Understanding the pedagogic underpinning and knowledge sources of game form coaching in high-level team sport coaches

ABSTRACT

Across a breadth of underpinning theories, the use of game form training has been advocated as a feature of effective team sport coaching. However, a number of issues have emerged, including naïve interpretations of game-based pedagogies. As a result, there are questions regarding the utility of theory in practice. Using semi-structured interviews with 17 high-level team sport coaches, this study explored the pedagogic knowledge informing game form training and the sources of this knowledge. Participants reflected on a range of different theoretical positions that influenced their practice and suggested that knowledge came from a variety of mediated, unmediated, and internal sources. Thus, the findings seem to suggest that at the highest levels of team sport coaching, game form training is used in a highly individual and contextual manner. Thus, we suggest the need for mediated learning to move towards a more nuanced and deeper pedagogic understanding of game form training.

KEYWORDS:

• Coaching
• pedagogy
• coaching knowledge
• games based approach
• constraints-led approach

Sport coaching is a complex pedagogical process, with learning design and the approaches adopted by coaches at the centre of this process (Cushion, Armour, & Jones, 2003; Dempsey, Richardson, Cope, & Cronin, 2020; Jones, 2006; Nelson, Cushion, & Potrac, 2006). Given the complexity, significant research attention has been devoted to the knowledge bases associated with coaching and how this knowledge impacts coaching practice (Cushion et al., 2010; Cushion, Armour, & Jones, 2003; Potrac, Brewer, Jones, Armour, & Hoff, 2000). In practice, it has been suggested that coaches have two main categories of methods they can use to influence athlete development, their “behaviours” and their “learning design” (Muir et al., 2011).

As part of a coach’s learning design, practice activities in team sports coaching have been further categorised as being more “game form” or more “training form” (Ford, Yates, & Williams, 2010). Training form activities are classified as physical training, technique, or skills practices and game form activities are phases of play or small-sided/conditioned games (Ford, Yates, & Williams, 2010). The use of game form training has consistently been advocated as a feature of effective practice across pedagogic approaches and underpinning theory, for example Teaching Games for Understanding (TGfU), Game Sense, Games-Based Approaches (GBA), a Constraints Led Approach (CLA), and Non Linear Pedagogy (NLP) (Bunker & Thorpe, 1982; Correia, Carvalho, Araújo, Pereira, & Davids, 2019; Miller et al., 2017; Pill, 2013; Renshaw et al., 2016; Tan, Chow, & Davids, 2012). Despite significant procedural overlap between these approaches, there is widespread debate in the sport coaching literature regarding how theory should inform practice (Bergmann et al., 2021; Harvey et al., 2018). As such, an investigation into the pedagogic underpinnings of game form training and the sources of these underpinnings in team sport coaching is warranted.

Whilst the notion of using games as a feature of practice has a significant history (Worthington, 1974), the development of the TGfU framework acted as a catalyst for change in how sports are coached with the original TGfU authors countering a trend in physical education teaching towards behaviourist approaches, devoid of game context, tactical understanding, or decision-making (Bunker & Thorpe, 1982). Many variations have since appeared in the literature, including the later development of a GBA and Game Sense (Light & Curry, 2021; Stolz & Pill, 2014; Tan, Chow, & Davids, 2012). Underpinned by a more cognitive stance, TGfU encouraged participants to think, make decisions and develop a deeper understanding of the game (Bunker & Thorpe, 1982).

As a contrasting theoretical position to a range of cognitive theories, ecological dynamics is a theoretical model for understanding and optimising human movement, emphasising a direct (and non-representational) relationship between organism and environment (Araújo, Davids, & Hristovski, 2006; Michaels & Palatinus, 2010). Despite different theoretical presuppositions, the Ecological Dynamics perspective and the associated CLA and principles of NLP, there remains a similar recommendation of the use of game form training (Wood, Mellalieu, Araújo, Woods, & Davids, 2023). Supported by a burgeoning body of academic literature, both CLA and NLP have become increasingly prominent in coach education (Renshaw, Davids, Newcombe, & Roberts, 2019). A CLA suggests the manipulation of individual, task, or environmental constraints to facilitate skill acquisition through the perception of affordances and through individual action (Button, Davids, & Bennett, 2008; Correia, Carvalho, Araújo, Pereira, & Davids, 2019; Davids, Button, & Bennett, 2018). In other words, constraints being used as the boundaries which promote the emergence of functional interactions between an athlete and their environment (Pill, 2021).

Participant learning experience

Beyond the aforementioned approaches, additional theoretical concepts have been developed across the coaching literature that can support the coach in optimising athlete learning, thereby helping coaches moves beyond activity design to a consideration of the athlete’s experience. One example, which may in part explain the popularity of game for training, regardless of the underpinning theory or particular pedagogic approach, is the concept of high fidelity, or representative practice (Davids, Araujo, Vilar, Renshaw, & Pinder, 2013; R. Taylor, Taylor, Ashford & Collins, 2023), with fidelity being the degree of realism created by an activity (Choi & Wong, 2019) and representative design being: “practice tasks being representative of the interdependencies between the performer and the environment to enable the performer to learn the correspondence between the probabilistic proximal cues and the variable of interest” (Gorman & Maloney, 2016, p. 113). Fidelity is a multi-dimensional concept and has been used by both those using both ecological and more representational theories. Amongst other dimensions used in the literature, fidelity can be considered in terms of psychological fidelity (the psycho-emotional similarity to competition), conceptual fidelity (the similarity of problem solving requirements), action fidelity (how closely the actions of performers match those of competition), and physical fidelity (the extent to which training activity represents the physiological demands of competition) (Adamson, 2015; Dieckmann, Gaba, & Rall, 2007; Hochmitz & Yuviler-Gavish, 2011; Riccio, 1995). From a more cognitive perspective, this perceived similarity between training and performance task has historically been proposed to trigger the retrieval of mental representation, resulting in an increased likelihood of transfer (Gick & Holyoak, 1987).

In addition to the type of challenge presented by task fidelity, there is also a need for the coach to consider the level of challenge experienced by the participant. This may be related to the difficulty of the task itself (Guadagnoli & Lee, 2004; Hodges & Lohse, 2022), or the coaches’ scheduling of practice. The latter being termed desirable difficulties, the: “conditions that create challenges and slow the rate of apparent learning [to] optimise long-term retention and transfer” (Bjork & Bjork, 2011, p. 57). This makes an important distinction between performance and learning where performance refers to the temporary fluctuations in behaviour or knowledge that can be observed and measured during or immediately after the acquisition process (Soderstrom & Bjork, 2013). Learning, on the other hand, can only be understood based on retention and transfer (Bjork & Bjork, 2020).

