Patrick Ewing or Elgin Baylor. The question becomes: how much weight should we give to stats, to winning, to skill-set, team impact, etc?

In the Basketball-Reference debate, one commenter took the position that Larry Bird has a better argument for G.O.A.T. than Michael Jordan on the grounds that Larry had a more immediate impact on team wins as a rookie and seemed to be more indispensable to team winning based on how Larry’s team performed when he missed time due to injuries versus how Jordan’s team performed without him in the 1994 season in which he was retired. It’s an interesting point that basically says, regardless of skills or production, Bird’s presence, his ability to help the team without amassing statistics, had a greater impact than Jordan’s. It’s sort of a plus / minus argument that uses wins and losses in place of points.

My response to this argument in the comments was:

Speaking more about the theory of how we determine who’s “better”, is team wins really the best measure? I guess the way I look at it, winning is largely circumstantial – a result of many factors rather than a reflection of a single player’s contribution. If you can’t explain how the team came to take its victories, then assigning credit to an individual and holding that as a trump over everyone else isn’t really determinant.

That’s why all these great statisticians are trying to find ways to attribute winning to players as individuals or lineups rather than just saying that this player’s team won so many games / titles, hence he’s greater than others. Longley has more rings than Hakeem, Robinson, Moses, Walton, but if we even glance at stat sheets we know he wasn’t as good a player. If you actually watch them play you can explain WHY he wasn’t better or more important. I think that factoring all three – results, stats, and observations is necessary. That’s why I don’t generally talk about players I didn’t get to see.

Obviously everyone falls on some continuum of talent and another of success, but it certainly needs to be looked into more deeply than team A + player X = Nwins vs. team B + player Y = Nwins-1.

It should be noted though that former NBA official Mike Mathis basically said that Larry Bird was the best player he ever saw because Larry did such a great job keeping his teammates in their sets and on the same page – enforcing team chemistry and compelling the group to be better than the sum of their parts. Again, it is an interesting point.

Another pick for G.O.A.T. in the debate was the great Bill Russell –

Bill Russell was a dominant player who, after arriving at USF, then with the USA Olympic Team and, eventually, with the Boston Celtics, always performed as though he knew exactly what the game of basketball is REALLY all about, i.e. #1. Rebounding, #2. Team Defense, and, then, #3. Team Offense. The consumate Winner.

It’s a concise argument that only includes one minor misspelling. The counterpoints include that the rebounding is massively inflated due to the pace and horrendously poor shooting percentages of the era which combined to create a ton of available rebounding opportunities (Russell’s per possession rebounds are leagues away from Dennis Rodman’s), and that it’s a little tough to look at a player who shot 44% from the field and 56% from the foul line and to call him the best to ever pick up a Spalding. Bill not only shot an abysmal percentage from the field and line, he also used a disproportionately low percentage of team possessions for a franchise player (which I guess is good since he shot so poorly). Career player comparison with Tim Duncan, who draws the most frequent style of play comparisons with Russ among active Hall of Fame level players (13 seasons each – both left college after 4 years):

Russ – 42Min / GM; FGA / GM 13 @ 44%; FTA / GM 6 @ 56%
Duncan – 37Min / GM; FGA/GM 16 @ 51%; FTA / GM 7 @ 69%

Considering that Russell must have scored at least a couple baskets a game on the offensive boards and a couple more as the trailer on the break for Coach Auerbach’s legendary fastbreak offense, and considering how few total shots he took, that shooting percentage really makes it look like Russell was not a player who could be relied on to score for his team in the half court. At all. The great team success starts to look like maybe Hall of Fame offensive stars Bob Cousy, Bill Sharman, John Havlicek, Tommy Heinson, and Sam Jones deserve a big chunk of credit. But I didn’t get to watch the ’60s Celtics, so I really don’t know.

But then again, Russell’s results are unmatched, so do we consider his production in this case, the fact that he didn’t have to be even a shadow of the offensive player that a Kareem or Jordan or Bird was in order to win 11 titles, or do we let it go since 11 rings are 11 rings, and he was still the unquestioned dominant force on the team?

On the other hand when you give too much credence to production, you get bogged down in what type of production is most valuable, and that is a particularly labyrinthine issue. We’ve gone over the various metrics and statistics in the past on Double Dribble, but to illustrate a particularly divisive metric, take a look at this post over at Wages of Wins where Professor Berri gives a list of the greatest individual seasons by a shooting guard as measured by his homemade metric, Wins Produced. Trying not to get into the math (which is really fairly simple), Wins Produced loves very efficient scoring, rebounding of all kinds, and steals, hates shot attempts and turnovers, and is relatively indifferent to assists, blocked shots, and fouls. I believe Berri would consider this a fairly definitive list based on statistical criteria he has proven relates to actual winning and losing.

