Ranking the Worst MVPs by the Metrics
The MVP has become the NBA's preeminent regular season storyline, but sometimes the analytics and the voters diverge. These are the 11 MVPs where the voters lost the plot
Last Tuesday night, Joel Embiid claimed the 2022-23 NBA MVP and ended a three-year pursuit of the league’s top regular season individual award. Because it’s 2023, where there’s no such thing as a private moment, we all witnessed Embiid, surrounded by his teammates, learn that he had just won MVP. It was a sweet moment, Embiid, overcome with emotion, attempted to fight back tears, whether he won or not is unknown, as his teammates cheered, chanted, and congratulated.
Awards are silly. But they mean a lot to those who win them. If you believe they’re simply a vessel for joy, stop reading because I’m here to be as cold and heartless as numbers can be. I’m glad Joel Embiid won the MVP. He’s a great player, and you cannot tell the story of the NBA’s early 2020s without giving him serious airtime, but I also don’t think Embiid should have come anywhere near winning the MVP for a simple reason. He has now joined a not-so-illustrious club of MVPs.
Since 1980-81, the season media members took over MVP voting, there have been 11 instances where an MVP didn’t lead the league in one of the four volume or rate metrics available on Basketball Reference (Win Shares per 48 minutes (WS/48), Win Shares (WS), Box Plus-Minus (BPM), and Value Over Replacement Player (VORP)), and Embiid is the 11th. If your natural inclination is to be distrustful of analytics, here are a few numbers to consider. There have been 17 MVP seasons where the winner made a clean analytical sweep of the four metrics, five where they led in both rate metrics (WS/48 and BPM), and two more where they led in both volume metrics (WS and VORP). Over the past 43 years, 32 MVP seasons have seen the recipient lead at least one of the four metrics.
Years before analytics joined the basketball discourse, they did a pretty bang-up job predicting the MVP. They’re not infallible, but considering almost 75% of MVPs have been won by a league leader, dismissing them is willful ignorance on some level.
I was fairly confident Joel Embiid was going to win the 2022-23 MVP, but I also knew that Nikola Jokic led the league in WS/48, WS, BPM, and VORP. Based on the analytics, Embiid’s MVP case was pretty weak, which gave me a thought. Of the 11 MVPs not to lead the league in one of the rate or volume analytics, which one was the most and least egregious based on the metrics?
Methodology for Rankings
There’s no perfectly objective way to rank these 11 MVPs winners by how undeserving they were, but I think I landed on a solid methodology. The first is simply taking their regular season ranks in WS/48, WS, BPM, and VORP, adding them all up, subtracting four, and then comparing who had the lowest total overall. I called this rank score. The MVP is relative, if you’re the second-best player in each metric, maybe you shouldn’t win MVP, but you deserve to finish at least second. (I subtracted four because I wanted it to be how many people were ahead of the player.)
The next factor I considered is how far away they were from first in WS/48, WS, BPM, and VORP. Instead of just taking the difference and adding up all the negatives, I calculated the percent of the league leader(s) in WS/48, WS, BPM, and VORP they achieved and then averaged the figures. I called this Total+. The reason I preferred division over subtraction was to not punish a player as WS/48, WS, BPM, and VORP rose, which is important because we have two lockout-shortened seasons and two pandemic-shortened seasons. For example, if a player led the league in WS with 10.0, and the MVP finished with 8.0, that’s a difference of 2.0, but let’s say in another season the league leader reached 20.0, and the MVP finished with 18.0, that’s still a difference of 2.0, but in the first example the player reached only 80% of league leader’s production, while the second reached 90%.
Once I calculated Rank Score and Total+, I ranked them all again, then added their ranks together to get one number and ranked them for a final time with the lower the number the better.
11. Karl Malone 1996-97
Rank Score : 5 (T-2nd)* // Total+ : 92.48% (1st)*
The best MVP decision that didn’t feature an analytical leader was Karl Malone in 1996-97. Yes, Michael Jordan secured a clean sweep of WS/48, WS, BPM, and VORP, but Malone finished second in WS, WS/48, and BPM and third in VORP. His season metrics were also all within 90% of Jordan’s, the league leader. So while Malone shouldn’t have won MVP, it wasn’t a total disaster by the voters. Malone would later win the MVP in the lockout-shortened 1998-99 season, where he led the league in WS and VORP.
*Rank out of 11
10. Magic Johnson 1989-90
Rank Score : 5 (T-2nd) // Total+ : 88.93% (2nd)
Magic Johnson’s third and final MVP is similar to Malone’s in that he finished second in three metrics and third in one, but he was hurt by going up against peak Michael Jordan. Jordan, like in 96-97, performed a sweep of WS/48, WS, BPM, and VORP, but he was better in 89-90 than in 96-97, which hurt Magic’s Total+ score. While Magic’s WS and WS/48 were almost identical to Malone’s 96-97 campaign, his BPM and VORP were significantly better, but so were Jordan’s.