From the cognitive perspective, it is also suggested that the cognitive load of high fidelity practice may impede learning (Choi & Wong, 2019; Norman, Dore, & Grierson, 2012). Similarly, the unguided exploration of a complex learning environment may also be detrimental to learning (Paas, Renkl, & Sweller, 2003). When applied to game-form training, high fidelity and minimal guidance might yield high activity levels but not necessarily learning (Kirschner, Sweller, & Clark, 2006). This would suggest that depending on the athlete, the coach should play an active role when using game-form training and use increasingly complex versions of the whole task in combination with the coach scaffolding and questioning (Cope & Cushion, 2020).

Pedagogic theorising

Despite this breadth of underpinning theory, naïve coach education approaches have often fallen foul of oversimplification in an attempt to impact coach learners (Cushion, Stodter, & Clarke, 2022; Stodter & Cushion, 2014). Similarly, a number of false dichotomies have emerged; for example, the need to abandon “traditional” methods in favour of those considered “contemporary”. These dichotomies have tended to focus on both dimensions of a coach’s practice (i.e. their use of coaching styles) and their learning design (Cushion et al., 2010). In terms of learning design, it is proposed that “traditional” coaching assumes that techniques must be developed before being implemented into a game scenario (Ford, Yates, & Williams, 2010; Kinnerk, Harvey, MacDonncha, & Lyons, 2018); therefore, the implementation of practice activities by coaches has typically progressed from reliance on training form activities towards more game form activities (Partington & Cushion, 2013; Partington, Cushion, & Harvey, 2014; Williams & Hodges, 2005). Where these dichotomies have emerged, they have been critiqued from a variety of positions (e.g. Cope & Cushion, 2020; Herrebrøden, Gray, Schack, & Bjørndal, 2024; Taylor, Ashford, & Jefferson, 2023), with the suggestion that there is “no single theory of learning that explains learning or the lack of it in all situations, and therefore, there can be no single approach to instruction” (Rink, 2001, p. 123). This is supported by recent work by Ranganathan and Driska (2023) who suggest that based on the current body of empirical work, choosing a single theory of skill acquisition may be detrimental; however, within the right context, activities that may seem suboptimal at one level (e.g. blocked practice which is often critiqued as lacking context for the learner) may be preferred at a different level of analysis (blocked practice allows the coach to view multiple athletes simultaneously). Solutions to these problems have been proposed in different forms with suggestions that coaches should be guided by a single theorical framework (e.g. Button, Seifert, Chow, Davids, & Araujo, 2020). Others suggesting a multi-theoretical perspective and greater conditional knowledge can support coach decision-making (Taylor, Ashford, & Jefferson, 2023). Conditional knowledge being an understanding of when and why different knowledge forms are useful (Schunk, 2012), and in turn, where some methods may work better than others (Collins, Taylor, Ashford, & Collins, 2022; Mosston & Ashworth, 2002; Pill, SueSee, Rankin, & Hewitt, 2021; SueSee, Pill, Davies, & Williams, 2021).

Sources of coach knowledge

Appreciating these perhaps diverging approaches to the intricacies of coaching, the differences between research and applied practice, and the construction and use of coaching knowledge are areas of importance to coach, researcher and coach developer. Coach learning has long been a topic of interest in the coaching literature with coaches being acknowledged as active constructors of their knowledge (Stodter & Cushion, 2017) by utilising a range of sources to inform their practice (Cushion et al., 2010; Dempsey, Richardson, Cope, & Cronin, 2020). These differences have been categorised as three distinct learning situations: mediated, unmediated, and internal (Werthner & Trudel, 2006). Mediated learning refers to situations in which the coach is directed to salient information (e.g. postgraduate coaching degrees), unmediated situations where the learner decides for themselves to-be-learnt material (e.g. observation of other coaches), and internal learning such as self-reflection (Werthner & Trudel, 2006). A prerequisite feature of effective unmediated learning being the coach’s ability to critically consider sources of knowledge (Stoszkowski & Collins, 2017). Critiques of the efficacy of the various forms of coach learning have regularly pointed to the inability to improve coaching expertise (Abraham & Collins, 1998; Nelson, Cushion, & Potrac, 2013; Vella, Crowe, & Oades, 2013). An example coming from North American coach education where more than 25% of programs were perceived to offer little in the way of application (Gano-Overway & Dieffenbach, 2019). In some cases, mediated forms of coach education have even been described as amounting to a form of indoctrination (Côté, 2006; Nelson, Cushion, & Potrac, 2006; Potrac, Brewer, Jones, Armour, & Hoff, 2000). The reason for this ineffectiveness, and the perception that research is divorced from the reality of coaching (Lyle, 2018), is attributed to a mismatch between academic research, its dissemination, and the reality of coaching experiences (Cushion, Ford, & Williams, 2012).

A particular issue reflective of these concerns has been naïve interpretations of the games-based and game form literature, regardless of theoretical underpinning. In practice, a result of this has been the devaluation of the role of the coach and the proliferation of memes such as “game is the teacher” and “just let them play” (Nash & Taylor, 2021). A consequence of oversimplification has been the sub-optimal interpretation of the nuances involved in the planning, design and application of games in training. Subsequently, “hands-off” approaches have been seen as being beneficial for learners and universally adopted by some coaches. This, despite multi-theoretical agreement that the coach should play a highly active role in shaping the learning experience of the athlete (Chow, Button, Lee, Morris, & Shuttleworth, 2023; Cope & Cushion, 2020). This leaves the coaching world with challenging questions, both regarding the extent to which coaching research reflects the reality of practice and, how coaches use theory in practice? As regards the latter, several studies have found a misalignment between some coaches’ intentions for practice, awareness of methods used and what they did during practice (Ashford, Cope, Abraham, & Poolton, 2022; Hewitt, 2015).

To better inform coach development and practice, there is a need to know more about the knowledge that coaches use and the depth of their understanding of any pedagogic underpinning of that knowledge (Stoszkowski & Collins, 2016). Addressing a recent call for further investigations into the use of “game form” practice by high-level coaches (Kinnerk, Harvey, MacDonncha, & Lyons, 2018), there is an opportunity to move beyond procedural understanding what coaches did and instead seeking to understand why they did what they did (Crandall, Klein, & Hoffman, 2006). Consequently, this paper had two specific research aims, firstly, to explore the use of pedagogic knowledge by a group of high-level team sport coaches in game-based practice and secondly, to examine the sources coaches used to acquire this knowledge.

