Due to the experiences players gather through deliberate practice and competitive match play, they develop a task-specific knowledge base which creates an opportunity to interpret events encountered in sporting situations with reference to those previously experienced (Williams, 2000). This knowledge base results in domain-specific adaptations in long term working memory, facilitating sophisticated strategic or tactical problem representations (McPherson, 1999, 2000; McPherson and Kernodle, 2007). In team sports, skilled players have superior strategic awareness compared to novices (i.e., soccer, Williams, 2000), which impacts player positional location (González Víllora et al., 2013) and the ability to predict future movements of others in the game (Brault et al., 2010, 2012).

A player’s strategic awareness is linked to the visual information they acquire from within the environment. For example, through watching filmed based scenarios in soccer, when anticipating pass direction, experienced players (EP) generally look more peripherally in the display, fixating on other players or areas of space that could be exploited and demonstrate a higher search rate (more fixations of shorter duration) compared to novice players (NP) who fixate the player with the ball (e.g., Helsen and Pauwels, 1992, 1993; Williams et al., 1994, 1999; Helsen and Starkes, 1999; Vaeyens et al., 2007). These adaptations in visual search behaviour result in the EP being more accurate in anticipating scenario outcome compared to NP in video-based paradigms.

Physically executing a task, results in a different visual search behaviour compared to when little/no motor component is required (e.g., watching a video Land and Hayhoe, 2001; Dicks et al., 2010). This adaptation of visual search behaviour can be attributed to the presence of two-visual systems; a dorsal and ventral stream (Goodale and Milner, 1992; van der Kamp et al., 2008). The ventral stream creates an allocentric representation of the world where the initial constraints of the action are defined. In contrast, the dorsal stream provides an egocentric perspective mainly associated with identifying contrast and movement related to the execution of the motor action. The separation of these pathways is debated (see Mann et al., 2021) but instigated the use of more representative experimental designs. When investigating visual search behaviour in sport, and to avoid decoupling of perception of information from the motor action (c.f. Mann et al., 2010) it is important to use representative designs (c.f. Araujo et al., 2007). If participants engage with physically executing the task(s) it helps to ensure that information is derived from both ventral and dorsal streams. This is of particular importance in self-paced tasks with varying environmental constraints where differences in expertise are examined. Expertise differences are evident in superior motor actions suggesting an enhanced ability to rely on information from the dorsal stream. This is reflected in findings that differences in performance and visual search behaviour between experts and novices disappear when simulated or uncoupled actions are used in comparison to in-situ tests (van der Kamp et al., 2008; Mann et al., 2021). In other words, expertise differences might be less apparent when ventral stream information is derived then dorsal stream information. As most studies examine expertise differences and visual search behaviour in anticipation skills (Connor et al., 2018), or use a simulation of the environment (Hüttermann et al., 2019), there is the need to examine expertise differences in visual search behaviour in the execution of discrete skills in dynamic in-situ environments. However, currently there are very few studies have utilised these “live” game scenarios to understand visual search behaviour in discrete skill tasks (i.e., throwing a pass) in a semi-predictable sport environment (for exceptions, see Martell and Vickers, 2004; van Maarseveen et al., 2018). In rugby, an attacking player with the ball is constrained with only being able to pass the ball backward (as opposed to most other team sports where the ball can be passed in any direction) to a teammate. Due to this requirement, it is likely that the attacking player’s movement patterns and visual search behaviour will be constrained by the movement patterns and positioning of teammates when passing the ball, as opposed to other team sports (e.g., soccer and basketball). To fully understand how expertise influences visual search behaviour in this dynamic relationship between players, the current study collected visual search data using a mobile eye tracker in “live” attacking rugby scenarios. Within this setting, both experienced and novice rugby players were placed in attacking scenarios and were required to score a try against a defender. In this 2 vs. 1 game play, we created four scenarios of varying complexity, where players had to decide to take “the defender on” or “pass the ball.”