9. Magic Johnson 1988-89
Rank Score : 4 (1st) // Total+ : 81.14% (4th)
In 1988-89, Magic finished second to Michael Jordan in each of WS/48, WS, BPM, and VORP, which is how he has the best rank score of the 11. Although, he and Charles Barkley both finished with 16.1 WS. The problem, again, is that Michael Jordan was even better in 88-89 than he was in 89-90, and Magic was a bit worse in 88-89 compared to 89-90, which dipped his Total+ to fourth. In another era, Magic would have performed a clean analytical sweep, but there’s a reason people worship at the altar of Jordan.
8. Derrick Rose 2010-11
Rank Score : 16 (7th) // Total+ : 84.77% (3rd)
To be frank, I thought the Derrick Rose MVP would rank much worse, but he performs well in this methodology because the field in 2010-11 was historically weak. LeBron James achieved a clean sweep of WS/48, WS, BPM, and VORP, but by his lofty standards, it was one of the weakest seasons of his peak. Rose’s rank score is hurt tremendously by grading out poorly in WS/48 (10th) and WS (5th), but his BPM (3rd) and VORP (2nd) help him make up ground. His Total+ of 84.77% shows just how tight the field was. His .208 WS/48 ranked 10th but was 85.25% of LeBron’s league-leading .244. For comparison, Luka Doncic finished 10th in WS/48 this season with .204, but that was 66.23% of the league-leading .304. People have long complained that LeBron was robbed of this MVP, and they’re right, but there are far greater robberies in MVP history.
T-6. Joel Embiid 2022-23
Rank Score : 7 (4th) // Total+ : 77.53% (7th)
We have our only tie, and it’s between Joel Embiid in 2022-23 and Charles Barkley in 1992-93. Embiid’s rank score carries the load here. He finished third in WS/48, WS, and VORP and second in BPM. However, Embiid suffers as the analytical gap between him and Jokic, who completed his third straight analytical sweep, was massive. His Total+ of 77.53% is the lowest of any player with a single-digit rank score. Embiid largely won this award for the same reason the next guy did 30 years earlier; the voters were hell-bent on preventing another MVP threepeat.
T-6 Charles Barkley 1992-93
Rank Score : 8 (5th) // Total+ : 78.15% (6th)
Contrary to popular belief, Charles Barkley was significantly better on the Sixers than on the Suns, but because he won the MVP in his first season in Phoenix, the story of his career changed. Barkley’s rank score of eight came via two second-place finishes in WS/48 and BPM, and two fourth-place finishes in WS and VORP. While Karl Malone and Hakeem Olajuwon were close to Barkley in the two rate metrics, they bested him in the volume metrics through sheer force of minutes. Looming over the battle for second-best was Michael Jordan, who once again, pulled off a clean sweep of WS/48, WS, BPM, and VORP. Because of Jordan’s excellence at the top, Barkley’s Total+ score is weighed down.
Intermission
We’re a little more than halfway through, but we’ve gotten past what I would call the voter fatigue beneficiaries. All of these MVPs were won despite a player completing an analytical clean sweep (Jordan x3, LeBron, Jokic), and they all have one thing in common, limiting the total MVPs won by a single player. From 1987-88 to 1992-93, Jordan led the league in WS/48, WS, BPM, and VORP every season but only won three MVPs. LeBron was in the middle of a five-season run leading the league in the four metrics when he lost to Derrick Rose (he did win four in five), and Embiid won despite Jokic’s ongoing three-season run of claiming an analytic sweep and back-to-back MVPs. In reality, Jordan should have claimed all six of those MVPs, LeBron all five, and Jokic all three and possibly counting. I’m not entirely against rewarding great players with awards, but voter fatigue has functionally watered down the quality of an MVP season. Not all great players will win a championship, and not all great players will win an MVP.
5. Allen Iverson 2000-01
Rank Score : 30 (9th) // Total+ : 80.09% (5th)
According to the metrics, Allen Iverson’s 2000-01 MVP is one of the worst in NBA history. While A.I. was an MVP-level entertainer, his production on the court was not quite to that level. He finished 12th in WS/48, 10th in WS, and sixth in BPM and VORP. His rank score of 30 ended up ranked ninth as a result. However, the field in 2000-01 didn’t do much to separate, which is how his Total+ of 80.09% ranks fifth. Shaquille O’Neal led the league in WS and BPM and finished second in WS/48 and VORP in 2000-01. Shaq was coming off an analytic clean sweep MVP and probably should have taken home the award for the second straight year. Shaq remains the only player in NBA history to lead the league in BPM in at least four consecutive seasons and not take home multiple MVPs. Iverson over Shaq is the ultimate style over substance MVP result.
