Stephen Curry, the Most Improved, Most Valuable Player in NBA History?

With four minutes left in last Thursday’s game against the Oklahoma City Thunder, Golden State Warriors’ point guard, Stephen Curry, stepped to the free throw line — he had just been fouled on a layup that extended the Warriors’ lead to eight points and sealed the team’s record-tying 44th-straight home victory — as a chant echoed down from the upper level of Oracle Arena. It wasn’t the now-ubiquitous “M-V-P” chant with which fans all around the league serenade their favorite players, no, this was something new. The crowd in Oakland was shouting “M-I-P…M-I-P!!”. Apparently, with Curry looking like a lock to win the NBA’s Most Valuable Player award again this season, the fans were shifting the focus of their free-throw-line chanting campaign to another award due Curry: the Most Improved Player Award.

OK, they weren’t really saying “M-I-P”.

But, Warriors fans could be chanting “M-I-P…M-I-P!” or, maybe “M-I-M-V-P…M-I-M-V-P!”, clunky though it may sound. The extent of the Warriors success this season has been surprising and a big part of their improvement has been Curry’s astounding ability to follow-up his MVP-winning season from  last year with an even better performance this season. From the opening game, Curry has been looking to score more aggressively, taking more three-pointers and driving to the basket more frequently than last year. Despite his increased usage, he’s actually scoring more efficiently than last year, thanks to his increased three-point and free-throw attempt rates. He’s the league’s highest scorer and the leader in two informative summary statistics, Win Shares and Value Over Replacement Player.  Here’s the comparison between this year and last:

Curry, 2014-15: 23.8 ppg, 1.28 pts/poss, TOV%=14.3, WS=15.7, VORP=7.9

Curry, 2015-16: 30.7 ppg, 1.37 pts/poss,  TOV%=12.9, WS*=18.6, VORP*=10.1

*Projected

Curry’s continued development this season begs the question: is he the most improved, most valuable player in NBA history?

Measuring Value

To evaluate whether Curry is even more valuable this year than he was last year, I’m going to use Win Shares, developed by Justin Kubatko of Basketball-Reference, as my measure of a player’s value. It’s not a perfect player metric but it is a handy measure for comparing players going back several decades. Win shares are designed to measure a player’s individual contribution to his team’s success. When a team wins a game it earns one win share which is allocated to its players based on points produced (scored + assisted), shooting efficiency, rebounds, defensive rating, etc. By design, players who make large contributions to teams that win a lot of games will collect a large number of win shares. As such, win shares are pretty closely tied to the consensus, if somewhat nebulous, definition of the Most Valuable Player. Indeed, the Most Valuable Player Award is consistently given to a player among the league’s win share leaders.

The NBA Most Valuable Player Award goes to…

If we assume that Curry will win MVP again this year, he will be the 33rd Win Share leader to collect MVP honors (33 of 61, 54 percent). In the past, even when the Win Share leader hasn’t won the MVP, the winner was usually in the Top 5 (20 of 61, 33 percent) or at least in the Top 10 (6 of 61, 10 percent) for Win Shares.  Only twice has the MVP come from outside the Win-Share Top-10: in 1977-78 when Bill Walton injured his foot after playing only 58 games, but was still voted MVP by his peers, and in 2004-05 when Steve Nash was awarded the MVP by media members, perhaps somewhat dubiously.

Is Curry the Most Improved, Most Valuable Player Ever?     

Last year, Curry earned 15.7 Wins Shares during the regular season. This year, after 60 games, Curry has collected 13.6 Win Shares.  Based on his current rate of Win Shares per game, his projected Win Share total for 82 games would be 18.6 for 2015-16 (which may be an underestimate given the way Curry has played in the last few weeks). That represents a year-to-year improvement of a whopping 2.9 Win Shares. Curry has transformed from mild-mannered MVP, Wardell Stephen Curry Jr., to Supersteph:

Faster [release] than a speeding bullet,

More [skill]ful than a[n] [auto]motive [mechanic],

Able to [stretch] tall [defenses] with a single [downed]…[jump shot],

humiliator of three-point lines and destroyer of usage-efficiency curves .

Below, Curry’s heroic post-MVP win share increase is compared to all previous post-MVP win share changes.

A Boxplot of Changes in Win Shares the Year After Winning the NBA Most Valuable Player Award, boxes represent the 25th percentile, median, and 75th percentile win share change.

Note: All win shares are adjusted for season length and prorated to an 82 game season

In the 59 post-MVP seasons prior to 2015-16, the reigning MVP experienced a decrease in Win Shares 44 times (75 percent), whereas only 13 (25 percent) of the reigning MVPs were able to improve on their Win Share total from their MVP campaign. One-time MVP winners were particularly unlikely to replicate their MVP success: 16 of the 18 (89 percent) had fewer Win Shares the year following their lone MVP. This group features some of the steepest Win-Share declines due to injury: Bill Walton, -14.4 Win Shares (non-consecutive seasons), Kevin Durant, -7.6 Win Shares, and Derrick Rose, -5.6 Win Shares.

In contrast, players who won multiple MVPs tended to perform more consistently: 13 of 41 (32 percent) had more win shares in the year following their MVP. In addition, there is an elite group of players who won consecutive MVPs: Bill Russell (three-in-a-row), Wilt Chamberlain (three-in-a-row), Kareem Abdul-Jabbar (two-in-a-row, twice), Moses Malone, Larry Bird (three-in-a-row), Magic Johnson, Michael Jordan, Tim Duncan, Steve Nash, and LeBron James (two-in-a-row, twice). Not surprisingly, the consecutive MVP seasons posted by these players were the most likely to produce increases in Win Shares; during the seasons that were the second-half of a back-to-back MVP run, there were Win-Share gains in 9 of 16 instances (60 percent).