Based on emerging evidence that the relationship between expertise and visual search behaviour in-situ is different from simulated environments (Mann et al., 2021) we hypothesised that experienced rugby players would (1) demonstrate superior performance compared to NP (2) direct visual search behaviour for more time, and more often, toward areas related to the successful executing of the action (i.e., throwing a pass) compared to NP.

The current study investigated the adaptations which occur as a function of expertise in rugby player’s visual search behaviour when tasked with completing attacking scenarios. Differences in visual search behaviour between experienced and NP were observed in how long and how often players fixated immediately ahead of the support player in preparation of completing a pass. Concurrently, the mechanism of acquiring visual information in preparation and completion of the action (e.g., passing or taking on the defender) were significantly different. In line with our hypothesis, it was found that EP visually attended more often and for longer to areas related to opportunities for action (e.g., gaps on the side of the defender) where novices visually attended significantly longer to the defender, an area that could reflect a more allocentric perspective in motor response planning. These findings were accompanied by a superior performance of the EP compared to the NP.

By using “live” game scenarios we were able to create a representative design where rugby players completed an attacking phase of play. This allowed for the examination of expertise differences in visual search behaviour in both the planning and execution of the motor action. Previous studies (van der Kamp et al., 2008; Dicks et al., 2010; Mann et al., 2021) highlighted that the inclusion of the motor action influences the balance in deriving information from the dorsal and ventral visual system and therefore adopted visual search behaviour. As expertise differences reflect superior motor performance it is expected that the adopted visual search behaviour is also different. Indeed, our results indicate that superior performance of the EP was also evident in their visual search behaviour. This contradicts findings from Connor et al. (2018) where visual search behaviour of elite and novice rugby players was examined in anticipating evasion manoeuvres of attacking rugby players. Connor et al. (2018) did not identify significant difference in various visual search behaviour variables (i.e., viewing time at different AOIs) that could support the significant improvement in anticipation skills in the elite players. However, Connor et al. (2018) did not include a motor component in the anticipation of the evasion manoeuvres and therefore predominantly examined ventral stream related visual search behaviour, excluding dorsal stream influences. Particularly the ability to derive visual information through the dorsal stream underpins superior performance of experts.

Within scenarios which required the participants to pass the ball (Spin Pass and Switch scenarios), a number of simultaneous processes occur which inform processes made regarding stimulus-related variables (e.g., travelling velocity of the person to pass to) and response-related variables (e.g., distance required to pass the object, Lim, 2015), all of which are informed based upon an egocentric perspective where the acquired visual information reflects the position and motion of the body with respect to the environment (Davids et al., 2004). The distance the ball was required to travel (particularly in the Spin Pass scenario) and the speed the support player was travelling meant that passing the ball at the support player would result in the ball arriving behind them and the ball not being caught; alternatively, the support player would be required to slow down/stop running to catch the ball, increasing the likelihood of the defender running across to make a successful tackle. Instead, successful interceptive throwing/passing tasks result from fixating at the interception point (Lim, 2015), which in the current study was immediately ahead of the support player. Indeed, despite the importance of the participant acquiring positional information of the support player (i.e., ensuring the support player is not positioned in advance of your position) the length and time EP spent fixating the support player only accounted for a small proportion of the trial (Table 1). Instead, EP fixated longer and more frequently immediately ahead of the support player and acquired information of the support player from their periphery (e.g., Piras et al., 2014a,b) in comparison to the NP. By not directly fixating the support player (turning the head left to centralise the fovea on the support player), this also maximised the area in the visual scene which could still be perceived from the opposite side of the periphery, which contained other key information sources (e.g., location of the oncoming defender). In the current study, the NP did not adopt this strategy. The NP spent an equally small proportion of the trial fixating on the support player and immediately ahead of the support player (Table 1). This finding supports that the EP initiated the pass earlier in the phase of play in comparison to the novices. A subsidiary analysis of all the trials completed in the Spin Pass scenario, when collapsed across group, demonstrated the importance of fixating the interception point and not the support player when passing. A successful pass and try resulted from fixating significantly longer at the area ahead of the support player [pass t(41) = 2.65, p = 0.011, try t(45) = 2.630, p = 0.012], but this was not apparent when looking directly at the support player (p > 0.05).