Methods

Research philosophy

Framed by a Pragmatic research philosophy, which embraces a plurality of research methods and the belief that researchers do not seek certainty in complex situations rather they look for courses of action which would increase our understanding of a given phenomenon (Choo, 2016; Kaushik & Walsh, 2019), an interpretivist epistemological position was deemed appropriate as we sought to understand individual coaching experiences (Giacobbi, Poczwardowski, & Hager, 2005). Interpretive research, which considers all knowledge fundamentally subjective, thus allows researchers to understand participants subjective experiences and interpret their meanings (Markula & Silk, 2011). Given the need for a deep interpretative investigation, the application of qualitative methods to generate in-depth, rich descriptions of participants’ professional practice was deemed most suitable for this study (Patton, 2002). While there are advantages and disadvantages associated with all modes and forms of interviewing, we chose to conduct individual, semi-structured interviews (Markula & Silk, 2011) as they were most likely to produce nuanced, contextualised and specific knowledge.

Participants

Following ethical clearance from the authors’ institutional ethics committee (REC/2022/057), and based on the need to include coaches in our sample who would be able to provide informed accounts of their coaching process, the sampling criteria was designed to ensure that coaches could be justifiably considered to hold a level of expertise. Reflective of the need for a breadth of domain perspective amongst “high level” coaches we also sought a balance of coaches working in talent development and high performance domains of practice (Lyle & Cushion, 2016). In order to sample an appropriate level of expertise, we used a combination of the recommendations of Crispen and Hoffman (2016) and Nash, Martindale, Collins, and Martindale (2012) to formulate a set of inclusion criteria. Coaches were therefore recruited based on the following criteria: a) at least ten years of experience coaching; b) at least five years’ experience working with their current level of athlete; c) holding the highest possible coaching qualification in their respective sport; d) if working at talent development level, they must have a track record of developing athletes to the elite level, or if working in high performance, they must have a track record of winning elite-level competitions. In addition, given the focus of the study, all must have followed a mediated coach education qualification route that promoted the use of “game form” training. Participants were recruited through a mixture of the professional networks of the research team and snowball sampling (Onwuegbuzie & Leech, 2007) and drawn from a variety of team-based invasion sports across Western Europe to allow for a broad range of experiences to be explored. Each participant provided written consent prior to any interviews. As interviews progressed, sampling was also guided by the notion of information power with the recruitment of additional participants ceasing when we were satisfied that the contribution of new knowledge in the analysis had ceased (Malterud, Siersma, & Guassora, 2016). This led to a total of n = 17 team sport coaches being recruited as participants. These coaches had an average age of 41 (SD = 5.9), included 13 male coaches and 4 female coaches and represented a total of 5 sports: Gaelic football (n = 2), Hurling (n = 2), Soccer (n = 4), Rugby Union (n = 3), Hockey (n = 3), Netball (n = 3). To illustrate the relative “eliteness” of athletes being coached (McAuley, Baker, & Kelly, 2022), Swann and colleagues’ descriptors (2015) were used to demonstrate their athlete’s standard of performance. (see Table 1)

Table 1 Participants

	Sport	Sex	Level of athlete coached (based on Swann et al., (2016)	Years of experience
Lucy	Soccer	Female	Talent Development	18
Ben	Soccer	Male	Top tier professional league	24
Scott	Rugby	Male	Top tier professional league	10
John	Gaelic Football	Male	Top tier league^1	20
Colin	Hockey	Male	International Olympic level	26
Tom	Rugby Union	Male	Top tier professional league	20
Brendan	Gaelic Football	Male	Top tier league	31
James	Hockey	Male	International Olympic level	11
Steven	Soccer	Male	Talent Development	14
Stuart	Rugby Union	Male	Talent Development	14
Peter	Hockey	Male	High performance	21
Callum	Soccer	Male	Talent Development	16
Donal	Hurling	Male	Top tier league	17
Zara	Netball	Female	International level	20
Martin	Hurling	Male	Top tier league	16
Sophie	Netball	Female	International level	10
Susan	Netball	Female	International level	11

Research design

Pilot interviews were undertaken with two coaches in order to prepare for the challenges likely to be faced in the substantive study (Malmqvist, Hellberg, Möllås, Rose, & Shevlin, 2019). The pilot interviews were conducted with two high-level coaches who met all inclusion criteria with the exception of 10 years’ experience. Following the pilot interviews and reflection amongst the research team, questions were refined so that coaches could comment on their general use of games and provide specific examples. Contact was then made with initial participants via the professional network of the research team with email requests to participate in the study. This was followed up by the first author with a phone call to discuss the nature and purpose of the study and confirm the participant’s perception that their mediated coach education journey had emphasised the use of “game form” training.

All interviews were conducted by the first author, with each coach individually, allowing them to relate their own experiences in a free and open manner. The semi-structured interviews consisted of questioning around the coaches’ use of game form training, their underpinning pedagogic reasoning and where they perceived this knowledge to have been acquired. The interview guide included key, open-ended questions and probes were used during the interview to clarify, expand, and deepen the coaches’ responses and, thereby, increase the richness of their responses (Arksey & Knight, 1999). This approach allowed the authors to gain an in-depth exploration of each participant’s story, rather than a systematically structured account (Turner, 2014) and facilitated an appropriate balance between (a) allowing participants to share their experiences in a way that best reflected their reality and (b) achieving a necessary degree of focus and consistency across the entire interview set. The interviews were conducted using Zoom (Zoom Video Communications, San Jose, California, U.S.A.) and were transcribed verbatim. (see Table 2)

Table 2 Interview Questions

Question	Probes
To begin, I would like to know more about you and your work in coaching.	Give me a brief run through your career, with an emphasis on your coaching roles
How often do you use games in your typical training week?	What does your typical training week look like? Does this change depending on phase of season? Typically, how much of your training time is taken up in game form training? Why is this amount of training time given to game form activities?
Why do you use game form activities in training?	Are there any specific reasons you use games? What decides the selection of game-based practices?
Outside of training, what methods do you use to prepare players for game-based practice?	Discussion of individual player roles/tasks/needs? Prebrief the session? Debrief the session?
What are the typical outcomes for players when using game form training?	How do games benefit players? Are there disadvantages to games based training? How do you know this? How do you measure this? How do you maximise/mitigate these?
What is the role of the coach in games?	What is your role when game form training is taking place? What are your reflections on coaching/providing feedback vs using the game? Do you stop the game to give instructions or ask questions?
I’d like you to think of a game that you have recently used in training. How did you choose or design it? (Specific)	What were your reasons for playing the game? Theoretical underpinning? Trial and error? Seen someone else use it? Planned as part of a block of training? Do you use the same game across multiple sessions? What was the game built upon?
What was your criteria for success when using this game? (Specific)	How the players coped with the demands or can it be deemed a good game even if the players struggle somewhat with it? Is this normal? Have there been any occasions when you have looked for different outcomes?
Where have you found out about the use of games?	Where did you learn about using games in training? Do you do it how you were taught it, or ‘by the book’? Do you mix approaches?

