Warm the Bench
Posts
Anthony Davis, the first teammate since Carlos Boozer to outshine LeBron

Anthony Davis, the first teammate since Carlos Boozer to outshine LeBron

Tinkering with Win Shares to calculate player burdens

08 Jul

Photo credit: Tim Shelby from Arlington, VA, US / CC BY (https://creativecommons.org/licenses/by/2.0)

In his rookie season, LeBron James finished third in Win Shares on the putrid 2003-04 Cleveland Cavaliers, trailing Carlos Boozer and Zydrunas Ilgauskas. For the next 15 seasons, he would never fail to lead his team in this stat.

Until now.

Win Shares, as calculated by Basketball Reference, attempt to summarize an individual’s contribution to a team’s total victories. This stat, like PER, VORP, BPM, RAPTOR, and my own proprietary E.Y.E. T.E.S.T metric, provide a useful if imperfect shorthand for player value.

LeBron led the entire league in Win Shares from 2008 to 2013, peaking with a truly bonkers 20.3 total in the 2008-09 season that earned him his first Most Valuable Player award.

In this current season, LeBron ceded his WS throne to his superstar running mate, Anthony Davis. Per Basketball Reference, Davis posted 10.3 WS this year, just beating out LeBron’s 9.5 for top spot on the Lakers.

In fact, Davis produced perhaps the single best season among any eligible LeBron teammates—defined here as making at least 10 appearances and playing with Bron for the whole season. He leads all players in WS per 48, which adjusts the total WS to an approximation of impact per 48 minutes played. WS per 48 works better in this case as Davis’s season got cut short by the pandemic.

Before the season, I asked if Davis would be the best teammate in LeBron’s career, and he’s definitively proven that.

Davis thrived in every holistic advanced metric. He finished first in BPM (box score plus-minus), first in PER, and 2nd in VORP (value over replacement player) for all historical LeBron teammates.

My research into LeBron’s teammates sparked another thought; can/should you adjust Win Shares for the overall performance of the player’s squad?

Adjusting win shares by team success

Instead of merely tracking WS by player, I’m curious how this metric transforms by dividing each total by their individual team’s win totals. This theoretically continues to track production but now also highlights the burden placed on each player. A like-minded Reddit user tried to pull something similar a few years back, but for our purposes, I’ll extend the window of analysis from 2003 to 2020.

This adjustment—let’s call it WS Share, for the laughs—can help unearth players with solid production stuck on terrible teams and quantify the workload placed on an individual to get his franchise a W.

Here’s how LeBron’s WS Share looks over the course of his career.

This visual makes sense logically, as LeBron carried his early Cavs teams to any success before teaming up with more productive stars later in his career. He didn’t get any worse, but instead found teammates who could help carry the load. In 2019-20, LeBron ended up with a WS Share of 19.4%, while Anthony Davis edged out a slight lead at 21.02%.

WS Share did unearth two oddities that deserve further analysis.

Carrying the team like Greg Jennings

Since 2003, two NBA players contributed roughly two-thirds of their entire team’s season wins, calculated by WS Share.

In the 2010-11 season, Kevin Love posted a WS Share of over 67%, meaning that he’s contributed roughly two-thirds of the total victories that year for his Minnesota Timberwolves. He joins Brook Lopez (65%) as the only players in the entire dataset to post a WS Share over 50%.

In the land of the blind and winless, any stat can be king

Love played for the worst team in the NBA in his WS Share leading season, with the 2010-11 Timberwolves ending up a miserable 17-65.

He averaged 20.2 points and a league-leading 15.2 rebounds per game for the Wolves. He went on a double-double tear, hitting that mark in 64 of his 73 games per Basketball Reference. Love dropped 31 points and 31 rebounds in a rare victory over the Knicks. A Bleacher Report piece covering the game summarized Love’s statistical brilliance and Minnesota’s cursed season perfectly:

Love seemed to be amazed by his work. “It seemed like no matter what anybody did I was going to go and get those rebounds. I don’t know what to tell you,” Love said after the game. “I impressed myself. I don’t even know what to say to be honest with you.”

The significance of Love’s accomplishment was not lost on teammate Michael Beasley, who also turned in a spectacular performance himself by scoring 35 points and grabbing six rebounds.

“Thirty years from now I can say I was on the floor while history was being made,” Beasley said. “It’s great. It’s amazing. Every rebound came off to the guy. 20 and 20 is one thing. But 30 and 30?”

Unfortunately for most Timberwolves fans, they were unable to see the game as it was not shown locally in the Twin Cities market.

I can think of nothing more fitting for the Timberwolves than their sole star’s career-best night not even showing up in their home market.

Brook Lopez’s year doesn’t boast the same statistical output of Kevin Love’s 2010-11 season, but he did put 18.8 points, 8.6 boards, 2.3 assists, and 1.7 blocks for the twelve-win New Jersey Nets.

His high WS Share more reflects rare competence on a terrible team than a league-leading output, which seems to follow a broader flaw with the calculation.

Is this stat useful?

WS Share surfaces top performers on abysmal teams, but the methodology here may reward the players who lose more often. The challenge in adjusting for team wins is that often, the teams with the fewest total (and smallest denominator) end up producing the largest WS Share seasons.

Since 2003, only three players with top-25 WS Share seasons played for a team with a winning record.

Damian Lillard tops the league currently with a 33% WS Share. His Portland Trail Blazers might stand at 29-37 pre-Orlando, but Lillard still set career highs in points, assists, true shooting, and PER.

We should celebrate that!

Seasons like Lillard’s might not get MVP consideration and may soon get forgotten like Kevin Love’s rebounding master class in 2010. WS Share helps highlight underrepresented gems from neglected teams, but shouldn’t be used as a sole measure of success.

Instead, this stat could be useful as a small component in a more complicated scoring model, like my MVP scoring methodology.