How Stats Are Helping Paralympic Teams Find An Edge

Wheelchair rugby used to be called Murderball, so that should tell you something about how the athletes feel about contact. Wheelchair basketball has its fair share of bangs and crashes, but it’s fairly similar to its “stand-up” counterpart. The two are among the most visible Paralympic sports, but how they inform their gameplay through statistics gives a window into how oft-used analyses are being adapted by disabled athletes.

Wheelchair basketball and wheelchair rugby have been fixtures of the Paralympic sports scene for decades. The two major material differences in wheelchair basketball are the equipment and the lack of a double-dribble rule. Wheelchair rugby is played on a basketball court, with the goal of pushing your wheelchair over the baseline between the edges of the key with a modified volleyball in your hands. The ball is thrown forward, with backcourt and shot-clock limitations along with hockey-like penalties — after all, the sport was concocted in a Winnipeg gym in the 1970s.

While advanced metrics like usage rate and player efficiency rating are commonplace for even the slightest of data-obsessed NBA fans, the use of advanced statistics in wheelchair sports is only relatively recent. In October 2019, Genius Sports, a software company that works with partners like the NBA and NCAA, announced a collaboration with the International Wheelchair Basketball Federation (IWBF) to integrate the sport into its platform. And both wheelchair basketball and wheelchair rugby are using advanced sports-specific data to inform their development processes.

Wheelchair Basketball Canada (WBC), the national organization for the sport north of the 49th parallel, has collated key statistical indicators for its athletes that date back to the first Paralympic games, in Rome in 1960. But the kind of work it’s doing now is a far cry from what it was when Mike Frogley, director of the WBC National Academy, first joined the program in the 1990s. Back then, simple physical testing, “subjective player ratings” and box score stats were the norm.

Subscribe to our sports podcast, Hot Takedown

“The types of data we have access to now are more refined,” Frogley said. “We are still learning how to use [them] as we’re developing some of these different tools.”

Frogley said the statistics that best predict the outcome of a game are 2-point field-goal percentage and turnovers; offensive rebounds are also a predominant factor on the men’s side, while the women’s game also pays close attention to assists and free throws. The program has split its analysis into operational components: technical skills, tactical skills, physical skills, mental skills, health, nutrition and equipment.

The Canadian wheelchair rugby team, under head coach Patrick Côté, has produced similar key performance indicators for its on-court success: goals scored per game, scoring rate (goals over time), turnover creation, plus/minus (or goals scored versus goals against), key offensive and defensive success rates,¹ and overall efficiency percentage (the team’s scoring rate versus its opponent’s). The team has per-game targets to hit in those and other categories. Côté said the team also measures the physical readiness of its athletes: “Power test, sprint test, agility test, endurance test — these are all part of our regular testing battery.”

Analysis in para sport can become complicated, however, when classification comes into play. Both wheelchair basketball and wheelchair rugby rely heavily on a classification model based on a person’s functional ability in a chair. Athletes are assigned point values based on their level of disability — from 1.0 to 4.5 in basketball and 0.5 to 3.5 in rugby. Low-point players, with a classification between 0.5 and 2.0, will generally have a lack of functional movement in their lower extremities. In rugby, all athletes have quadriplegia, so low-pointers will also have only some use of their hands. Wheelchair basketball athletes at the higher classifications often have disabilities that the classification system considers as having a lesser impact on performance. In rugby, many high-classification players are multiple-limb amputees, but they have full control of their trunks and can twist and turn quickly. In the middle of those two extremes are athletes with conditions like cerebral palsy or incomplete spinal cord injuries that cause inconsistency in how the effects of functional ability are measured in sports.

The points assigned to each player limit who can be in the game for one team at any given time. In rugby, 8 total points are allowed on the floor over four players. For basketball, it’s 14 points spread among five players. The aim is to allow for fair play in sports that feature athletes with different disabilities playing together. This system differs from sports like para swimming, in which all athletes of the same level of ability are in the same classification and thus the same race, or para ice hockey, which has only one classification system.