Data analysis

A reflexive thematic analysis (RTA) (Braun & Clarke, 2019, 2021) was conducted to analyse the content of the interviews. RTA “is about the researchers reflective and thoughtful engagement with their data and their reflective and thoughtful engagement with the analytic process” (Braun & Clarke, 2019, p. 594). Open coding of data was focused on deriving semantic and latent codes. Both coding and theme generation used an inductive and deductive approach. This is coherent with our purpose and pragmatic epistemology given that inductive analysis, rather than being free of theory, is simply grounded in the data (Braun & Clarke, 2019). However, it has been argued that it is not possible to conduct a purely inductive analysis, as the researcher requires criteria to identify whether a piece of information is meaningful enough to code (Braun & Clarke, 2012). Therefore, deductive analysis, based on existing pedagogic theories, was also employed to ensure that coding and theme generation were both linked to the data and existing pedagogic theories (Braun & Clarke, 2019, 2021).

Once interviews had been transcribed, data was analysed using Braun and Clarke’s six-phased approach to RTA (Braun & Clarke, 2019, 2021). To facilitate deep immersion in the data, each transcript was re-read several times, and familiarisation notes were taken to ensure familiarity and understanding of the data. At this point pseudonyms were assigned to de-identify each participant. In the second phase, both semantic and latent codes were generated through multiple sweeps of analysis by both the first and second authors. Semantic coding being driven by coaches’ explicit reference to pedagogic theory, for example: “the forgetting curve” (Colin). Latent coding being the categorisation of implicit meaning behind text, for example “he’s had too much information. He’s now overloaded” (Steven). Qualitative analysis software (QSR NVIVO-12) was used to assist in the structuring, organising and analysis of raw data into their thematic hierarchies. Initial themes were generated from the codes as the third phase of analysis, though the process of coding and theme development was flexible and organic, evolving throughout the analytical process (Braun & Clarke, 2019). At the fourth step, these initial themes were reviewed and refined before the fifth stage which involved defining and naming themes based on shared meaning with the third author acting as a critical friend, challenging theme generation (Faulkner & Sparkes, 1999). The sixth and final phase consisted of writing the report, although given the reflexive nature of RTA report writing was recursive and woven into the entire process of the analysis (Braun & Clarke, 2019, 2021). Member reflections were solicited by email after each stage (Smith & McGannon, 2018). A total of 11 coaches responded to the member reflections. Some participants simply acknowledged the themes presented while others discussed theme generation offering further reflections. These discussions were then integrated into the main body of analysis and labelled as such. Given the interpretive and reflexive nature of RTA, analysis was predominantly conducted by the first and second authors, with the third author acting as a critical friend to audit the analytical process by sense-checking and exploring alternative interpretations.

Reflexivity

In all cases, the positionality of the authors can be framed as being neither insider nor outsider (Dwyer & Buckle, 2009). Whilst not being an “insider” of any of the coaching environments of the participants, none of the authors would consider themselves to be “outsiders” to the team sport coaching environment as each of the authors were experienced in team sport coaching.

Beyond the benefits of the research team’s experiences, a number of procedures were adopted during collection and analysis to optimise the trustworthiness such as the member reflections as described previously. This encouraged reflexivity by challenging interpretations of the data with a collaborative focus on rich interpretations of meaning, rather than consensus (Braun & Clarke, 2019; Byrne, 2021). Additionally, constant comparison was used throughout the analyses, both within and across different participants and their different interviews (Tuckett, 2005).

Results and discussion

The specific aims of this study were firstly, to explore the use of pedagogic knowledge by a group of high-level team sport coaches in game-based practice and secondly, to examine the sources coaches used to acquire this knowledge. Three themes were generated against the types of pedagogic knowledge used by coaches (learning design, coaching styles and desired player experience). Coaches also reflected on a range of different knowledge sources as being influential on their practice which also generated three themes (unmediated, internal and mediated). These themes and sub-themes are subsequently presented in detail with exemplar quotations.

Theories informing practice

Thematic labels were generated through a mixture of espoused theories, those that were specifically referenced by the coaches, and those that were actively generated through the process of interpretive analysis. Themes are presented in Table 3 and in italics throughout the text. A wide variety of disciplines were used to inform coaches’ practice ranging from educational psychology to motor learning. Coaches also demonstrated significant differences in theoretical underpinning with no coach in the sample holding a single theoretical position (Rink, 2001; Taylor, J.Taylor, Ashford, & Collins, 2023).

Table 3 Theories Informing Practice

Theme	Sub-Theme	Raw data exemplar
Learning Design	Constrain to afford	Yeah, so we would use constraints where necessary…for example, I wouldn’t say “you can’t play backwards”, because I think sometimes that takes the realism out of the game, it forces unnatural habits…. We might put constraints or restrictions, we might put rewards on that as well. So every time I receive and play forwards, I’ll get a point whatever it might be. (Callum)
	Perception- Action Coupling	If you define skill as both the binding of perception with action, so what you’re perceiving and your ability to execute on what you’re perceiving with action and for me, then the principle of specificity, so it’s got to replicate the demands of the game. (Colin)
	Fidelity and representative design	For me the best design of sessions are if your session can coach any topic or any player any particular skill (A) it’s realistic to the game (B) it’s got context and (C) it’s challenging. So that’s how I kind of judge success of a game as well is actually Can you and could you coach anything within that as well? If you can’t, it might not be realistic. (Callum)
	Layering of complexity	I’ll try and connect the warm up with the fact that we’re wanting everybody to defend. So like, the warm up game would generally be something to do with having to turn the ball over and score from it and in the session, as well, the technical skill part of the session will also be to do with their body position and that’s where it’ll be broken down a little bit, probably a little bit less games, but then the games probably would be reintroduced once we’ve gone through some of the technical skill. (Susan)
Coaching Styles	Appropriate feedback	I try not to do that (pull players out for feedback) because they want to play football and if you pull a player out of a game, I mean… I don’t do a lot, I’ll only do that if I feel it’s burning, like I really need to tell you something, because you’re not understanding it. Like, if I’ve tried to call it out a couple of times, like as they’re playing, but you know, this needs to be addressed now. (Lucy)
	Questioning	You can’t become a problem solver by just being told what to do all the time, so I’d ask a couple of questions. If we’re not getting there well then I probably haven’t asked the right questions. (Brendan)
	Explicit Instruction	Like we’re going to be on the pitch for 77 minutes, they’re going to do 41 minutes of work. So stopping a session or even in the break, to ask a question and develop an open conversation that like is just, it’s just not happening. Those conversations need to happen before the session or after the session or in a meeting or at lunch when you’re chatting or whatever. So a lot of the time it will be an instruction. (Tom)
Desiredplayer experience	Shared mental models	At least feeling a bit more confident about what’s the idea of these kinds of let me say “connected play”? How can we attack together? How can we defend together? How can we manage certain situations within these games, so that they are more, they are not always on one page but that their expectations and awareness become more aligned. (Peter)
	Cognitive load	If you’re just going into the gym to strength test, just literally turn up and zombie your way through it and just push out as big numbers as you can. And there’s, there’s no cognitive load with that. So actually, you know, is this a week where we can get them thinking lots on court, because we’re going to be working physically hard in the gym. And so, the cognitive load can be high on court, but we don’t need it to be super intense and always on the go. Or is it a week where we just want to, you know, you know, deload them cognitively or whatever. But trying to really think about that a little bit more, because I think previously we didn’t so much. (Sophie)
	Challenge point	I think errors are crucial. Like, if errors aren’t happening and an exercise is working too well in training then the exercise isn’t appropriate for the group, it’s too easy… I think errors have to be happening because they’re just so valuable. (Martin)