Expertise specific differences in visual search behaviour where also apparent in examining the environment in guiding motor actions. The EP exploited gaps to the right of the defender significantly more than NP in the more complex Switch and Dummy Switch scenarios. Similarly, there were clear significant differences in the length and frequency of attending to the defender. NP visually attended significantly more time and more often to the defender in general and specifically in Spin Pass and Take-Player-On scenarios. Both these findings reflect how expertise influences the extraction of visual information with constant environmental constraints.

In both Switch and Dummy Switch scenarios, the support player was required to arc their run (from the left of the participant) behind the participant, eventually running into gap right (Figure 1C). To maximise the opportunity to score a try in these scenarios, the participant was required to create space for the support player to run into. Insufficient space between defender and attack/support player allowed the defender to “tackle” the ball carrier. During Dummy Switch and Switch scenarios only, EP fixated more frequently in gap right compared to NP. With no associated increase in time spent looking at gap right, it appears that EP were “checking” that sufficient space was being created for the support player to run into, something that NP did not do.

Concurrently to differences in exploiting the space around the defender it is equally important to acquire visual information regarding the movement and intentions of the defender. The defender was consistently and significantly more frequently (NP 23.5 ± 6.5 vs. EP 16.5 ± 7.7%) and for longer (NP 34.8 ± 17.4 vs. EP 18.2 ± 12.2%) visually attended to by the NP than the EP (see Figures 3A,C). The importance of directing visual search behaviour at the defender is supported by previous research which has reported that during 1-on-1 situations in rugby, experienced defending players focus on the pelvic region of the opponent (Brault et al., 2010), as this provides the most accurate cues relating to the opponent’s future running direction (Brault et al., 2012). This contrasts our findings and highlight the importance of task constraints on visual search behaviour (c.f. Rienhoff et al., 2016 for an overview of the influence of constraints on visual search behaviour). In our attacking scenarios, in contrast to defending scenarios in Brault et al. (2010)), visually attending to the defender was less important to our EP than our NP, specifically in Take-Player-On and Spin Pass scenarios. An explanation of this can come from the influence of expertise on dorsal and ventral visual systems. The increase in visually attending to the defender suggests that the NP use the defender to plan and guide their motor action based on defender movement and anticipation of defender movement (i.e., visual anticipation via the ventral stream). In contrast, the NP attend less to the defender but direct attention more to areas that guide their motor actions (e.g., area ahead of support player) reflecting a greater reliance on the dorsal stream. These marked differences in visual search behaviour as a function of expertise in a “live” scenario provides further support for the use of representative designs as these differences disappear in an anticipation video-based paradigm (Connor et al., 2018).

When completing attacking scenarios in rugby, differences in visual search behaviour are apparent as a function of expertise. Differences in visual search behaviour between experienced and NP related to the more refined awareness and functional understanding of the strategic role of space immediately ahead of the support player by experts. This space is exploited to guide motor action from an egocentric perspective in contrast to the novices who seem to rely more on information from defender movement and actions to build an allocentric perspective to predict their future movements.

Future research may wish to consider how to expedite player learning through cueing visual attention to key strategic areas during attacking play in rugby. Whilst approaches to and effectiveness off perceptual-cognitive skills training are varied (see Klostermann et al., 2015) and initial understanding of the adopted gaze behaviour of experts can form the foundations of a successful intervention. Such approaches are recommended to be designed with a representative design reflected in the inclusion of relevant motor actions and an information rich environment with a large variety of perceptual stimuli.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Faculty of Science and Engineering Ethics Committee at Anglia Ruskin University. The patients/participants provided their written informed consent to participate in this study.

KP, IB, and MT contributed to conception and design of the study and wrote sections. MT and JL performed the data collection. KP, JL, PR, and MT contributed to data and statistical analysis and interpretation. MT wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.