Patterns in All-NBA Team voting
Patterns in All-NBA Team voting
The NBA does not explicitly define what it takes to earn an All-NBA Team selection; there are no criteria to help media members fill out their ballots. Informally though, the criteria for an All-NBA selection probably align pretty nicely with the summary statistic, Win Shares, developed by Justin Kubatko of Basketball-Reference. Win Shares factor in a variety of individual statistics: points produced (scored + assisted), shooting efficiency, rebounds, defensive rating, etc. to divvy up team wins. Thus, players who make large contributions to teams that win a lot of games collect a large number of Win Shares. These are usually the same kinds of players who win All-NBA Team selections. We can empirically examine the relationship between Win Shares and All-NBA Team voting using the plots below. First up: the guards.
The relationship between Win Share ranks and All-NBA guard selections, 1996 to 2015.
[Note: Win Shares from 1999 and 2012 are prorated to an 82-game season in all plots]
In the plot above, green circles (and to a lesser extent, yellow and blue ones) indicate a high correlation between Win Shares and All-NBA Team voting. Hypothetically, perfect correlation would mean that the players selected to the 1st-Team were ranked No. 1 and No. 2 in regular season Win Shares, 2nd-Team selections would be No. 3 and No. 4, and 3rd-Team selections would be No. 5 and No. 6. If this were true for each of the past 20 years, the entire plot above would be reduced to six rows of green circles. Overall, there is a lot of green; 46 of 120 (38 percent) All-NBA guard selections matched their Win Share ranks exactly. Last year (2015, furthest column to the right), the six All-NBA guards were each among the Top-8 in Win Shares.
In contrast, red circles and their complementary white circles indicate a discrepancy between Win Shares and All-NBA Team voting for a given player. For example, in 2000, three of the six All-NBA guards finished outside of the Top-10 in Win Shares. In the past 20 years, several guards earned All-NBA Team honors, despite ranking 20th or worse in Win Shares.
The relationship between Win Share ranks and All-NBA forward selections, 1996 to 2015.
All-NBA Team voting has been even more correlated with Win Shares for forwards; 59 of 120 (49 percent) All-NBA forwards matched their Win Share ranks exactly (green circles above). For example, in 2000, 2005, and 2008, each of the All-NBA forwards was in the Top-6 for Win Shares among forwards. However, the extremes were also wider for forwards; Carmelo Anthony earned a 2009 3rd-Team selection, despite being the 43rd highest-ranked forward in Win Shares.
The relationship between Win Share ranks and All-NBA center selections, 1996 to 2015.
Every All-NBA center for the past 20 years has been ranked in the Top-8 for Win Shares. The relatively narrow band of Win Shares for the center position is due to the All-NBA Team structure — each team has only a single center to elect — as well as the relative dearth of quality centers in the modern NBA.
In summary, Win Shares are generally predictive of All-NBA Team voting. However, it’s also informative to focus on the instances where Win Shares did NOT predict All-NBA Team voting (the red circles above), to understand how the media votes.
Respect your elders
Ostensibly, the All-NBA Team selections are meant to recognize the best performances of an individual season; they are not career achievement awards. However, the data show that veteran players who have earned All-NBA Team honors in the past are more likely to overcome Win Share deficiencies than their less decorated, younger counterparts. For example, look at the centers plot again, with Dwight Howard’s career highlighted.
The relationship between Win Share ranks and All-NBA selections for Dwight Howard.
As you can see, from 2007 to 2011, Howard was either first or second in Win Shares among centers and he was rewarded with four 1st-Team and one 3rd-Team All-NBA selections as a result. Subsequently, from 2012 to 2014, Howard’s production dipped and he fell out of the top-4 in Win Shares among centers. Despite his diminished performance later in his career, Howard was still honored by the media with three additional All-NBA Team selections.
Injured stars get the benefit of the doubt
Along the same lines, great players who miss part of the season due to injury are sometimes given the benefit of the doubt when it comes to All-NBA Team voting. For example, here is the guards plot again, with Kobe Bryant’s career highlighted.
The relationship between Win Share ranks and All-NBA selections for Kobe Bryant.
Like Howard, Bryant has an impressive run of green and yellow circles during the peak of his career — representing All-NBA Team selections that were vindicated by Win Share totals — but, at the tail end of his career he produced more orange and red circles. Notably, Kobe also had a dubious 3rd-Team All-NBA selection during the middle of his career, in 2005. He played only 66 games during that season due to a sprained ankle and the Lakers missed the playoffs. Despite being the 13th-ranked guard by Win Shares, Bryant kept his All-NBA streak alive (his 7th of 15). Likewise, Scottie Pippen and Shaq O’Neal were given similar leeway by the All-NBA voters in their own injury-shortened seasons.
One Dimensional Standouts
In some cases, the voters seem to have favored players that were outstanding on one side of the court over players with better overall skill sets (and corresponding summary stats).
Career Win Shares and Box Plus-Minus Stats for Five All-NBA Selections, Offense vs. Defense.
For example, Carmelo Anthony, Tracy McGrady, Allen Iverson, and Tony Parker each earned All-NBA Team honors, despite being ranked 20th or worse in Win Shares for their respective positions. From the plot above, you can see that these four players have something in common; over the course of their careers, each of these players made huge impacts on offense and relatively smaller contributions on defense. As a result, their overall Win Share numbers were underwhelming relative to their reputations among the media. On the other hand, a few players like Jermaine O’Neal, got the All-NBA Team nod thanks to their outstanding defensive prowess.