Learning design

This theme represents a range of theories that coaches used to inform activity design. A common tool was the manipulation of the conditions of games, or the use of constraints to adaptive player outcomes. Rather than being used as an entire approach to practice, constraints were used as a means of shaping the athletes’ opportunities for action, akin to the notion of constrain to afford (Renshaw, Davids, Newcombe, & Roberts, 2019). As an example, Callum outlines how constraints were used to generate individual player:

I do a certain type of game, because one player needs to work on his awareness and receiving skills. So my game might be constrained so that they have to play through a midfield area where he is

Constraints were also used as a medium of presenting affordances for groups:

we did a lot of work on lateral movement … I designed a couple of games … the grid was very wide and in terms of depth, very narrow. The only way they could move to receive the ball was laterally, they had no space to move forward to receive the ball, which meant you have to create space by moving lateral. [Donal]

Key to a CLA is the mutual relationship that emerges from interactions of each individual performer and the performance environment (Renshaw et al., 2016). In some cases, coaches seemed to use constraints in terms of reinforcement of learning in a behaviourist sense where the learner is characterised as being reactive to the conditions in the environment rather than taking an active role in exploring the environment (Ertmer & Newby, 2008). In other cases, coaches used constraints in a manner procedurally akin to the changing of conditions in practice (Wade, 1967).

We had a constraint around if you could kick and pin the opposition and put it in their third of the pitch, that was your score as if you’ve scored a try. So that was a constraint, you didn’t have to kick but if you did kick well and chased well, you would be rewarded for it. In-season our games would often try and link to something tactical that we’re maybe going after that week. [Tom]

The actions of Tom highlighted above could be interpreted with a variety of theoretical lenses, constraints were used to present players with opportunities for action that were closely linked to a specific game plan that players had generated a shared understanding of prior to practice. This finding may be a result of practitioners lacking understanding of the CLA (Renshaw et al., 2016), or as appeared to be the case for others, it was the blending of theoretical stances with affordances being seen under a theoretically, but perhaps not practically different lens (cf. Friston, 2022). This is supported by recent work done with swimming coaches which established that coaches use variations of a CLA while possibly lacking an appropriate theoretical understanding of such an approach (Brackley, Barris, Tor, & Farrow, 2020).

Participants also discussed the use of learning design geared towards the provision of perception-action coupling. Using an Ecological Dynamics lens, the coupling of perception and action is used as a means through which athletes engage in skilful behaviour as a result of the properties of the performer(s)-environment relationship (Tribolet et al., 2022). Here Colin describes how Ecological Dynamics informed their view of skill:

I would quite often describe the difference between pushing a ball to a teammate is just an action of pushing the ball, that isn’t practicing passing. As soon as you have an opponent and you’ve seen you’ve got limited space, and you understand whether the ball can get there or not, and then whether you can use it that now becomes a pass because there’s a decision to be made about how, when, how fast, the angle you use.

Yet, the perceived outcome of the coupling of perception and action was a change in athlete understanding. A mixed theoretical view that appeared to characterise the work of coaches. This blending of theories was exemplified by Colin who simultaneously uses elements of an ecological dynamics approach in the form of perception-action coupling, alongside cognitive psychology with explicit references to frequency of recall and the forgetting curve: “I felt very strongly around perception-action coupling and ensuring that you’ve got the right context in front of you. So, really using that theory, then using sort of theories of learning, so you know, frequency of recall and the forgetting curve”.

A central feature of a variety of motor learning literature is the importance of high fidelity practice to support transfer of learning (Tribolet et al., 2022). This theory suggests that practice environments should adequately replicate the performance environment to ensure the same degree of functionality and fidelity exists between both domains (Gorman & Maloney, 2016; Pinder, Davids, Renshaw, & Araújo, 2011). This mirrors decades worth of recommendations in coaching pedagogy, with the suggestion that greater “realism” of practice will support transfer (Worthington, 1974). Whilst conceptually different from an ecological dynamics perspective recent work indicates that coaches believe practice should be representative (Brackley, Barris, Tor, & Farrow, 2020). Yet in practice, achieving consistently high fidelity was a prominent challenge for coaches:

Ideally, it’s as close to the game as possible. It looks as close to the game as possible, it feels as close to the game… How can you get the game looking like the game and feeling like the game without the contact level being up to game standard? That’s a challenge. [C10]

Interestingly a number of coaches described the barriers to achieving fidelity in practice. Peter outlined barriers to true fidelity in training sessions:

Because of knowing each other so well … everybody knows what the other one won’t do, will do and is capable of doing … we cannot copy the way of behaving from umpires and opponents and pressured situation with having these consequences in mind about the results. We try to be as realistic as possible … then if you’re always training with the same players, I know that this guy has a big reach, and I know I have to do this and I’m used to this. But if I’m playing someone who I haven’t played against, and he’s a tall guy, I’m assuming he’s capable of doing this and that, but I might be surprised if he is fast, far more speedy, far more skilful or whatever.

