As we seek to prepare our athletes for competition, understanding and preparing them for their most intense demands is an important pursuit. We aim to add objectivity to this process through the analysis of time motion data. Such data has been analysed with increasing resolution; from entire games to halves, 15 minute, 5 minute, and 1 minute periods. As ever, having a perspective of both the big picture and the minutia is vital to practitioners.
A multitude of research now exists that explores such periods, which are frequently referred to as the “peak demands” and/or “worst case scenario” (WCS). A variety of methods exist, all with the intention of quantifying the most demanding passage(s) of play (interested readers are directed to this open access systematic review by Whitehead and colleagues, 2018). Moving averages have become a common approach, as fixed time epochs underestimate demands compared to rolling averages (Cunningham et al., 2018; Oliva-Lozano et al., 2021).
In light of this evidence base, the WCS has become a popular tool in applied practice. Expert consensus – via a Delphi survey – agreed that the WCS for high-speed and sprinting should be used as a benchmark for the training week (McCall et al., 2020). However, the group expressed concern over the definition and quantification of the WCS, and urged future work to explore the concept.
Analysis of the Worst Case Scenario
In response, a group of heavy hitters published a study in the Journal of Sports Sciences earlier this year, analysing the WCS approach. [You can request the full text on ResearchGate here] Novak and colleagues (2021) analysed optical tracking game data for a Premier League team across the 2018/19 season. The WCS was identified for total (TD), high-speed running (HSR), and sprint distance (SD) from the peak 3-min period for each player in each match. Interestingly, a 5-min rolling average was also analysed but omitted due to similarity with the 3-min results. Variability was assessed using the coefficient of variation (CV). A mixed effects model approach was used, with the three WCS as dependent variables and a variety of contextual factors set as fixed effects.
Uni vs multivariate approach
This study demonstrates that the WCS is not necessarily concurrent across different variables. For illustration, the WCS for a player in a single game may be approximately the 6th minute, the 35th minute, and the 80th minute for TD, HSR, and SD, respectively. Practically this makes sense in my opinion, as the most intense high-speed running period, for example, would not necessarily align with the most intense period from a total distance perspective, or even alternative demands such as deceleration or changes of direction.
This finding questions the ability of a single drill to replicate the WCS. Even if we acknowledge that we use the WCS to design conditioning exercises with a single variable in mind (e.g., HSR), we are not appreciating the multivariate nature of sport. Therefore, may be creating an unrealistic environment in terms of other demands, be them physical or technical, tactical, and/or mental.
High intra-positional variability
Practitioners may seek to use the WCS on a position level to design training drills according to the peak demands of each position. However, this study demonstrated high variability within position groups, with the variability increasing with higher speed demands. On a positional level, the CV as a percentage varied from 6.5-6.9%, 21.0-30.6%, and 35.0-56.1% for TD, HSR, and SD respectively. For example, one striker may have a SD WCS of 51m in a 3-min period, whereas another may have only hit a maximum of 15m SD in their WCS period. Based on this variability, positional benchmarks would not be recommended.
"Past outputs do not necessarily reflect, nor guarantee, future demands."
High intra-individual variability
Variability was not just high within position groups, but within individuals too, with variability also increasing speed demands. For example, the intra-individual CV for SD WCS varied from 21 to 76%. To illustrate this in a practical sense, one player’s (median) average SD WCS was 43m but their highest was 127m.
It is important to respect the variability in competition demands and not just focus solely on preparing players for their “average” demands. However, we should consider that past outputs do not necessarily reflect, nor guarantee, future demands. Indeed, the authors describe the danger of a recursive cycle in utilising the WCS as a benchmark, whereby physical preparation is based on prior demands and therefore, athletes are underprepared for atypical conditions. Yet again, we are in danger of naïve interventionism if we do not consider the potential downsides of our approach.
Considering collisions
Much WCS analysis, both in the literature and applied practice, has been calculated using time motion analysis. Many professional team sport leagues use local positioning systems to track athletes during competition, despite high contact demands, such as ice hockey and American football. Focussing on speed and distance demands underestimates the impact (no pun intended) of contact demands.
In one analysis of contact demands, Rugby League players were exposed to periods of 3 or more collisions per minute between 8 and 22% of the time in games (Johnston et al., 2019). In addition, collision and running demands may interact, with increasing collisions associated with a reduction in match speed. Practically, this makes sense as athletes are slowed down by the demands of contact. Therefore, focus on the WCS within the context of running demands alone, underestimates the true demands of such contact sports.
WCS should consider internal responses too
Novak and colleagues doubt that a true WCS can be determined without understanding the individual internal response. The individuality of the WCS necessitates the monitoring of internal load. In my experiences, internal load is often overlooked in the applied environment, possibly due to the relative ease of and/or access to external load technologies. Once again however, we are reminded that the combination of both internal and external load measures is important.
Final Thoughts
The WCS analysis has stemmed from practitioners and researchers seeking to understanding the most demanding passages of play. Quantification of this concept is an appealing prospect and this terminology – “worst case scenario” – utilises practical and easily understood language.
As is often the case with measures of training load, we seek to simplify the complex in order to action it in applied practice. However, in doing so, we can be at danger of an overly reductionist approach. As this study highlights, using benchmarks based on univariate measures from the past may overlook notable variability and actually limit an athlete’s ability to meet demands in atypical conditions.
Understanding an individual’s true WCS would require comprehension from multiple dimensions, potentially include time motion, velocity changes, collisions, technical/tactical demands, as well as the measurement of individual internal responses.
That said, I believe there is still a place for the current WCS approach, as long as we appreciate the limitations of the methodologies and the underlying complexity to measuring and monitoring such demands. I am in agreement with the authors that the use of specific terms, such as “peak locomotor demands”, “duration-specific running demands” or “high intensity phases”, may be better practice that the grandiose and all-encompassing title of “worst case scenario”.
