Still Chasing the Ghost from Chicago? A Look from WS/48 on LBJ’s Distance to MJ
Around the end of 2020 I participated in a debate on the “LBJ v.s. MJ, who is the GOAT?” topic in a podcast. Here I share the research notes are better communicated in a written format and also share some further diggings after the debate.
What are the Issues?
The two issues this article intends to settle are the followings,
- Who has/had better teammates?
- Who has/had tougher opponents in the finals?
The motivation behind answering these two questions is that fans on both sides of the GOAT debate often use the same logic to attack the other side by saying either MJ or LBJ had/has better teammates or tougher final opponents to invalidate/validate the other side. It goes without saying that it is almost impossible to compare NBA basketball players played in different eras by just looking at the traditional stats, due to different rules applied, progress of playing styles, physical conditioning, technique developments, etc, and how these differences result in how the stats are distributed differently in different eras. Instead of looking at traditional stats, win share per 48 minutes (WS/48) is the advanced stat that is going to be used in this article to answer the aforementioned two questions.
Why WS/48?
A detailed description of WS/48 can be found on basketball-reference. In a nutshell, Win Share is a statistic that attributes the proportion of one win to a player, based on points generated and possessions. Win Share itself is an accumulated stats, and WS/48 turns it into a per game basis statistics. Since the calculation of Win Shares takes into account the league points per possession, league points per game, league pace, team pace, etc. and a win has always been a win, Win Shares is more invariant to changes of rules, progression of basketball strategies/philosophies, physical conditioning and techniques development. And the invariability of Win Shares makes WS/48 suitable for our discussions in this article, which is to compare players between different eras.
Before diving into the discussions of the two questions at hand, here is a quick comparison between MJ and LBJ with WS/48. Since his retirement, MJ has been sitting on the #1 seat at career WS/48 with 0.2505 while LBJ is at the #6 with 0.2331. Notably, Chris Paul is ranked #4 career wise and is #1 among active players. For career playoff WS/48, MJ ranks #1 with 0.2553 and LBJ ranks #3 with 0.2454. For single season WS/48, MJ ranked the #1 8 times in his 15 seasons of playing, and LBJ ranked #1 5 times in his 18 seasons of playing so far.
Who Has/Had Better Teammates?
To have a controlled comparison and focus on comparing MJ and LBJ and their peaks, we limit our comparisons of teammates and opponents to the seasons when MJ and LBJ reached the NBA finals (whether or not they won them). In the table below, teammates of either MJ or LBJ from these selected seasons are listed and ordered by their career WS/48 descendingly (only players with career WS/48 in the all-time top 250 are included, according to https://www.basketball-reference.com/leaders/ws_per_48_career.html). Rows in yellow are LBJ’s teammates and red rows are MJ’s teammates. The decision to use career WS/48 instead of WS/48 over the selected seasons might seem odd in the first place, but the reasoning is that career WS/48 summarizes a player’s contribution to a win no matter which teams he was on or what roles he played and thus better represents the players’s capability. In the meantime WS/48 for specific seasons are more correlated with team composition, line ups, injuries, etc.
By observing the table, it is obvious that LBJ has/had better teammates than MJ throughout the years, by the fact that the first 6 players on the list are all LBJ’s teammates. The high WS/48 does show that players such as A. Davis, D. Wade, C. Bosh, K. Love and K. Irving are/were all the top stars whether or not they are/were with LBJ. On the contrary, the distribution of MJ’s teammates on the list does represent the fact that they were never the go-to guys or 1st options whether or not they were with MJ or not. Even as great as S. Pippen was, Pippen was known for his elite “Swiss knife” capability to help a team on both ends of the floor, but never for his ability to draw attention offensively (due to the lack/inconsistency of shooting touch) and create opportunities for his teammates. In short, LBJ had better teammates when he reached the NBA finals than MJ had by looking at their career WS/48.
Who Faced Better Opponents in the Finals?
The same criteria are used to include opponents for either MJ or LBJ and the results are shown in the table below. Players in green rows on the left are LBJ’s final opponents, and blue rows on the right are MJ’s final opponents. To better visualize the results, different color shades are used to place players into three buckets in terms of all-time WS/48 ranking. Light shade (1~50th), medium (51~150th) and dark shades (151~250th). From the table, we can see that both LBJ and MJ had 18 opponents ranked higher than 250th on the all-time WS/48 list. LBJ had more top opponents (8 players in 1–50th) than MJ (5 players) had. It just so happened that the top opponents LBJ played against included all his 10 final appearances, while MJ played against his top opponents in 4 out of his 6 final appearances. Although LBJ only won 4 in 10 of his final appearances while MJ won all his 4 final appearances against the top opponents, we can say that LBJ did play against tougher opponents in general in his NBA finals.
If Career WS/48 Does Not Make Sense
The use of career WS/48 to compare players normalizes (to a degree) how injuries, lineups, roles, coaching, etc., may affect each player differently. But in our cases of comparing teammates/opponents for MJ and LBJ, it does sometimes over/underestimate the players’ Win Shares contribution when they played with/against MJ or LBJ. For example, the WS/48 of S. Pippen’s first and last three seasons are all below 0.1, which means his career WS/48 is definitely lower than his primetime during the Bulls’ two three-peats. The same could also be applied on D. Wade. To remedy the situation and take the discussion further, we use the playoff WS/48 for MJ or LBJ’s final appearance seasons to look at their respective teammates/opponents.