The provision and extent of practice fidelity was informed by coaches using an approach best understood as the layering of complexity; how coaches planned and sequenced learning design over time, or over the course of a session. In the latter case, participants discussed the use of different sequencing of whole-part activities where part practice refers to breaking a practice activity into smaller, more manageable units and whole practice involves practicing an activity in its entirety (Fontana, Furtado, Mazzardo, & Gallagher, 2009; Van Merriënboer & Kirschner, 2017):

I’m gonna do a drill, based on a receiving skill and then I put that into an opposed setting to challenge the execution of that skill. So giving the player tools and a toolbox, and then within that competition of me against you, that then challenges my decision making skills to pick the right skill from my toolbox, depending on the context [Callum]

Yet, it was also apparent that regardless of the perceptual-cognitive complexity presented to athletes, that coaches were informed by the need to ensure that skills acquired in training are similar to those in performance, or the notion of action fidelity as originally suggested by Riccio (1995) and developed by Stoffregen, Bardy, Smart, and Pagulayan (2003). Coaches also used different sequences of whole-part activity to layer desired complexity. Here Steven reflects on the failure to appropriately layer complexity:

The disadvantage is the game becomes the game … that might sound ridiculous because we are playing the game, so we should be teaching the game but I think if you are just playing the game, and there’s no kind of idea behind it and you just keep doing that, can people lose clarity of direction? I was always for game based practice, themed game practice, teaching the game in that way. That was my bias. I would never use drills. As I’ve got a little bit older, I’ve started to use drills, but I’ll only use them with multiple outcomes. I wouldn’t put generic practice on because it’s passing and receiving and I want to get loads of passing and receiving out. I want to get repetition of pass and receiving here … What are my tactical outcomes? Can I use eight minutes of drill based practice to get repetition to get a base layer understanding, to then evolve it into the themed game where you get three outcomes, but you also might find another outcome, your kids might discover something different.

Steven implies that the inadequate layering of complexity may lead to the idea that “the game is the teacher” with coaches using game form activities, with little underpinning reasoning.

Coaching styles

Additional to coaches’ manipulating learning design, they also reflected on their coaching styles and the deliberate intent behind their approach (Pill, SueSee, Rankin, & Hewitt, 2021). Finding an appropriate balance between coaching input and time on task created a tension between the need to create game-like conditions and appropriate feedback: “I think you lose a lot of opportunities to teach and coach because it’s become more, let the game flow. Let the game be the teacher kind of attitude” [Callum]. As a result of the need to balance coaching input and time on task, coaches reflected on the need for a bandwidth of feedback (Patterson & Lee, 2013) to ensure that neither the game became the teacher, nor did they overwhelm players with a large volume of interventions:

We don’t want to overload them … what are the key take-homes I want to get out of this session and make sure then that I’m not throwing too much at them. So I think two key outcomes coming out of a session is plenty. I would never go for more. [Brendan]

Theoretically, ensuring that athletes engage in problem solving processes themselves, rather than becoming reliant on coaching input is a key concern in the literature. In addition to knowing when to give feedback, the type of feedback used by a coach is a core feature of support required by athletes (Mason, Farrow, & Hattie, 2020). Much of the GBA literature promotes questioning as a vital mechanism in player learning (Forrest, 2013; Light & Harvey, 2017). Ben reflected on their use of questioning to better understand decisions made by players, to inform their coaching approach:

That player just thought he saw something and the question for me “What did you see?” I think it’s just as powerful as me telling him what I saw because I know now I need to change something around his kind of perception around what he’s looking for and knowing where to look and how to look.

Despite questioning being an integral part of the implementation of a GBA (Harvey & Light, 2015), coaches frequently deployed the use of explicit instruction in conjunction with their use of games. Lucy reflects on the use of explicit instruction to ensure clarity:

In an actual session direct instruction I think is really helpful. Personally, for myself, and our group, I feel like it helps. We need to be clear what we want from them … we have to sometimes just tell them what we’re trying to do.

Interestingly, the criteria for selection of explicit instruction or questioning by coaches seemed to be individualised with some using explicit instruction due to procedural concerns, such as time limitations on pitch with their players. Others, for example Donal, reflecting on the needs of the athlete, described how the level of athlete affects coaching style:

I’m definitely trying to ask better questions, as opposed to give more (direct) feedback because (current team) have high game IQ. They understand the game a lot, whereas my time with (previous team) it was a very novice team so I gave a lot of (direct) feedback… Now I’m trying to ask more questions or ask better questions because of the group I’m working with. They’re at a level where their understanding of the game is quite high … they’ve also played to a higher level so they just have that experience.

This approach to the provision of feedback is supported by the work of Kirschner, Sweller, and Clark (2006) who note that minimally guided instruction is likely to be ineffective when dealing with novice or intermediate learners.

Desired player experience

The final overarching theme, desired player experience, related to the theories that the coaches used as a means of underpinning their pedagogic choices and how these theories impacted the desired learning experience of the player. One of these approaches was the leveraging of game form training to enhance shared knowledge and shared mental models (SMMs) to enhance coordinated behaviour (Ashford, Taylor, Payne, Waldouck, & Collins, 2023; Richards, Collins, & Mascarenhas, 2017). SMMs have been defined as the “knowledge structure(s) held by each member of a team that enables them to form accurate explanations and expectations … coordinate their actions and adapt their behaviour to demands of the task and other team members” (Cannon-Bowers, Salas, & Converse, 1993, p. 228). For example Tom explains how game form training was integrated with off-field learning design, to facilitate a shared understanding:

We wanted to design a game that would help them make good decisions and be able to execute effective kicks … help them work on their kick strategy. So that started with … looking at clips with them around situations and games and how might you create a kick here? You know, what sort of kick might you use?…Video review with them and then go into our (training) game.

Game form practice was used to develop SMMs; for example, games were used with players to provide opportunities to play in different positions with the aim being to better effect a shared understanding:

We’ll allow everybody a go at it. So when we’re doing the build-up stuff … , he’s in that position (atypical team role) for a little period of time to get an understanding of what we’re doing as a team. So all the players within that game design, were able to get some kind of exposure to things that they will work on regularly, things that they won’t work on that regularly. But everybody’s got a shared understanding.

A key concern for many participants was managing Cognitive Load in training sessions (Sweller, 1988) as game form practice could lead to high extrinsic load that could impact learning. This meant that coaches were concerned with how their learning design and coaching approach might add to extrinsic load:

If I can see that they don’t understand the game … it’s too complicated, then for me, that’s a worry because I’ve overcomplicated the game design, they’re too worried about understanding how this game design works, rather than just playing. [Lucy]

In addition to the need to manage pedagogic strategies, CLT would also suggest that learners can only retain new information effectively provided it has been processed sufficiently to pass into long-term memory. Thus, a key concern was the extent to which learning was truly embedded. This theory was even used by coaches with a more ecological lens. For example, Colin using testing to check for embedding:

I have my particular biases of particular models or skill acquisition theory. I felt very strongly around perception-action coupling and ensuring that you’ve got the right context in front of you. So, using that theory, then using theories of learning: frequency of recall, the forgetting curve, and then test. So the (testing event) would allow me to understand what’s gone in, what you learn and what’s falling off.

