A new trove of data looks at who is touching the disc in elite mixed play, throw-by-throw, across all filmed games from the 2019 season.
April 17, 2020 by Charlie Enders and Steve Sullivan in Analysis with 0 comments
Over the last half-decade, issues related to gender equity have frequently been front and center in the ultimate community. As we continue to grapple with the subject, one place in the sport that has received intense attention — both positively and negatively — has been the mixed division.
For some, seeing men and women competing with each other on the same field at the highest level is evidence of progress — or, at worst, a desire for progress — toward more equitable footing between genders. For others, the division only serves to highlight disparities and pay lip service to the cause, as men are more frequently team leaders and on-field focal points. Either way, conflicts of team chemistry, player relationships, and on-field conduct in the division need to be addressed through a gender-relations lens, primarily through one of equity. While these can certainly go both ways, women are often the ones feeling marginalized, both on and off the field.
We know that gender inequity exists in ultimate, just as it does everywhere in our society and world. But is the mixed division itself inequitable? Is it improving the disparity, perpetuating it, or exacerbating it? How can we know?
There has been plenty of anecdotal discussions of these questions for years, as well as a growing set of efforts to quantify it with stats. Five years ago, Jane Carlen laid out some thoughts about expected and actual contributions based on rudimentary goal, assist, and block tallies at 2013 Mixed Nationals. Then in 2018, Kathy Frantz conducted an undergraduate research project that presented a much more detailed set of data about the quantity and type of touches compared between genders in the mixed division. Both were limited in their own way by the quantity and type of data that was available, and both were subject of heated debate for their conclusions.
Now, we have another data point to add to the conversation: relevant statistics for every Nationals-qualifying mixed team in 2019. Over the course of last club season, Ultiworld, ESPN, and USAU filmed 55 mixed games across nine1 tournaments. Not only did this include footage of every team that participated at Nationals, it also included games from 14 additional Elite or Select Flight teams. I was able to watch and collect data from all of these games, building out a profile of each individual team filmed as well as the cohort in aggregate.
This new data may not provide any definitive answers to the tougher questions related to mixed and gender equity, but it does serve as an additional data set in the ongoing conversation.
Important caveat up-front: for simplicity’s sake, I will be referring to players and stats as only men and women. While I acknowledge that gender is not cleanly binary, the rules of the mixed division treat it that way.
In order to get a better understanding of how touches play out in the mixed division, I tracked two things in particular: total touches by gender and who is throwing to whom. Total touches by gender is self-explanatory and it’s superficially the easiest comparison to make. Originally it was the only data I planned on collecting, but halfway through the process, I realized that touches alone don’t really tell the full story. So I went back and recorded the gender breakdown of who throws to whom. Let’s look at the data piecemeal.
First, what qualifies as a touch? A touch is recorded when a player takes possession of the disc with a team’s intention. For example, a touch is recorded when a player catches a pull, picks up a dead disc, or receives an under. A touch is NOT recorded if a player receives the disc by accident, for example by making a catch-block or collecting an overthrow meant for someone else.
It is also important to note that the number of touches by a player doesn’t account at all for the value of the touch. Short reset passes don’t impact a team’s likelihood of scoring nearly as much as a long huck. Detailed throwing and receiving yardage statistics could provide additional context here, but they were beyond the scope of this research.
Gender Touch Breakdown 2019% of Touches by Gender on Nationals Mixed Teams in 2019
|Team||Games Logged||% of Touches by Men||% of Touches by Women|
In 2019, the average ratio of touches on mixed teams that qualified for Nationals was split 68.54% for men to 31.46% for women. Of course, some teams balanced touches more than others. Compared to the same measure in Frantz’s earlier study, elite 2019 clubs had a much closer split than those from 2014-2017 (74% for men to 26% for women).3 While I had a lot more concentrated data to work with, it does seem to indicate that the division, on average, is becoming less men-centric.
The teams studied in 2019 that failed to qualify for Nationals were slightly more balanced, at 66.72% of touches going to men to 33.28% going to women.4 However, given that the sample sizes are generally much smaller among this group and that there are a huge number of variables that factor into results, we make no suggestions that gender balance in touches is at all correlated with success one way or the other.
One explanation for at least some of the overall imbalance could have to do with role. Handlers typically account for a majority of touches on a team — these individuals can rack up 20+ touches in a single point against a zone look — which quickly skews numbers. While there were numerous teams that frequently operated with all-men backfields, I didn’t see a single team that relied exclusively on women in this role. In fact, of the 30 teams studied, only one — Seattle Birdfruit — had a game where their touches by women outnumbered the touches by men. Why? Because Birdfruit was the only team to feature a mostly women backfield. On almost every other team, women handlers were few and far-between.
Does that mean that the women in the mixed division aren’t as talented of throwers as their teammates who are men? Of course not — there are numerous incredible women throwers throughout the division. But it is clear from the data that very few teams put many women in positions to operate as central handlers and absorb lots of touches.
Given both anecdotal evidence and previous data sets, it is not surprising to find that men tend to touch the disc more frequently than women in mixed. If teams generally have more men as handlers, we can expect them to have more touches. But who are those handlers throwing to?
Several high-profile teams in recent years have been accused of playing “dude-ball” — ie, exhibiting a tendency for men to throw mostly to other men. But what qualifies as “dude-ball”? Is there a statistical threshold we can compare teams to, or is it more observational — just a gut-feeling that a team isn’t distributing the disc between genders fairly? While I think that defining “dude-ball” still falls squarely into the court of public opinion, we can share data on how frequently men throw to other men as opposed to women in elite mixed — and vice versa, in terms of who women throw to.
