Women’s soccer velocity thresholds: statistical techniques or physiological metrics – context is critical.

I read with great interest the recent study by Park et al. (Fothcoming) entitled, ‘Velocity zone classification in elitewomen’s football: where do we draw the lines?’ As a researcher who has published extensively on women’s soccer demands (Vescovi 2012a, 2014, 2016; Vescovi and Favero 2014; Vescovi and Falenchuk 2018), sprint performance of female soccer players (Vescovi and McGuigan 2008; Vescovi et al. 2011; Vescovi 2012b), and recently published a commentary on the same topic (Bradley and Vescovi 2015) as Park and colleagues, I felt compelled to highlight several issues with the outcomes reported. First, it is commendable, the authors have used very sophisticated statistical methods – their expertise in this area is impressive – to develop different velocity thresholds for elite women’s soccer. However, they provide no supportive evidence that the new thresholds are an improvement over existing thresholds which are based on physiologically associated parameters in female soccer players (Bradley and Vescovi 2015). In fact, the newly proposed sprint velocity threshold resulted in 32–37 m of sprinting for a half of soccer in these elite female players. In contrast, when applying a threshold of 23 kph (comparable to the 22.5 kph recommendation by the authors (Park et al. Fothcoming)) in professional women’s soccer matches the average sprint distance was approximately 230–290m and 110–300m for players in the top league in the U.S. (WPS) (Vescovi 2012a, 2016) and Europe (UEFA) (Bradley et al. 2014), respectively. Even using themen’s sprint threshold (25 kph) resulted in an average of 170 m (Datson et al. 2017). Moreover, a player performing the typical number of sprint efforts permatch (i.e., 19–20) (Trewin et al. 2018a, 2018b) covering the average distance per sprint (i.e., 10–15 m) would cover approximately 200–300 m per match ... nearly 3-4X more than what the authors identified. Based on these data the face validity of the newly proposed sprint threshold (22.5 kph) is highly questionable as the authors seem to make inferences from statistical models without context, which results in an underestimation of sprint distance. Second, there is a substantially large velocity band for ‘highspeed’ running (i.e., 12.5–19 kph). It is difficult to believe that all locotmotor activities within this range can be qualified in the same manner. Indeed, the midway point of this zone (15.75 kph) is aligned with the YoYo scores presented in the methods by the authors (16–16.5 kph); closely approximatesmaximal aerobic velocity in similar cohorts (Ingebrigtsen et al. 2011; Haugen et al. 2014); and is nearly identical to the recommendation for high-speed running in women’s soccer based on physiological parameters reported previously (15–16 kph) (Bradley and Vescovi 2015). Thus, it is unclear how values that are approximately 20–25% above and below maximal aerobic velocity can be classified as the same locomotor activity (i.e., ‘high-speed’ running). Lastly, a methodological limitation of this study should also be noted that could have impacted the outcomes; namely, to treat data as missing when velocity and acceleration limits from male 100 m sprinters were exceeded. There are two issues to consider, first, it is not sex-specific and second, it is not soccer specific. Applying thresholds from male 100 m sprinters seems equivalent to applying velocity thresholds from men’s soccer to women’s match analysis. Moreover, sprinters are trained to accelerate for longer distances and can maintain top speed for up to 50–70 m. Unfortunately, this capacity differs in team sport athletes, where top speed is achieved <40m and sprints during a match are often <20 m. It would have been preferable for the authors to utilize maximum sprint/acceleration qualities from female soccer players. Indeed, maximum sprint velocity was 31–32 kph in female soccer players, which occurred during the final 15 m of a 35 m sprint test (Vescovi 2012b). As a result of the authors’ approach, there would likely be data included that exceed the reasonable limits for female soccer players. In turn, higher velocity thresholds would be identified and subsequently eliminate some proportion of sprint distance that should have been included, which is the likely reason for observing 32–35 m sprint distance in their cohort. I wholeheartedly agree that a consensus around velocity thresholds for women’s soccer matches should be developed so that the intensity distribution can be accurately quantified and direct comparisons between studies can be made. This should be based on experimental evidence that is interpreted within context and provides clear evidence for improvement beyond existing approaches. The current approach, albeit sophisticated, seems to be a time-intensive statistical technique that does not overcome the challenges of applying other thresholds and fails to demonstrate superiority when compared to female-specific physiological metrics that can be easily evaluated in cohorts across the entire developmental spectrum (i.e., sprint and YoYo test for maximal velocity and maximum aerobic velocity, respectively).