Whilst not fully reflective of concepts emanating from cognitive psychology, this suggests an appreciation of learning as a process rather than an event (Kirschner, Sweller, & Clark, 2006). Another key concept of CLT is the notion of expert-novice differences and the need to consider different approaches depending on the level of the learner (Sweller, 1988). Therefore, the relative challenge point of the learner was a key consideration for coaches:

[when] a player comes [to academy], technically, tactically not at the level of his peers, the group that he comes into. He might be that early physical maturer, very quick, let’s say. By putting them into games without the toolbox of skills and options of an understanding of decisions to be able to cope with the demands of a game against those particular peers, could be detrimental. [Callum]

In this sense, coaches were aware of the need to manage the challenge levels experienced in “game form” training (Guadagnoli & Lee, 2004). Coaches reflected on the need to challenge athletes and awareness of this mediating their pedagogic decisions. Here the influence and importance of challenge point in game design is explained in the context of the development of knowledge:

If there’s errors at the start of practice, but I’m happy with the way in which the game is designed, the errors that showed to me that the boys are thinking and trying to overcome these problems, then I think that’s sort of like the perfect place to be [Lucy]

The coaches also reflected on what they perceived an appropriate number of mistakes in a training activity to be, a matter that is of debate in the wider literature (Wilson, Shenhav, Straccia, & Cohen, 2019). Coaches reflected on this, explicitly referencing the challenge point framework (Guadagnoli & Lee, 2004):

We’re trying to get to that optimal challenge point where there is a level of error or else the challenge isn’t right. You know you’re looking for a 70% success rate and 30% failure [Stuart]

I’d always hope the players achieve five out of 10. So it’s kind of 50/50, there’s some level of success, and there’s a good level of challenge there as well but if that was one of the 10, like, it just becomes very disheartening from the players point of view and very frustrating. It’s trying to get the balance [Donal]

Notably absent from the reflections of the coaches were a series of highly evidenced concepts from the motor learning and broader literature such as desirable difficulties and contextual interference (Bjork, 2018). Thus, whilst coaches appreciated the importance of appropriate task difficulty, embedding variability and contextual interference did not appear to be prominent features of their practice.

Knowledge sources

The second research question aimed to examine the sources of knowledge used by coaches with generated themes including mediated, unmediated, and internal sources. Seven sources of knowledge were identified as being impactful among this cohort of participants (see table 4). Although atypical for RTA research, the number of coaches identifying a particular knowledge source as being impactful for their practice is added to help the reader make sense of the relative weighting.

Table 4 Knowledge Sources

Theme	Sub-theme	Raw Data Exemplar	Number of coaches identifying as an impactful knowledge source
Unmediated sources	Observing other coaches	I spent a lot of time observing really, coaching schools’ day to day where I’d coach but then I’ll be as an observer on the academy and watching what the kids responded really well to and what the kids do. (Ben)	14
	Social media	I’m always looking at, yes, I’m always looking online. For other sports from warm up drills to skills drills, that kind of stuff. Little snippets. [Sport] for [skill], anything that can give me an edge, I’ll always chase down and I’m always seeking to learn. (John)	9
Internal sources	Reflection on personal practice	I remember my first coaching job at (Football club’s) community department doing after-school clubs. Before I did an after-school club, they put me in a camp and I’d never worked with kids before and I’ve never coached. And then I had to. So basically my first experience of coaching was taking 15 kids on holiday camp for like five hours a day. So I was just like, “Oh, my God, what the hell do I do with these children?” And I remember learning in like, day one, that kids love games. So… all I was thinking, how am I going to entertain these kids, I used to come up with these games. (Lucy)	10
Mediated sources	Mediated coach education programmes	It has but probably not in a traditional sense, you know? … The B license and the A license generally pushed phases of play… I suppose the game to a degree, but in a very directive and specific manner, it wasn’t free play in the games that we’re kind of discussing at the minute or it wasn’t the sort of ebb and flow if you want, it was probably more directive to “you are trying to get this outcome explicitly”. (Steven)	8
	Academic qualification	I’ve done my masters in sports psychology that’s helped with my coaching and my coach education, having that kind of background too. (Donal)	6
	Mentoring	It came from a conversation with (coach) and so myself and (coach) would have phone conversations that would go on forever and stuff like that and he just started talking about you know using different balls and then throughout the course of the conversation anyway we came up with the idea… then you just start building on it, you start trying that and see how it goes. (Brendan)	8
	Coach Developer influence	I invited him to come in and observe some of our sessions way back in 2007 and he just observed some of our sessions, and just sort of asked a few questions about, you know, how valid is your environment? And by that he meant, you know, is what the players seeing, feeling, perceiving (is it) replicating (the game)? Ever since then, I’ve had a massive curiosity. (Colin)	6

Unmediated sources

The most widely referenced and impactful source of knowledge was coaches’ observation of other coaches. This supports previous work which suggests that coaches preferred informal learning sources (Stoszkowski & Collins, 2016):

There would have been like a genuine curiosity, just chatting to other coaches and watching them coach and how they use games … the biggest influence would have been watching other coaches use them and use them well [Scott]

The observational learning described here is associated with the acquisition of attitudes and values and can lead to more creative coaching (Bandura, 2008). These observations were not limited to within their own sport and indeed many coaches displayed a hunger to learn from other sports:

I can go and watch the rugby guys coach and rob an invasion game and then change it to (coaches own sport)… So yeah, I mean, I would use rugby games, basketball games. Definitely football, I would use all of those things in order to stimulate thoughts and to draw pictures and then copy them [Zara]

This example highlights observation as an inductive process of reflection and action (Marsick & Volpe, 1999; Schön, 1983). Further to the point of informal, unmediated learning, nine coaches pointed to social media as a source of knowledge.

I was forever ripping stuff off Twitter and Instagram … just started as you do going into the deep dark hole and you find some awesome people and some awesome stuff [Tom]

While informal learning was a preferred source of learning, research suggests that often the preferred method of learning is not always the optimal method (Kirschner & van Merriënboer, 2013) and as such unmediated learning may not be without its limitations: “there’s a lot of people designing and sharing a lot of stuff with probably not a lot of context and understanding behind it” [Callum]. This poses two issues; first, with such an abundance of content available online, coaches may simply consume bad coaching practice on social media and look to replicate it (Stoszkowski, Littrell, & Collins, 2022). Second, where coaches are exposed to higher-quality knowledge sources, this might be used sub optimally, focusing on the “how?” or procedural knowledge rather than the declarative and why this might be appropriate in a given setting (Stoszkowski & Collins, 2016).

