An Introduction to Complex Systems in Sport: Why It Matters for Sports Scientists
- Jo Clubb
- Apr 15
- 7 min read
This post explains the characteristics of complex systems and their importance within sports science.
Sometimes in applied sport, we implement what the literature suggests is 'best practice'. We've seen the evidence, we've heard others do it, and we feel we're being evidence-based. Yet the outcome does not behave as expected.
Performance plateaus. Injuries still occur. Or an athlete responds in a way that seems disproportionate to the stimulus or intervention applied.
In a previous British Journal of Sports Medicine editorial, myself and Will Greenberg explored why “best practice” is not always best in sport. The central argument was that many challenges in sport do not sit in simple or even merely complicated domains. They sit in the complex domain. And complex problems require a different way of thinking.
This article provides an introduction to complex systems in sport and outlines why this perspective matters for sports scientists, strength and conditioning coaches and clinicians working in applied environments.
From Reductionism to Systems Thinking
Sports science has been built, quite appropriately, on reductionist foundations. We have isolated variables, identified mechanisms and quantified relationships between inputs and outputs. It is the scientific process at work to test a single hypothesis, a single variable.
The move from monocausal thinking towards multifactorial models of injury, as I explored in my post on the evolution of injury causation models, represented an important step forward. Rather than searching for a single cause, we acknowledged that injuries arise from multiple interacting contributors. This stemmed from Dr Willem Meeuwisse's multifactorial model for assessing causation of injuries in sport (1994), in which intrinsic risk factors predispose an athlete, and extrinsic risk factors make them susceptible.
However, much of our methodology still rests on an implicit assumption: that if we measure enough parts and understand them precisely enough, we can explain and predict the whole.
The difficulty is that many biological and performance-related phenomena do not behave in a linear, additive manner. Bittencourt and colleagues (2016) proposed that sports injury emerges from a “web of determinants” rather than from isolated predictive factors. In this model, it is the pattern of interaction between variables, not the variables themselves, that gives rise to the outcome.
This distinction is subtle but important. Improving one component of the system does not guarantee improvement in the system as a whole. In some cases, it may even destabilise it.
Athletes as Complex, Adaptive, Dynamical Systems
Through my involvement as a mentor on the ALTIS Masters degree programme, I have been exposed to the ALTIS CADS lens: viewing athletes as Complex, Adaptive, Dynamical Systems. This framework has helped clarify why many applied challenges resist simple solutions. As frustrating as that may be!
Each component of CADS adds a layer of understanding:
Complex
Athletes are composed of interacting subsystems across physiological, biomechanical, psychological and social domains. These interactions are interdependent and frequently nonlinear. A change in one domain propagates through others in ways that are not always proportional or predictable.
Adaptive
Athletes continuously adapt to training stimuli, environmental stressors and competitive demands. Adaptation is not a fixed response to a fixed load, which is why I call this the Training Adaptability Prediction Problem. It is shaped by training history, recovery status, emotional state and contextual factors.
Dynamical
Athletes are time-dependent systems. The configuration of the system today is influenced by everything that has occurred previously. Training responses, injury history and accumulated fatigue alter the landscape of future responses. Athlete status is even changing minute-to-minute based on both training and life stress.
Systems
Athletes do not exist in isolation. They are nested within teams, coaching structures, organisational cultures and competitive calendars. This is nicely illustrated by Kate Yung and colleagues in their 2022 Sports Medicine Open editorial on Characteristics of Complex Systems in Sports Injury Rehabilitation.
Taken together, this lens encourages us to zoom out. It shifts the question from “Which variable caused this?” to “How is this system currently organised?”
The Cynefin Framework and Decision-Making in Sport
In our BJSM editorial, we drew on the Cynefin (pronounced kuh-nev-in) framework to illustrate that not all problems are the same. This framework is a Welsh-derived "sense-making" framework, developed by Dave Snowden, used to guide decision-making by categorising problems into domains based on cause-and-effect relationships. It helps leaders distinguish situations to match responses to the context.
The framework distinguishes between four domains: clear, complicated, complex and chaotic.
In the clear domain, cause and effect are obvious. Protocols are appropriate.
In the complicated domain, expertise is required, but relationships remain largely predictable.
In the complex domain, cause and effect are only apparent in hindsight (if even then). Interactions are nonlinear and outcomes are emergent.
In the chaotic domain, immediate action is required to stabilise the system before analysis can occur.
In the figure below, we provided applied examples to each of these domains from across sports medicine, sports science, nutrition, and strength and conditioning.