The Canadian rugby and basketball programs use many of the same drills to measure physical performance, including the 20-meter sprint.² While it’s rare to see a full sprint of that distance in one direction more than a few times per season — dodging at least seven other wheelchairs requires a lot of avoidance — the test tells coaches important information about the athletes’ power-to-speed ratio.

Côté observed that while the sprint times did not vary widely among classifications, the time measurements at 5 meters, for the first push of the athletes, did. The velocity of a first push — something like a first-step advantage between a cornerback and a wide receiver in football — is affected more by an athlete’s functional ability to turn the chair quickly using rotational speed. But many rugby athletes don’t have full trunk-muscle control, and the test also does not allow for measurements of deceleration, which plays a large part in the game. As Côté put it, “There’s only so much a stopwatch can tell you.”

Both teams use electronic sensors to measure splits. Wheelchair rugby has also started integrating this test after practices to measure recovery rates. And the teams are currently collaborating on a project to measure rotational speed using an inertial measurement unit sensor. But wheelchair setup is also a crucial component in how an athlete performs — less like a new pair of shoes and much more like a new car in a sport like Formula One — and that’s something that’s still being studied. And physical tests are only part of the puzzle.

“On more complex levels, for example in technical skills, we have something we call our ‘individual technical quality skill-measuring tool,’ and … that takes all the different skills that we teach on the basketball court,” Frogley said, “and it breaks down the behaviors in those skills, and they are assigned a point value to each of those behaviors.”

Frogley gave the example of determining court vision with four indicators and an assigned point value to go with each. Those factors are whether the athlete is seeing the ball; seeing the ball along with seeing the other players, which would require more head rotation; seeing open space and predicting the flow of the game, like where pick and rolls might set up; and looking at the rim to predict where a rebound might land. Skills like these are tracked for multiple possessions a game and allow for context. For example, a junior athlete’s court vision scores can be compared with those of a senior athlete to show where the junior athlete can improve.

Typical methods of sports evaluation like a lineup analysis are complicated in wheelchair basketball and rugby. Where an NBA team might track players in certain lineups in certain positions, rugby and basketball are forced to go by classification if they want a simple analysis.

Let’s take the recent International Wheelchair Rugby Federation (IWRF) Paralympic Qualification Tournament, which Canada won. Wheelchair rugby’s data featured six lineups.

The higher-classification players tend to have the ball in their hands more, much like a point guard in basketball, while lower classifications often play a role similar to a football team’s offensive line in wheelchair rugby. For example, according to the team’s in-house analysis, Zak Madell, Canada’s main 3.5, scored 132 of the team’s 394 goals, while the next closest, class 3.0 Fabien Lavoie, scored 50. Accordingly, rugby analysis has adapted to a wider view of the game so that the focus is not solely on the statistics most often recorded by those in higher classifications. Côté says the previous system of analyzing data, which was in place for Team Canada prior to 2014, did not create clarity for athletes who didn’t show up as much in the box score. This led rugby to focus on metrics like plus/minus.

“Low-pointers were disadvantaged all the time because they don’t handle the ball, and usually when they do handle the ball, it’s not a desired situation,” Côté said. “The plays were not meant for them to handle the ball, so it’s a bad situation to start with … and negative stats would be amplified.”

Wheelchair basketball has also begun to use a plus/minus model, from both a lineup and individual perspective. The complexity of lineup creation — due to classification, wanting to swap out one player might mean multiple substitutions to stay under the 14-point requirement — has also led to analyses of your lineup versus your opponent’s lineup, though Frogley says this sort of work is in its earliest stages.

Both coaches acknowledge that different program leaders place an emphasis on different statistics. But both say that the use of data allows para-sport coaches to look beyond what has been available and move toward a more data-centric decision-making process. They’re creating more data points and a fuller picture of how their teams are performing across the four-year time period all Paralympic sports are judged on, both in medal count and funding, and there’s hope among many in the leadership structure that the foundational tests being used can filter down into the junior and domestic game.