Gender Connections in Mixed 2019
|Team||M-M% Avg||M-W% Avg||W-M% Avg||W-W% Avg||WTEEM5|
In aggregate, almost 44% of touches on the mixed field among elite teams in 2019 are from one man to another. Both men and women are more likely to throw to a man than a woman, with just 8.36% of passes connecting two women.
One final data point I thought it would be interesting to highlight is the percentage of points a team played where women had more touches than men, labeled WTEEM.6 Fewer than 20% of the points logged in my study saw a team’s women touch the disc more frequently than their men. Of the 30 teams studied, only Seattle Birdfruit had the women touch volume exceed the touch volume of their men in a majority of points logged.
While this data is certainly informative, I don’t think we can make any broad statements about representation inequity or the state of mixed ultimate in general from it. The purpose of this study was more for collecting a baseline, an average touch ratio we can expect from elite mixed teams, rather than solely relying on circumstantial or anecdotal evidence. To be sure, looking at a single variable (like gender) runs the risk of oversimplifying an issue, especially one this dynamic and variable from team to team, level to level, region to region, etc.
It is important to recognize that this project is primarily one of education, not persuasion. We have assigned no value judgments to any of these data points, we are not suggesting that specific strategies or ratios are correlated with success, and are we certainly aren’t drawing any sweeping conclusions about what these on-field statistics mean about the state of gender equity for these teams, elite mixed, or the ultimate community generally.
We invite mixed teams to review the data, both in aggregate and in team-by-team comparison, and decide how much they align with team strategies, goals, and values. We hope that more — and more comprehensive — studies can be undertaken in the future so we have comparisons over time and more nuanced conversation about the data we do collect. The more we know about how opportunities are playing out on the field, the better we can address issues of equity.
Opportunities For Further Research
With more time to collect and dig into the data, there’s a number of other statistics I would’ve loved to look at and that could help paint a more complete picture of how genders are involved on the field. Several would even allow for more direct comparisons to previous data sets. These include:
- MORE GAMES in general – First and foremost, all of this data is more compelling across more games, where we can limit the impact of any weirdness related to weather conditions or mismatched competitive incentives in individual games. One game is just not a sufficient sample size to extrapolate anything. Sorry, Polar Bears. Let’s get more mixed games on film!
- Point-by-Point Gender Ratios – Knowing how many players of each gender were on the field for any given point would help us understand expected touch ratios if all players were considered equally likely to be involved. While that is obviously not the case, it’d also be fascinating to study which teams prefer which ratios and whether or how that correlates to their touch distribution. Observationally, the teams I saw that dictated four women most often were Birdfruit, Moondog, Drag’n Thrust, and AMP.
- Roster Composition – Similarly, looking at gender ratios another level up — how many players of each gender teams roster — could shine more light on both whether roster breakdown correlates to on-field contributions, and on historical trends. Anecdotally I think that elite rosters are trending towards a more even balance of men and women over time, but hard numbers could be useful in interpreting whether this has contributed to the slight uptick in touch balance over the past few years.
- Individual Stats – Cross-referencing touch data with individual stats like points played, goals, assists, blocks, etc. A “usage rate” could help us see if one or two individuals — regardless of gender — are skewing touches one way or the other for a team.
- Yardage gained by player – In terms of impact on win probability, the number of touches by a player isn’t nearly as important or valuable as the quality of the touch. Handlers can rack up a lot of touches, but if they’re just swinging the disc back and forth, little is really being gained. Meanwhile, the gigantic under cut that chews up huge yardage is a single completion, but it’s worth numerous uplines and dump-swings in a team’s probability of scoring. Ideally, we’d have access to yardage numbers along with the other data collected here.
- Stats from other divisions – We suggested that on-field role can heavily influence the breakdown of touches on a team. It would be interesting to compare player-level touch data from split-gender divisions to see just how much position/role can explain the imbalance in touches independent of gender.
While our studies didn’t look at exactly the same things, a similar 2018 study from Kathy Frantz looked at 18 filmed Nationals games from 2014-2017. I won’t go over all the details here (read it!), but her findings included:
- Men had roughly 74% of their team’s touches, while women had only 26%
- Teams that had more rostered women typically also had higher women touch percentages
- Men picked up a stopped disc 93% of the time when playing 4M/3W (84% when 4W/3M)
- 65% of hucks were thrown to men
In the end, Frantz offered three main takeaways, none of which we can confirm or disprove, but that may be relevant in a review of our new data:
“Overall, there is significant lack of parity between male and female players in every statistic. This is most noticeable in touches and hucks, where it’s apparent men have a much more dominant role offensively. Men also make more defensive plays overall, but they have a lower defense to offense ratio. This might support the traditional notion that women have more of a defensive role than an offensive one in mixed.
Not only are women underrepresented in play, but there’s no substantial difference in their contributions when playing 4M vs. 4W. In particular, women average roughly 20% fewer touches than men in both game settings. This might suggest that in-game play is not in fact determined by gender ratio but is more prominently affected by other variables.
Finally, having more women represented on a roster consistently leads to increased female involvement. Teams with a higher female roster percentage choose to play 4W more often, and they have a higher female touch percentage in both ratios. In all fields where roster percentage is considered against game play, there’s strong correlation between a higher female roster percentage leading to higher female participation in game play.”
Counting all Regionals as a single event. ↩
Why are there decimal points for games? Unfortunately, some points were missed in filming. To keep our per game results consistent, the amount of points missed was factored into a team's game total. For example, a fully-filmed game has a game value of 1. A game that missed a single point filmed (eg 25/26 points filmed) would have a game value of .96, as 96% of the game was filmed. ↩
See more about the main takeaways from Frantz’s study in an appendix below. ↩
See the Overall Sums & Averages document linked above for details. ↩
% of points where Women Touches Equal or Exceed Men ↩
Women Touches Equaling or Exceeding Men ↩