Many technical processes in sport, such as calibrating equipment or diagnosing a concussion, generally fall within the complicated domain. However, decisions around performance optimisation, load management across a season or return to play after injury often sit firmly within the complex domain.
The error arises when we treat complex problems as though they are merely complicated. More data and more precision do not necessarily create predictability if the system itself is inherently nonlinear.
Key Characteristics of Complex Systems
Several features distinguish complex systems from simpler mechanical ones.
Interdependence
Variables do not operate independently. Strength influences movement strategy. Movement strategy influences tissue loading. Tissue loading interacts with fatigue and recovery. Psychological state influences movement quality and risk-taking behaviour. Each element both influences and is influenced by others.
Nonlinearity
In complex systems, inputs are not proportional to outputs. A small change in sleep or stress may significantly alter performance or injury risk. Similarly, increasing training load by a fixed percentage does not guarantee a predictable shift in training response or injury risk.
Emergence
Injury and performance are emergent phenomena. They arise from the configuration of interacting variables at a particular point in time. No single factor is sufficient on its own. The outcome reflects the pattern.
Recursive Loops
Outputs feed back into the system as new inputs. An injury alters strength, confidence and movement strategy. These changes modify future exposure and risk. The system reorganises itself in response to its own outputs.
This dynamic, recursive nature is one reason why injury prediction remains so challenging. The system is continuously evolving.
A Practical Analogy: Cooking as a Complex System
ALTIS SEO and coach Stu McMillan often uses cooking as an analogy for performance.
Consider a dish composed of five ingredients. We start with each ingredient in isolation. However, the final outcome depends on proportion, timing, temperature and sequence. The ingredients are collectively elevated (hopefully!) depending on the interactions between them.
Too much of one ingredient alters the balance. Changing the order of addition changes the texture. Increasing the heat modifies the interaction between components.
In sport, our “ingredients” might include training load, strength, sleep, psychological state and environmental context. We frequently optimise each in isolation. Yet performance and injury risk reflect how these ingredients interact within a given context.
The quality of the outcome is not determined by the presence of individual components alone, but by the way they are organised.
As I'll present this weekend at the Isokinetic Conference, we also need to view rehabilitation and load restoration in RTP through this complex lens. While there is a place for our GPS targets and tracking tools in rehab, these are like our individual ingredients.
It is how these interact with each other, plus the wider integration of other aspects of load (i.e. psychological, cognitive etc) that determine successful re-integration to the complex and chaotic team enviornemnt. Ultimately, load restoration is not a linear rebuild of capacity, but a non-linear process of reintegration into a complex performance system.
Implications for Sports Scientists
Adopting a complex systems perspective does not mean abandoning measurement or mechanistic understanding. It means interpreting them within a broader frame.
First, it encourages humility in prediction. Screening tests and dashboards provide partial insight into a dynamic system. They do not offer certainty.
Second, it shifts the focus from isolated risk factors towards risk profiles. Rather than asking which single variable predicts injury, we can begin to examine which patterns of interacting variables tend to precede adverse outcomes (Bittencourt et al., 2016).
Third, it supports an iterative approach to practice. In complex domains, the appropriate strategy is often to introduce small changes, observe system response and adapt accordingly. This probe–sense–respond approach aligns with the realities of applied sport.
Finally, it reinforces the importance of context. The same intervention may produce different outcomes in different organisational environments. Equally, the same intervention may produce different outcomes in different athletes.
Final Thoughts
Sport, especially team sport, is not a simple mechanical system. Athletes are not machines whose outputs can be fully explained by isolated inputs.
Many of the challenges we face in performance and injury management sit within the complex domain. Recognising this does not weaken sports science. It strengthens it. It expands our mental models and aligns our methods more closely with the nature of the systems we are attempting to influence.
If we want to improve decision-making in sport, we must not only collect better data. We must also think more carefully about the systems those data represent.
FAQs
What is a complex system in sport?
A complex system in sport refers to the interconnected and dynamic network of physical, psychological and contextual factors that interact to influence performance and injury outcomes. These systems are nonlinear and adaptive.
Why is injury difficult to predict?
Injury emerges from patterns of interaction between multiple determinants rather than from single risk factors. These interactions change over time, making prediction inherently uncertain.
How does the Cynefin framework apply to sports science?
The Cynefin framework categorises problems into clear, complicated, complex and chaotic domains. Many athlete management decisions sit within the complex domain, where outcomes are emergent and not fully predictable.
What is a risk profile?
A risk profile describes a configuration of interacting variables that together increase the probability of injury. It reflects systemic interaction rather than isolated factors
Comments