They're fast. They're pretty. They can't possibly be beat. They must be the greatest!


It passes the smell test at first. Jordan, Jordan, Jordan, Jordan, Jordan, Jordan, Jordan… sounds right to me! But when you look further you see that, even excusing the seasons for Lever and Richardson who should probably be considered point guards rather than shooting guards, the next best true shooting guard by popular opinion, Kobe Bryant, doesn’t appear on the list until Sidney Moncrief has shown up twice, Clyde Drexler four times, and the immortal Brent Barry and Rodney McCray once each. Was Kobe’s best season really not as good as Brent Barry’s best season? According to Wins Produced, that is the truth, which raises questions. Is a metric that can find the value of a center or power forward really the right measure of a shooting guard? The positions in basketball may not be as distinctly defined as in other sports, but there is a distribution of labor in effect, and the weight attached to a given statistic for one position may not make sense for another. Wins Produced would prefer for all guards to abandon perimeter defense and go chase down defensive rebounds. Coaches would not like this as much. Wins Produced would also prefer for shooting guards to chase down defensive rebounds rather than get out on the fastbreak. That makes sense for bigs but not for smalls.

And not to pick on Wins Produced, which I think is a very clever and fairminded metric, but what does it tell us that Michael Jordan’s Wins Produced dropped from 32.95 down to 28.69 between ’89 and ’90 while his team wins jumped from 47 to 55. I think the metric could easily explain that increased Wins Produced from Pippen and Grant and the rest of the team made up for Jordan’s drop in production. But Wins Produced isn’t interested in whether Jordan remained capable of getting 32.95 wins or more but declined to for the good of the team. Phil Jackson introduced the triangle offense that year and moved primary ball-handling duties from Jordan to Scottie Pippen, and everyone thrived. Jordan’s efficiency increased. The team efficiency increased. This is a good thing.

Which brings us back to the commentator’s point regarding Bird being more valuable than Jordan. Sometimes what a player doesn’t do is as important as what he does and behind the scenes factors have a bigger impact on winning than an individual’s statistical production. I made this very point using Larry Bird (and towards the end Isiah Thomas) not very long ago in this post.

We clearly can’t base our Greatness solely on winning or solely on statistics and metrics. Both are important, but both ultimately are quite limited and ignore the other. It all comes down to weighting criteria.

It’s all a continuum between success and production. Bill Russell has an argument as the G.O.A.T. even though he had a massive gap in his game on offense that few other G.O.A.T. candidates did. His defense was so strong, and his team’s successes were so impressive that he ranks up there anyway. Wilt Chamberlain has an argument as the G.O.A.T. even though he has relatively limited championship success because his production was unrivaled. Then there are all the players in between like Michael Jordan, Larry Bird, Magic Johnson, Kareem Abdul-Jabbar, and Tim Duncan who have had lots of team success and great production to varying degrees.

The completeness and efficiency of a player’s game were always big things for me. Players who scored, rebounded for their positions, had decent assist to turnover ratios for their positions, and who could defend and perform in the clutch were the ones who caught my votes in rankings. That’s why I generally wind up firmly in the Michael Jordan court (or did I start out firmly in the Michael Jordan court and develop my criteria to back up the position – I can’t remember anymore). Here’s a player who can carry an offense with 30 points per game, make plays for others, get big boards at both ends, and play dominating disruptive defense or lock down a hot hand on the perimeter. He does it in the playoffs too, and he makes big plays at both ends in crunch time.

Peak Years Reg Season
Peak Years Playoffs

Of course you could say a lot of those things about a lot of players, with more weight in one direction or another depending on what you value.

Advertisements

One Response to “G.O.A.T. – The Limitations of Rankings”

  1. pmadavi Says:

    It’s hard for me to think of anyone besides MJ as the GOAT. LeBron’s the only other guy who has come close to completely dominating a game of basketball the way Jordan did – of willing a team to victory. You get the feeling Bird would have won at least one ring, on pretty much any team. But having a few hall of famers running with him sure helped. When you break it down, the combination of rings, stats, athleticism, and sheer will are really unmatched.

    That being said, I have no basis to say MJ is better than Russell. The game is so different. The circumstances they were raised in as well. Each generation has a GOAT. Jordan will likely be the GOAT for those born in the 70s until about the recent bunch of youngins who have not seen him play in his prime. Russell the GOAT before him. LeBron, likely, the GOAT after him.

    I do grimace each time Mark Jackson makes the argument that Kobe has surpassed MJ. The argument is ridiculous.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: