Win Shares: A Brionna Jones case study
Why there is more to the advanced metric than meets the eye
Thanks for reading the Her Hoop Stats Newsletter. If you like our work, be sure to check out our stats site, our podcast, and our social media accounts on Twitter, YouTube, Facebook, and Instagram. You can also buy Her Hoop Stats gear, such as laptop stickers, mugs, and shirts!
Haven’t subscribed to the Her Hoop Stats Newsletter yet?
If you told me you thought Breanna Stewart was the best player in the WNBA, I wouldn’t have a problem with that opinion. If you then told me that Brionna Jones was the second-best player in the WNBA, I would ask if you misspoke and meant to say Jonquel. Upon confirming that you did indeed mean Brionna, I would understand that you, like many others, had fallen victim to the trap of holistic metrics.
Statistics in sports have come a long way since the first baseball box scores in the late 1800s. Since then, the discipline of analytics has evolved from basic count stats to tracking data and wearable technology. Through every era of advancement, one thing has remained constant: the desire to create all-in-one comprehensive stats that can easily rank players or teams.
To preface the points I am about to make, I am very much a sports stats nerd, and I know that these metrics are invaluable to understanding player efficiency, individual contribution, and which on-court skills translate to winning.
However, it is immensely important to understand each statistic in the context in which it was created so as to not draw incorrect conclusions from their output.
Going back to the initial example, Breanna Stewart sits atop the WNBA leaderboard halfway through the season with 4.0 win shares, a value that attempts to attribute a chunk of each team’s wins to its players. Stewart is deservedly leading the way, currently averaging 22.0 points, 7.6 rebounds, 3.3 assists per game, and 1.9 steals per game, a stat line that has only been achieved once in league history (Maya Moore, 2014).
Not far behind with 3.7 win shares is Brionna Jones. Jones is having an incredible season, especially from an efficiency standpoint. She is on pace to be the first qualified player since 2017 with a field goal percentage greater than 56% and a free throw percentage above 88%. Factoring in her propensity to get to the free-throw line and stellar defense and Brionna Jones has rightfully earned her second career All-Star selection.
What makes her placement in the win shares leaderboard perplexing is that she is not even the best player on her own team. Or the second-best player. And arguably not even the third-best player. In fact, Brionna Jones does not even start for the Connecticut Sun.
Win shares are broken into offensive and defensive components. Jones leads the league with 2.9 on offense and has 0.8 defensive win shares. While the math behind how each number is calculated can get confusing, offensive win shares look at the total points a player helps produce relative to the number of possessions they are on offense and then compares that individual output to overall league scoring.
So what’s going on? Is Head Coach Curt Miller making a glaring mistake by not giving the best offensive player in the WNBA more minutes?
The Sun have a loaded front court with reigning MVP Jonquel Jones, four-time All-Star DeWanna Bonner, and two-time All-Star Alyssa Thomas, which keeps Brionna Jones on the bench for tip-off. While Connecticut does utilize a lineup where all four bigs are on the court at the same time, Jones still plays a considerable percentage of her minutes (46%) with the second unit.
What this rotational pattern means for Jones’s productivity is that she is often going up against the backup centers and forwards for opposing teams, who she is considerably better than. She does find herself on the court against better competition when playing alongside the Sun starters, but in these cases opponents devote more of their defensive resources to stopping Jonquel, Alyssa and DeWanna, often leading to easier shots in the post for Brionna as well as offensive rebound opportunities against smaller defenders.
I am in no way slandering Brionna’s abilities. She has maximized her opportunities in the role she has been given and would be a starter for almost any other team. It is just important to understand that the situation around her plays a role in her offensive contribution. If another one of the Sun front-court players swapped places with Jones, they might be even more productive than she has been.
And Jones is not the only example of a player benefiting from her teammate’s abilities. Attempting to separate individual contributions in a team sport is extremely difficult. Dearica Hamby currently sits one spot ahead of teammate Kelsey Plum in the win shares leaderboard despite the latter being a potential MVP candidate.
It’s even more challenging to evaluate players on the defensive side of the statistic. Assigning credit with just a box score is exponentially harder on defense, especially when steals and blocks are highly random outside of the few players that excel at each skill.
What defensive win shares, defensive rating, and the defensive component of John Hollinger’s PER all have in common is that they are more so a team statistic than a tool for comparing individual players.
Unlike on offense, basic game data does not attribute each basket allowed to a singular player. Therefore, each defender on the floor is deemed equally responsible for a made attempt. With few ways to separate teammates’ defensive ability, players who play a lot together generally have very similar defensive win shares, even if one is known as a better defender.
Last year, Los Angeles Sparks guard Brittany Sykes finished 2nd in Defensive Player of the Year voting and was 9th in defensive win shares. This season she is ranked 57th, with her two highest-rated teammates Liz Cambage and Nneka Ogwumike at 48 and 54th, respectively. Has Sykes gotten much worse at defense? Maybe she’s taken a step back from her standout year in 2021, but the Sparks as a team are allowing the second-most points per game and the third-highest offensive rebound rate, which Sykes cannot solely control.
Win shares are a great metric for quickly grasping a player’s contribution to their team, but it is important to always understand the full context in which a holistic individual statistic is calculated when evaluating a player.
Thanks for reading the Her Hoop Stats Newsletter. If you like our work, be sure to check out our stats site, our podcast, and our social media accounts on Twitter, YouTube, Facebook, and Instagram.
Thank you for the much needed additional context. I used to produce performance statistics for an insurance company. People's annual bonuses were largely based on my numbers. Yes, the metrics were useful to have, but they also contained obvious flaws, which I repeatedly pointed out to management. And you can bet the employees being evaluated would give me an earful about everything wrong with the measurements. Trying to find a balance between having data (so that you're not operating blindly) and using it properly is always a problem. Whenever I released a new batch of statistics, I always prefaced the post by saying my numbers were the starting point for a discussion, not the whole discussion.
The other thing I would point out about holistic sports statistics is that the vast majority of people who love to quote them have never looked at the calculations. The computation process is hugely complicated, and way beyond the comprehension of most sportswriters. So, people are often arguing vehemently based on numbers they barely understand.