4. Hakeem Olajuwon 1993-94
Rank Score : 14 (6th) // Total+ : 65.91% (9th)
Hakeem Olajuwon is one of the greatest players in NBA history, but he was never really an MVP-level regular-season performer. His rank score of 14 saw him finish seventh in WS/48, third in WS, fifth in BPM, and third in VORP, and his Total+ of 65.91% shows just how far off he was from the league’s best. Adding insult to the voting is David Robinson completed an analytical clean sweep in 1993-94. However, Robinson would repeat the feat the next season and take home his only MVP. Hakeem did routinely perform better in the playoffs, which is amazing, and part of his legend, but his regular season production never should have led to an MVP.
3. Kobe Bryant 2007-08
Rank Score :18 (8th) // Total+ : 67.07% (8th)
Kobe Bryant’s lone MVP was a preemptive lifetime achievement award. He finished eighth in WS/48, fourth in WS, seventh in BPM, and third in VORP for a rank score of 18, and his Total+ of 67.07% shows he was miles off of the league leaders. As the league’s biggest star, Kobe winning an MVP felt like a forgone conclusion, but he never graded out well with analytics. Not once in his career did he lead the league in WS/48, WS, BPM, VORP, PER, or any of their offensive or defensive components. While no one completed a clean sweep in 2007-08, it wasn’t because no one was performing at an all-time level, it was because two players were. Chris Paul led the league in WS/48 (.284) and WS (17.8), and LeBron led the league in BPM (10.9) and VORP (9.8), which are all well above the average for an MVP season.
2. Steve Nash 2005-06
Rank Score :49 (10th) // Total+ : 63.56% (10th)
Fool me once, shame on you. Fool me twice, I can’t get fooled again unless it’s NBA voters and Steve Nash. Yes, Nash’s MVPs are the two most fraudulent in NBA history, but his second one was less fraudulent than the first, which I guess is a good thing? His rank score of 49 is hilariously low for an MVP. He finished 10th in WS/48, 10th in WS, 16th in BPM, and 17th in VORP, and his Total+ of 63.56% further illustrates just how far away from the MVP he should have been. Part of me understands how Nash conned the voters. He produced a 50/40/90 season, led the league in assists, led the Suns to the second-best offensive rating, and their league-leading pace of 95.8 possessions helped juice his traditional box score statistics. However, Dirk Nowitzki led the league in WS/48 and WS, leading the Mavericks to 60 wins and the league’s best offensive rating, and LeBron James led the league in BPM and VORP.
1. Steve Nash 2004-05
Rank Score :50 (11th) // Total+ : 61.85% (11th)
Nash’s first MVP simply doesn’t make any sense. He finished ninth in WS/48, 15th in WS, 11th in BPM, and 19th in VORP for a rank score of 50. There’s a real case to be made that he wasn’t even an All-NBA caliber player. Making his Total+ score of 61.85% even worse, was Kevin Garnett completed an analytical clean sweep that season, and Dirk and LeBron also had excellent seasons. Steve Nash owes KG an MVP, and he can pick between LeBron and Dirk over who gets the other one. You can’t tell the history of the NBA without Nash, but his winning even one MVP are the worst MVP decisions in NBA history.
Summary
From this exercise, I found two trends. The first is the presence of an overwhelming statistical behemoth that won multiple MVPs but should have won more. The second is there not being not a single clear-cut dominant player. The notable exceptions to this are Steve Nash in 2004-05 and Hakeem Olajuwon in 1993-94.
In 2004-05, the metrics pointed to Kevin Garnett winning his second consecutive MVP, but the Timberwolves won 14 fewer games than the season before. It wasn’t KG’s fault, but team success has always weighed heavily on MVP voting.
However, David Robinson missing out on the 1993-94 MVP to Hakeem is perplexing. The Rockets did win three more games than the Spurs, but Robinson had a season for the ages. He won the scoring title and secured an analytical clean sweep in a resounding fashion. The season ranks 18th all-time in WS, 15th in WS/48, ninth in BPM, and fourth in VORP (BPM and VORP go back to the 1973-74 season). His 19.98 Win Shares are the best single-season mark post-1980-81 behind three Jordan seasons and LeBron’s 2008-09, all of which ended in MVPs. Outside of Jokic’s 2022-23 and Jordan’s 1988-89, Robinson’s 1993-94 season is the greatest season since 1980-81, to not end in an MVP, made all the more surprising considering he had yet to win the award.
I hope this trip down memory lane and basketball history was informative and fun. All 11 of these seasons were fantastic and performed by all-time great players. The problem with basketball is a single player can dominate the league for years, and then there are stretches without a single dominant player extending their reign for more than a season or two. However, by and large, the voters did a good job without the help of widely available analytics. As more and more all-in-one analytics crop up and voters become better versed in their strengths and flaws, hopefully, we will never have another Steve Nash situation on our hands.