If only looking at the playoff WS/48 in their final appearances, MJ has 0.2674(6 seasons) while LBJ has 0.2466(10 seasons). For teammates and opponents, players averaged more than 15 minutes of playing time in the playoffs for those seasons are considered. MJ’s teammates averaged 0.141 in terms of playoff WS/48 in the Bull’s 6 final appearance seasons. The top three MJ’s teammates in terms of average WS/48 within the 6 seasons are Horace Grant (0.1843, 3 seasons), Ron Harper (0.1598, 3 seasons) and Toni Kukoc (0.1595, 3 seasons). LBJ’s teammates averaged 0.1175 in terms of playoff WS/48 in his 10 final appearance seasons. The top three for LBJ are Anthony Davis(0.284, 1 season), Chris Anderson(0.2404, ) and Daniel Gibson (0.211, 1 season). For opponents, MJ’s final opponents averaged 0.1209, and the top three opponent players were Charles Barkley (0.215, 1 season), Terry Porter (0.202, 1 season) and Shawn Kemp (0.198, 1 season). LBJ’s final opponents averaged 0.146, and his top three opponents were Kevin Durant (0.2452, 2 seasons), Stephen Curry (0.2103, 3 seasons) and Dirk Nowitzki (0.21, 1 season). By the above numbers, we might have a slightly different conclusion than the previous section where it’s actually MJ had better teammates and LBJ still had tougher opponents. But here comes a potential issue if you use the playoff WS/48 from specific seasons. The total wins of each playoff run weigh more than it does with the career WS/48, and in our context of comparing MJ’s or LBJ’s teammates or opponents, the teams that won the NBA championships in the end will likely have a higher WS/48 than the team on the other side of the finals. And is it possible that since MJ won all his 6 final appearances and LBJ only won 4 out of his 10, the Win Shares attributed to their teammates/opponents are tilted accordingly. To test this hypothesis, we will look at LBJ’s 4 NBA championship playoff WS/48 and compare them to MJ’s 6. In LBJ’s 4 NBA championship playoff seasons, his teammates averaged 0.1344 WS/48, and the top three are Chris Anderson (0.309, ), Anthony Davis (0.284, 1 season) and Kyrie Irving (0.21, 1 season). Opponent wise, they averaged 0.1303, and the top three are Kevin Durant (0.231, 1 season), Jimmy Butler (0.209, 1 season) and Kawhi Leonard (0.194, 1 season). LBJ himself averaged 0.2718 in his 4 NBA championship finals. The results are aligned with our hypothesis that for LBJ’s Championships, his teammates averaged higher WS/48 and his opponents averaged lower than the ones from his final appearances seasons. But the results still indicate that MJ had better teammates while LBJ had tougher opponents, albeit the differences are smaller now. The discrepancies between the conclusions from career WS/48 and specific seasons playoff WS/48 could still be explained by the smaller sample size of the later. In MJ’s 6 championship playoffs, his winning rate was 0.776 (90/116) while LBJ’s winning rate was 0.727 (64/88) and that could explain why MJ’s teammates had higher WS/48 than LBJ’s teammates had and the other way around for their opponents. In short, the 90’s Bulls with Jordan was just a better team comparatively to the other teams in that era, in a bigger margin than LBJ’s team compared to the teams LBJ faced. And now it makes sense if we look back at the conclusions from the career WS/48. Even though MJ’s teammates had lower WS/48 than LBJ’s, but the 90’s Bulls had always been absolutely dominating as long as MJ was on the team, and MJ had always been having the highest WS/48. LBJ, on the other hand, in his champion seasons, his team were never as dominating as the 90’s Bulls, even though he had teammates that were better than MJ’s teammates by looking at the career WS/48. Even LBJ was not the player on these teams having the highest WS/48 (Chris Anderson and Anthony Davis had higher WS/48 than LBJ).
Conclusions
For short, from the above discussions, the answers to the two issues we post in the beginning should be
- LBJ had better teammates considering teammates’ career WS/48.
- LBJ had tougher teammates according to the same data.
But the differences are small, and to settle the big GOAT debate (despite already tons of different arguments, fights, back-and-forths out there), we will recap with the direct comparison between MJ and LBJ in terms of WS/48 in the below table. Let’s skip the obvious part that MJ in general just has better WS/48 than LBJ, no matter if it’s regular season or playoffs, and talk about the only item that LBJ was higher than MJ, the championship playoff WS/48. Although LBJ had a higher WS/48 than MJ did, Chris Anderson and Anthony Davis had higher WS/48 than LBJ did in the respective playoffs as well, so it is actually the team composition that contributed more than LBJ’s individual contribution. For example, the 2019–20 Lakers heavily relied on Anthony Davis and LBJ on both ends of the floor while MJ’s Bulls were always more balanced.
A final note on WS/48. Win Shares is a stat that is suitable to measure achievements but not a good tool to understand what the players did or did not do. It is difficult to compare players in general, albeit players from different eras, even more difficult with players from an era where modern stats such as on/off impact, offensive/defensive ratings were not available. The source code used can be found here.
