BertBrain Classics: Stats in Mixed Ultimate – Analyzing Team Progress with a Match-Up Ratio Lens

This article first appeared appeared in the BertBrain blog on Upwind Ultimate’s site in July 2017. The language has been updated to reflect current terminology and rulesets governing playing ratio in the division.

In 2016, Mixtape scrimmaged BFG, and the captains were discussing how a specific player performed. I went through the plays of the game in my mind, reflecting on the times I noticed his impact, and responded “I thought he did a really great job of attacking poaches on offense.” Then, the link to our UltiAnalytics stats was posted, and it turns out the player only had one touch during the entire game. Granted, he got that one touch through smartly attacking a poach, which facilitated effective movement by our offense, but my brain took the one event I saw and generalized it to be a theme. 

For Mixed teams that have set an intention to be equitable on and off the field, it’s often difficult to measure success objectively. At the root of the gender equity movement is the idea that we all have implicit bias ingrained in us by society’s messages that favors cis men as athletes. This means that we might either exceptionalize or minimize the contributions each segment of our roster makes on the field. We may generalize from a few touches women-matchers have on offense and think we’re involving the women-matchers on our team really well, or may miss seeing the routine contributions women-matchers make. We might also magnify turnovers made by the women-matchers on the team, since that plays into the internal biases we hold. 

It’s a little unpredictable what way our minds will spin the events we see on the field, and our perceptions will always be individualized based on our experiences and the lens we view the world through. I think this is actually a great thing, since every player will find a different facet of the game we can work on – I see that we need to do a better job of clearing space laterally, while another player points out that we can attack and clear poaches in a more systematic way. Both help the team in the end, and come from our different vantage points on the game! 

When we’re trying to evaluate our progress in a more concrete way, however, we can use tools like statistics to provide an important piece of the puzzle of what actually happened on the field. 

What to Record

When we’re thinking about assessing progress on equitable involvement of men- and women-matchers on the mixed field, we need to know more than just assists, goals, blocks, and turnovers. The best form of stats taking I’ve found is using the UltiAnalytics app, which lets you track touches by each person on your team for offense, plus blocks. An even better method would include having data from both teams’ use of the app, which would allow a more representative picture of defensive opportunities.

Why touches? Let’s take my role on Mixtape as an example. At the Pro-Elite Challenge, I maybe had one or two assists on the whole weekend, but that is by no means representative of my contribution to our D-line’s offense. Much of the time, I’m the one who picks up the disc after a turn and starts our offensive flow, and I act as the reset that switches the field for us. I end up touching the disc 2-3 times per offensive possession, but very rarely get involved in the formally recorded stats for the tournament. Recording and analyzing touches provides a more clear picture of true involvement in the offense by all players. 

Analyzing with a Gendered Lens

Here’s where we’re going to dig into a little math. This is just one way to play with the numbers, and if people with more of a handle on math have other thoughts on how to go about this, please let me know! You can download the csv file to get fine grain information if you’re handy with Excel, or get a more broad view of things through their web platform. 

The first piece of the puzzle is looking at handler versus cutter touches. This differs greatly based on what kind of offense you’re running. If it’s cutter-dominated, you’ll expect more cutter touches than in a handler dishie-based offense. I think a fun exercise is to estimate expected touches, then validate through the stats and see if you’re truly playing the kind of offense you think your team should be. Label each player as a handler or cutter based on where they play the most (swing players will add error to these calculations) and total up their touches as a group and find the ratio of touches of each group. If your O line and D line run different offenses, you can also subdivide this into two groups. This may be important if your O line generally has a 2:1 handler gender ratio, while your D line uses only men-matching handlers, for example.

Now we’re going to find the expected contribution of women-matchers to your team’s offense. This is where a lot of the error comes from, unless you go really deep in this analysis. Not all points are created equal, especially when comparing O points and D points together. If your player positions are significantly different between lines, it may be worth the reduction in sample size to get a clear picture of how each line uses all their players.

I can find the proportion of points played by women-matchers in each role and pair that with the handler and cutter touch ratios to get an expected percentage of touches by women-matchers. 

Assuming a game played solely with 4MM / 3WM, in an offense that is 50% handler, 50% cutter that always has 2 men-matching handlers and 1 women-matching handler, I’d expect 41.6% of touches to be made by women-matchers. In an offense that’s 20% handler, 80% cutter, I’d expect 46% of touches to be made by women-matchers.