Internal sources

Internal learning was also discussed with coaches’ reflection on personal practice identified as a source of knowledge. This is in line with the idea that reflective practice is a critical aspect of coach learning (Knowles, Borrie, & Telfer, 2005). Tom discussed how his previous career in teaching was as an influence on the use of games within his coaching:

I suppose coming from a teaching background. I found out pretty quickly with teenage boys that if they didn’t get a game, they lost interest. And so I guess through years of teaching, basketball, badminton, whatever it was, I got the most focus and attention and effort and best sessions, where there was some competitiveness in session. So I guess, that just evolved in my teaching, and then into my coaching that I could see, just lifted everything

A coach’s experience of playing their sport has been shown to have an impact on their coaching (Cushion, Armour, & Jones, 2003; Lemyre, Trudel, & Durand-Bush, 2007). Reflecting this, Sophie explained how their experiences as a player, specifically their enjoyment of games within a session, helped shape their coaching approach and use of games: “all the sessions that I enjoyed had competition in them, they all had games in them, or game related activities”. In contrast, Zara reflects on their experience and specifically the lack of games within a session as an influence on their subsequent coaching approach:

So I had a couple of (international) coaches who came over and worked with me with as an athlete, who were only drill based coaches, you know, (repeating isolated technical skill) a million times and running from cone to cone and stuff like that and for me, as an individual, it completely turned me off

Mediated sources

Formal sources of learning were also strongly influential on coaches’ use of games in their coaching practice. Mediated coach education programmes were referenced by the majority of coaches as a source of knowledge as Callum explains:

The (advanced coaching certificate) has taught a variety of ways to manipulate games and informed coaches of different ways to adapt the area size for particular needs or adapt the constraints or the duration, or the time or the shape or the design of the session, the design of the game

This, however, differed between individuals and sports. Some participants perceived a total lack of impact of mediated coach education programmes on their practice and understanding. This lack of impact was supported in member reflections with a view that “game form” training, whilst being strongly promoted, was too simple to influence practice: “coach education does not exist on a high level of quality within our sport, so it does not influence the high performance area” [Peter]. It was instead the more social and unmediated elements (Stoszkowski & Collins, 2014) that were perceived as more influential: “The only thing I did enjoy about them was the interaction with the people there, I actually learned more from the people doing it rather than the course itself” [Donal].

Six coaches described how their academic qualifications influenced their coaching approach: “What’s influenced my understanding of games is my PhD … and I’ve then tried to apply that to my coaching in terms of game design” [Lucy]. This finding was interesting in that it highlights the potential for coaches to build a more evidence informed approach to their practice and challenge the oversimplifications that the coaches themselves reflected upon.

Mentoring also proved to be an important source of knowledge (Leeder & Sawiuk, 2021). Coaches described the influence of mentors as more experienced others, mediating what would otherwise be unmediated or internal learning processes. This reflects the idea that rather than being a formal education process where learning is imparted on learners, mentoring allows knowledge to be co-constructed and requires expert practice to be effective (Olsson, Cruickshank, & Collins, 2017). The mentoring process impacted how participants viewed learning design:

I invited [mentor] to come in and observe some of our sessions way back in 2007 and he just observed some of our sessions, and asked a few questions about how valid is your environment? And by that he meant, you know, is what the players are seeing, feeling, perceiving, replicating (the game)? Ever since then, I’ve had a massive curiosity [Colin]

In addition to mentors, coaches also drew on the different role of coach developers (Taylor & Nash, 2023) as practitioners who work across the spectrum of mediated and unmediated deliberately shaping the learning experiences of coaches (Brackley, Barris, Tor, & Farrow, 2020). Zara illustrates this influence:

We [sport] have some really good coach developers who, in my opinion, allow us to coach in this way, but get us to think more about the constraints and thoughts and how we might create and innovate around those games in order to get better outcomes

Coach developers have the potential to strongly influence a coach’s pedagogic choices and understanding (Stodter & Cushion, 2019). Although these experiences were consistently perceived as impactful, it is important to recognise that our data does not suggest that coach mentoring or development is the way that coaches should be developed, or inherently superior to other approaches.

Conclusion

This research attempted to construct an understanding of the pedagogic knowledge informing the game form training activities of high-level team sport coaches and the sources of this knowledge by moving beyond a procedural approach, understanding what coaches did and instead aimed to understand why they did what they did (Crandall, Klein, & Hoffman, 2006). High-level team sport coaches, reflected on a range of different theoretical positions as influencing their practice some of which might appear incongruent, yet seemed to offer significant pragmatic value (Raab & Araújo, 2019; Stodter & Cushion, 2017). These findings seem relatively coherent with a similarly high-level sample of swim coaches who used a variety of approaches to practice without holding a single theoretical underpinning (Brackley, Barris, Tor, & Farrow, 2020). In addition, none of the coach participants reflected on a single approach to their “game form” practice (Harvey, Cushion, & Massa-Gonzalez, 2010; Light & Harvey, 2017; Light & Robert, 2010) or the perceived need to marry learning design and coaching style (Light, Harvey, & Mouchet, 2014). Thus, the findings seem to suggest that in a sample of the highest level of team sport coaches, no coach utilised a single pedagogic or theoretical orientation (Rink, 2001). Instead “game form” learning design was used in a highly individual and contextual manner (Collins, Taylor, Ashford, & Collins, 2022). This was evidenced in the number of coaches who adopted a range of different practices, apparently conflicting with each other. An example being a participant discussing the importance of maintaining the coupling of perception and action, whilst managing cognitive load. Importantly, we are not suggesting that coaches should be using multiple theoretical stances, or would have been better using a single approach in their practice. We simply provide evidence that this set of coaches, all of whom could justifiably be classified as having a high level of expertise, used multiple models to understand their practice (R. Taylor, Taylor, Ashford, & Collins, 2023). Indeed, the critical reader might suggest that the variety of knowledge sources that the coaches were exposed to led to the mixed model views that they espoused. Rather than this being emblematic of poor coaching we suggest that this could instead be coaches finding pragmatic value in mixing theoretical positions based on a variety of knowledge sources, something more recently seen in the literature (Ramos, Coutinho, Davids, & Mesquita, 2021). If anything, coaches were critical of coach education that promoted the universal use of game form activity design as being too simple. Consequently, coaches made critical comments regarding the naïve and prescriptive promotion of GBA, in a manner contrasting with GBA literature (Pill, 2020). At risk of following a well-trodden path criticising formal coach education, here our critique is more subtle, noting that it is not possible (or appropriate) for coach education to provide knowledge of every pedagogic theory. Instead suggesting the need for mediated coach education to provide a platform for the growth of expertise (Nash & Sproule, 2012). Thus, we suggest the need for coach education to move beyond a procedural focus on the use of game form training to a more nuanced and deeper pedagogic understanding that does not rely on a single theoretical orientation.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data is not available on request due to participant anonymity concerns.