top of page

Thanks for submitting!

  • Writer's pictureJo Clubb

Moving from Reactive to Proactive Training Load Management

You’re Mauricio Pochettino. It’s Saturday 23rd December 2023. As of tomorrow, your Chelsea side faces three Premier League games in just seven days, including two early, away kick-offs (at Wolves and Luton). Every Premier League game is a “must-win”, so do you opt for player consistency or rotation? How do you ensure sufficient recovery for starters and adequate stimulus for substitutes, while preparing everyone tactically for each game? Special cases, like Romeo Lavia who seems close to making a long-awaited Chelsea debut after injury, pose additional challenges.


Across the pond, Darvin Ham, Head Coach of the L.A. Lakers, faces a similar dilemma over the holiday season. From Monday 18th to Sunday 31st December, the Lakers face eight games, including five away games across two road trips. How do you manage the workload of star player LeBron James? Although James has been granted an exemption to the NBA’s new Player Participation Policy, it remains vital to maximise your star player’s involvement. But what does this involvement entail, given his age and recent injury history. Plus, how do you prepare others to play more if James plays less? For players not exempt from the league’s new policy, their load management strategy also requires careful planning to comply with regulations.

 

Training Load Projections

Effective load management anticipates and, where possible, prevents issues before they arise. I’ve often discussed the importance of careful planning around the team’s fixture calendar on the blog. Although establishing cause and effect in complex outcomes like performance and injury is difficult, a thorough planning process can offer a competitive advantage.


When I worked in the NHL, I projected season loads for players using rudimental training and game averages. I used their averages according to the training periodisation (Game Day -1, Game Day -2, etc.). In a league that, like the NBA, plays 3-4 games a week, training load patterns are predominantly driven by the competitive schedule, and this analysis helped to visually highlight the most congested periods.


A proactive approach to training load planning is a blend of science and art. We project physical load while also accounting for stressors like travel, jet lag, environmental conditions, and opponent, along with potential life stressors. Flexibility and gut instinct are also necessary to adapt to game outcomes, performances, injuries, and life situations.


Considering team schedules has never been more pertinent, particularly within periods of fixture congestion like those described above. This is exemplified by Erik ten Hag’s recent comments about the football schedule crossing “the limits of what players can handle”, with Premier League injuries said to be suffering a hangover from last year’s World Cup


There have been similar concerns in the women’s game. A recent Fifpro player survey reported that the majority felt they had insufficient rest before and after the Women’s World Cup earlier this year. Clearly, training load management requires a more calculated approach than ever before.

 

The Shift to Proactivity

Traditionally, load management in sports has been largely reactive. We analyse training load reports to ascertain who was “high” or “low”. Perhaps we assess athlete readiness the following day and then decide if adjustments are needed. In cases of injury, we retrospectively examine training load data for clues.


While these are all important aspects of a comprehensive load management system, a purely reactive approach is limited. It is akin to only looking in your rear-view mirror to see what’s past, rather than looking out your windscreen to see what’s ahead. Training load accumulated across a season requires ongoing manipulation to achieve adaptation, promote recovery, and reduce injury risk. Just like with driving, we need a combination of looking back and looking forward.


With the vast data now available, we can surpass my simple approach to load projections described earlier. We can do this on an individual player level, across a range of metrics, and with increasingly advanced data analytics. Artificial intelligence (AI), for example, is now being harnessed for this challenge.


A pioneering study (Connor et al., 2021) evaluated control strategies, based on control system theory, for adapting future training loads. The findings demonstrated that an intelligent closed-loop feedback controller outperformed other strategies (i.e. random and proportional control) at reducing the deviations from a training plan goal.


In collaboration with Zone7, we've also explored how AI can forecast injury risk and support training load management. Our series on AI in high performance sport is available on my blog here and in an e-book that can be requested here.


Zone7’s new tools put proactive workload management at the forefront. The Periodisation Planner, for example, allows teams to upload their specific training periodisation model, informing the algorithm of their typical routine (see figure below). We’ve previously emphasised the importance of context in AI-driven analysis. As such, the more information we can provide the algorithm, including a team’s schedule, periodisation, coaching needs, and time of the year, the more nuanced its analysis.  


A screenshot of Zone7's training load periodisation planner. Different microcycle lengths (e.g., 3 days, 4 days etc) can be selected and then the different periodisation, according to a sport team's training philosophy, can be entered for different external training load metrics.
Zone7's Periodisation Planner

The Periodisation Planner serves as a reference for the team's and position groups' periodisation. Complementing this, the Micro Cycle Simulator offers in-depth analysis. 


The Simulator is a 7-day bespoke planner (below), initially populated by typical workload based on the Periodisation Planner, and adjustable to explore how varying training loads might impact injury risk. For me, this tool could be used to support multidisciplinary discussions on the training schedule. Rather than dictating the schedule, it offers an additional resource to examine the possible consequences of different training approaches for the upcoming week. 


Different training load external load metrics, like distances, speed, accelerations, and decelerations are shown over a week's periodisation, with a coloured estimate of injury risk shown as determined by AI-driven data analytics
Zone7's Micro Cycle Simulator

These tools support a proactive approach to load management. But while injury risk mitigation is a key focus, we must balance this with the risk of becoming overly cautious. In our previous AI in sport series, we discussed how injury is not the only outcome we’re interested in. The training process should be evaluated and planned by the multidisciplinary team, with a hybrid approach, through the lens of both performance and injury risk.


Final Thoughts

Even in the era of pen and paper notational analysis, coaches and sports scientists aimed to fine-tune training programmes. With today’s data and technology, this process has evolved. We still deal with complex entities and limited knowledge of cause and effect, but data analytics can support training load assessment and manipulation. This integration of analysis and AI doesn't replace but enhances the expertise of coaches and trainers, offering deeper insights that complement their understanding of athletes' unique needs. Given today's growing demands on athletes, proactive load management should become the new norm in sports science.


コメント


bottom of page