Implications for 2016
So, turning our attention to this year’s All-NBA Teams. Is it possible to use the patterns from past All-NBA Team votes to predict this year’s squads? Well, let’s start by looking at the Win Shares; here are the Top-11 guards:
- Stephen Curry, 17.9 WS, two straight All-NBA selections
- Russell Westbrook, 14.0 WS, All-NBA selections four of last five seasons
- James Harden, 13.3 WS, three straight All-NBA selections
- Chris Paul, 12.7 WS, five straight All-NBA selections
- Kyle Lowry, 11.6 WS
- DeMar DeRozan, 9.9 WS
- Kemba Walker, 9.9 WS
- Isaiah Thomas, 9.7 WS
- Damian Lillard, 9.2 WS, one previous All-NBA selection
- Jimmy Butler, 9.1 WS
- Klay Thompson, 8.0 WS, All-NBA selection last season
Given the pedigree of the Top-4 Win Share earners, I think it’s safe to say that Curry, Westbrook, Harden, and Paul will be All-NBA honorees, but their relative order is uncertain. Curry, the back-to-back MVP, seems like a lock for the 1st-Team; whereas, the acrimony in Houston has probably soured voters on Harden a bit, dropping him to the 2nd if not 3rd-Team. To me, the debate between Westbrook and Paul for the other 1st-Team slot is a good one. Paul made a compelling case for himself in carrying the Clippers to the playoffs this season without his injured superstar teammate, Blake Griffin. However, his steady, no-nonsense brand of winning basketball has been overlooked by voters in the past. For example, in 2009 when his 18.3 Win Shares — best among guards and 3.6 Win Shares more than 2nd-best Dwyane Wade — was not enough to earn a 1st-Team nod. I suspect that Westbrook’s more highlight-reel-friendly style of play will again guarantee that Paul is bumped to the second tier.
With Jimmy Butler unlikely to crack the All-NBA echelon for the first time this season because of his injury problems, there are three other worthy guards remaining to battle it out for the last two openings: Lowry, Lillard, and Thompson. Lillard and Thompson each have three advantages over Lowry. Both are aided by their previous All-NBA selections (Lowry has none) and their exceptional ability for making long three-point baskets (in contrast to Lowry’s more all-around solid play). Moreover, I believe Lillard will receive a bump from the “he deserves it, he got snubbed during the All-Star voting” train of thought; whereas, Thompson will no doubt benefit from taking part in the best regular season ever. One caveat is that voters may be reluctant to pick all six guards from the Western Conference.
There are lots of great choices for forwards too, here are the Top-14 by Win Shares:
- Kevin Durant, 14.5 WS, All-NBA selections five of last six seasons
- Kawhi Leonard, 13.7 WS
- LeBron James, 13.6 WS, eleven straight All-NBA selections
- Draymond Green, 11.1 WS
- LaMarcus Aldridge, 10.1 WS, three previous All-NBA selections
- Paul Millsap, 10.1 WS
- Al Horford, 9.4 WS, one previous All-NBA selection
- Paul George, 9.2 WS, two previous All-NBA selections
- Gordon Hayward, 8.9 WS
- Tristan Thompson,8.7 WS
- Kevin Love, 8.5 WS, two previous All-NBA selections
- Marvin Williams, 7.8 WS
- Jae Crowder, 7.3 WS
- Anthony Davis, 7.2 WS, one previous All-NBA selection
Again, the Top-4 Win Share leaders look like solid candidates for the All-NBA Teams, with James and Durant likely to couple on the 1st-Team as they have done in five of the last six years. The outstanding seasons turned in by Leonard and Green (#1 and #2 in Defensive Player of the Year voting, respectively) seem to qualify both for an inaugural All-NBA selection. The serious contenders for the last two forward spots appear to be Aldridge, Millsap, Horford, George, and maybe Love. I’ll send Horford to the centers below, leaving Aldridge and George as the two most likely repeat honorees.
The list of Top-12 centers by Win Shares is noticeably bereft of previous All-NBA honorees:
- DeAndre Jordan, 11.5 WS, All-NBA selection last year
- Hassan Whiteside, 10.3 WS
- Karl-Anthony Towns, 8.3 WS
- Enes Kanter, 8.1 WS
- Greg Monroe, 7.5 WS
- Andre Drummond, 7.4 WS
- Marcin Gortat, 7.3 WS
- Jonas Valanciunas, 6.9 WS
- Nikola Jokic, 6.7 WS
- Dwight Howard, 6.6 WS, eight previous All-NBA selections
- Steven Adams, 6.5 WS
- Rudy Gobert, 6.4 WS
Given the lack of decorated centers, it seems likely that the media will be shifting forward-center hybrids into one or two of these center slots. Horford seems like a smart bet as a pseudo-center. I also like Jordan to improve on his 3rd-Team All-NBA selection from last season and I think either Whiteside or Gobert could be selected as a “best defensive center in the League” type pick.
My predictions:
1st Team: Curry (G), Westbrook (G), James (F), Durant (F), Jordan (C)
2nd Team: Paul (G), Harden (G), Leonard (F), Green (F), Horford (C)
3rd Team: Lillard (G), Thompson (G), Aldridge (F), George (F), Whiteside (C)