It’s a simple matter, from here, of calculating the percentage of touches by women-matchers for the games you’ve selected and comparing that to your expected contribution. 

If you’re close to the mark of expected involvement, that’s generally a good sign that your offense is working to make space for everyone on the field. If you’re missing the mark, you have an evidentiary foundation for gameplay analysis, which can in turn guide the structure of your practices with more specific goals in mind. 


In 2016 at US Open, Mixtape played against Slow White twice. The first time, we won 15-12 in pool play; the second time we lost 15-10 in the finals. After the finals game, many women-matchers on the team, especially the O line, expressed frustration that they were getting looked off or didn’t have enough space to make their cuts. In going back and analyzing the stats, we were about 2% away from meeting our expected contribution by women-matchers in our pool play game, but almost 10% off the mark during finals. This specific statistical evidence framed the way we looked at the video from our finals game against Slow White, as well as the previous semi-final against Drag’n Thrust where we squeaked out a universe point win. As a result, we restructured our offense and how we taught downfield cutting skills for the remainder of the season, emphasizing how to clear space for each other. We also took a stronger focus on mental toughness and holding to our systems in higher stress situations. 

Continued re-assessment of your stats breakdown can help you understand whether or not your adjustments are having the impact you’re aiming for. Over time, your team will hopefully get closer to the kind of involvement you hope for across match-ups, and have actionable steps toward improvement and accountability. 

Bonus Steps

This system definitely isn’t perfect, since many players swing between playing as a handler and playing as a cutter, and increasingly teams will play four women-matchers as opposed to four men-matchers. If you have the luxury of resources (people, video, time), there are ways you can eliminate some of this error through calculating out each point’s expected contribution compared to the actual based on player positions during that specific point. One huge way to reduce error is to look at a larger bank of data than just one game, although as in the example above, sometimes looking at one game is extremely helpful. 

If you’re using UltiAnalytics, you can also click on individual players to see how they’re doing compared to what you’d expect from them. This isn’t a place to start throwing stones, but rather to analyze the way you’re setting up your offense and find places to teach important skills. For example, we expect a centering handler on offense to throw equally to men- and women-matching cutters if there are two of each downfield. If they’re throwing disproportionately to men-matchers in the cutting space, check how you’re setting up pull plays or the way you set up the stack at the beginning of the point. Do you provide equal opportunities for men- and women-matchers to initiate in the cutting space? Similarly, if a men-matching cutter is the initiator, they will need to throw a continue pass ⅔ of the time to a women-matching cutter. If they’re throwing to those women-matchers rather than needing to reset to handlers or find the other men-matching cutter on a regular basis, that’s one sign your offense is flowing successfully. What is reflected in the stats? If they’re not proportionately throwing to women-matching cutters, is there a way they can build capacity to throw to women-matchers better? 

One thing I wish we could track more effectively is data on turnovers. Who was the player trying to throw to? What kind of turnover was it – a poach block, execution error, or 1:1 defense block? This is something you can track independently if there are enough eyes on the game, or if you have film you can go back and review. Understanding what type of turnovers different players have would lend toward better practice design, whether those drills are along match-up lines or not. For example, a team could give those extra reps to men-matching cutters trying to throw to a tightly defended women-matching cutter, run a huck drill where women-matchers are throwing to men-matchers, or go through different poaching situations as a team.

As the division standardizes the use of rules where the team starting in a given endzone chooses the match-up ratio, it would be fascinating to use stats to decide whether it’s to your advantage as a team to go four women-matchers on defense. If you had detailed stats for opponents as well as your own, it would be possible to see relative efficiency on defense of blocks versus chances for each match-up type. For example, for points played with person defense, if women-matchers generated six blocks off forty throw attempts to women-matchers, while men-matchers generated eight blocks off eighty attempts to men-matchers, it might be in the interest of that team to play with four women-matchers on defense.


Stats can only take us so far, and should be used as a quantitative piece among a ton of qualitative data. Stats don’t take into account the role players have in clearing space on offense, differences in the strength of match-ups from team to team, and if players completely shut down their person on defense. 

They can, however, provide a starting point for your team’s analysis of a tournament. Combined with player perceptions and any film you’ve got to look back at a game, you can get a more complete picture of how you’re involving the whole roster in your play.