Even among this group of NBA legends, Curry’s projected 2015-16 Win-Share increase would stand out as elite. His 2.9 Win-Share gain would be the third-most ever following a MVP season. The only post-MVP seasons to feature larger improvements were 1963-64 Bill Russell and 1971-72 Kareem Abdul-Jabbar. In his consecutive MVP-winning seasons during 1970-71 and 1971-72, Kareem sky-hooked his way to 22.3 and 25.4 Win Shares, respectively (+3.1 win shares). His 1971-72 win share mark is the highest in NBA history!

Bill Russell increased his Win Shares (adjusted for an 80-game season length) from 13.8 during his 1962-63 campaign to 17.7 in 1963-64 (+3.9 Win Shares). Interestingly, the 13.8 Win Share total of his MVP season represented a low-water mark during a four-year stretch of 15.9, 13.8, 17.7, and 17.3 Win Shares (adjusted for season length). Moreover, Russell was only the seventh-highest Win=Share earner in 1962-63, when his peers awarded him the MVP. Of course, Russell was also coming off back-to-back-to-back-to-back championships that season and was on his way to his fifth of eight consecutive championships, so, perhaps, some forms of value transcend advanced statistics. At any rate, Curry’s improvement from last year is clearly one of the biggest leaps made by any MVP winner.

Where Does Curry Go from Here?

With no end in sight to Curry’s brutal assault on the three-point line this season, it’s hard to predict where his career goes from here. Could he shoot more threes? Could he score more efficiently? Could he be even more valuable? To answer these questions, it may be useful to compare Curry’s career arc to past MVP winners.

Career Arcs of Past NBA Most Valuable Player Award Winners

Note: All win shares are adjusted for season length and prorated to an 82 game season

On average, players tended to peak during their first MVP season. This is perhaps not that surprising, since, as discussed above, 18 of the 30 past MVP winners only won the Award once.  So let’s separate the arcs into two plots: one for winners of a single MVP Award and another for winners of multiple MVP Awards.

Career Arcs of Past NBA Most Valuable Players, Stratified by Number of Awards Won

Note: All win shares are adjusted for season length and prorated to an 82-game season

First, some historical notes. The two players who waited until their 12th season to win their first MVP Awards were Karl Malone (1996-97) and Kobe Bryant (2007-08). The player whose career continued 18 years after his first MVP Award was Kareem Abdul-Jabbar (1988-89). Kareem also won his first and last MVPs nine years apart (1970-71 and 1979-80)! What you can see from the plots is that the 18 one-time MVP winners (left) had much less success following their respective MVP seasons — during only 2 of 142 (1 percent) post-MVP seasons was one of these players able to surpass the win share total he amassed during his MVP season.

In comparison, the 12 players who won multiple MVP Awards (right) had more consistency after their first MVP award: during 26 of their 128 post-1st-MVP seasons (20 percent), these players were able to collect more Win Shares than they did during their inaugural MVP campaigns. That amounts to a rate of about two seasons per individual in which the player beat his Win Share total from his first-MVP season. Russell and Chamberlain beat their initial MVP-Win-Share mark most frequently, seven times each. However, each player won his first MVP early in his career — Russell in his second year and Chamberlain in his rookie year — so, it may be unreasonable to expect a similar level of sustained improvement from Curry. Indeed, generally, modern players tend to win their first MVP later in their careers than players from the 1950s and 1960s, which limits the extent of their post-MVP career arcs.

If Curry is able to continue his torrid shooting and win his second consecutive MVP award, he will join the elite group of back-to-back winners. Some of these players were able to continue improving, pushing their value to dizzying heights. Will Curry be able to maintain his new-found performance level like past legends, Russell, Wilt, and Kareem?

How Will Stephen Curry’s Career Arc Compare to those of Past NBA Most Valuable Players?

Note: All win shares are adjusted for season length and prorated to an 82 game season

Is Curry the Most Improved Player This Year?  

Okay, so Curry is definitely among a handful of the most improved, Most Valuable Players in NBA history, but does he really deserve to win this season’s Most Improved Player Award too? Well, Most Improved Players tend to start with relatively few Win Shares in the year before their MIP seasons, only 3.4 win shares on average. That leaves them with a lot of room to improve. The median Win Share increase during the 30 MIP seasons in NBA history was 4.5 Win Shares over the previous year (interquartile range: 3.7 to 5.4 Win Shares) with the maximum Win Share gain of 8.8 posted by Boris Diaw in 2005-06. One typical MIP season was turned in by Mahmoud Abdul-Rauf in 1992-93, when he went from 0.6 Win Shares the year before to 5.1 Win Shares during his MIP season (+4.5 Win Shares). By comparison, Curry’s projected 2.9 Win Share increase seems pretty modest (so, I guess that’s what Phil Jackson was talking about?).

Still, Curry’s 2.9 Win Share  improvement this year would be at least as impressive as six previous MIP winners. This year there are actually several MIP candidates with higher win share gains than Curry, including front-runners: Kyle Lowry (+6.7), Kawhi Leonard (+6.2), Jae Crowder (+5.1), Kemba Walker (+5.1), C.J. McCollum (+4.4), Hassan Whiteside (+3.9), Will Barton (+3.8), Evan Fournier (+3.5), and Reggie Jackson (+3.3). So, no, Curry won’t really be the Most Improved Player this year, but he’s sure having a great season.

All stats provided by Basketball-Reference