Evaluating interaction mechanics in virtual reality gaming: from user studies to design recommendations

Abstract

By utilizing six-degrees-of-freedom (6DoF) interaction, virtual reality (VR) gaming often places greater demands on both the user’s motor skills and cognitive resources compared to gaming on non-immersive platforms. Satisfying VR interaction mechanics (IMs) should facilitate a state of flow by balancing challenge and enjoyment while minimizing negative VR-induced symptoms and effects (VRISE) that threaten user comfort and overall experience. In this paper, we present the results of three user studies involving a total of 90 participants (N = 30 per study), examining how different customizable parameters of commonly encountered IMs (slash, pick-and-place, shoot) affect both objective performance measures and subjective measures of player experience, workload, and VRISE. The studies were conducted in a laboratory environment utilizing a specialized testing platform. Our results indicate a general agreement between subjective measures of challenge, workload, and competence and measures of task performance. Observable variations in preferences among participants suggest there may not be a one-size-fits-all approach to implementing each of the tested mechanics. While participants most commonly preferred parameter configurations they deemed comparatively less demanding, for certain parameters a level of added challenge appeared to contribute toward a more fun experience. Despite a level of physical exertion experienced during the study session, participants reported a relatively low intensity of VRISE. Based on these findings, we propose a set of IM-specific recommendations for implementing the three mechanics investigated in this study. The proposed recommendations provide practical guidance for designing configurable VR IMs that balance player experience, performance, and comfort.

1 Introduction

Unlike non-immersive interfaces, which often rely on two-dimensional input modalities such as mouse, keyboard, or touch interaction, virtual reality (VR) systems position the user at the very center of the virtual experience. By adopting a first-person perspective and utilizing six-degrees-of-freedom (6DoF) motion tracking, VR applications facilitate a direct mapping of the user’s physical movements onto their avatar. This confluence of shared perspective and shared embodiment establishes an avatar-self merging (Müsseler et al., 2022), wherein the user gains a deep sense of body ownership and agency over their digital representation, driven by multisensory and sensorimotor contingencies that occur as they observe virtual body parts in the position of their own (Slater et al., 2010).

As explained by LaViola et al. (2017), methods that enable the user to accomplish a task via the user interface are referred to as interaction techniques. They involve software and hardware components responsible for translating the user’s input into a particular action within the system, as well as the resulting system response, presented to the user via output devices. While this term is applicable across different systems and use-cases, we observe its relevance in the context of VR gaming. As described by Fabricatore (2007), when a gamer interacts with the game interface, their actions are mediated by what the author refers to as the player-token. Whether embodied as an avatar or presented in a more abstract way, the player-token functions as a proxy mapping user controls onto in-game behavior, thereby defining the core gameplay and, consequently, the core mechanics of the game. In VR games, due to self-merging and immersive input and output modalities, the player-token usually provides a very straightforward translation between the gamer’s movements and the behavior of their virtual body. Because of this tight coupling, interaction techniques in VR games are closely intertwined with core game mechanics, which is why we choose to merge the terms and refer to them collectively as interaction mechanics (IMs).

We introduced the term interaction mechanics in our earlier work (Vlahovic et al., 2022), where we analyzed and classified different IMs depending on multiple factors, such as interaction fidelity, hand usage, tool-mediation, and target characteristics. We highlighted the importance of detailed and systematic exploration of task performance and user preferences in this context, particularly when considering that user experience and wellbeing depend on the developer’s ability to reconcile competing demands of VR game design. First, there is the balancing act that is central to gaming regardless of platform — ensuring that challenge and required skill are carefully calibrated to facilitate entering the flow state (Csikszentmihalyi, 1990). Second, the embodied quality of VR content necessitates further consideration of additional trade-offs. While physical exertion during active gameplay may serve as a contributor to enjoyment (Yoo et al., 2018; Pallavicini and Pepe, 2020), it needs to be carefully fine-tuned, as it can make users feel awkward and uncomfortable (Rogers et al., 2019). As described by Cobb et al. (1999), VR-induced symptoms and effects (VRISE) encompass a broad spectrum of ergonomic issues, cybersickness symptoms, and alterations in physiological state, postural stability, psychomotor control, and visual perception. Moreover, with prolonged use, active immersive gaming comes at a risk of neuropathy and repetitive injury (Belyea et al., 2025), posing an industry-wide challenge that threatens not only the livelihoods of VR game developers, but the safety of end users.

In the aforementioned work (Vlahovic et al., 2022), we proposed the INTERACT framework—a set of guidelines for implementing platforms for the evaluation of VR IMs. The framework highlights flexibility, repeatability, customization, automation, and avoidance of nuisance factors, while further underlining the importance of providing suitable interfaces and facilitating the export of multiple collected measures. INTERACT-compliant platforms — or testbeds — are intended to be used alongside subjective measures (e.g., those pertaining to player experience, workload, and VRISE) to provide developers with additional insights into user experiences, allowing for the precise mapping of IM configuration onto user preferences and performance.

Arguably, the idea of controlled experimental testbeds in immersive VR is not novel — in the academic context, their importance has been recognized since the early work by Poupyrev et al. (1997), who formalized a framework for systematically studying manipulation techniques, focusing on a subset of tasks including positioning, selecting, and orienting a virtual object, as well as tasks pertaining to inputting characters or numerical values. They defined a number of parameters pertaining to different tasks, as well as criteria for assessing performance, including error rate, completion time, accuracy, ease of use, ease of learning, and sense of presence. Similarly, Bowman et al. (1999) implemented a testbed for studying selection and manipulation tasks across multiple interaction techniques, measuring selection speed, selection errors, placement speed, and qualitative measures of user comfort. Bowman (2002) later defined a set of guidelines for designing selection and manipulation techniques. More recently, Weise et al. (2020) evaluated and compared VR interaction techniques in terms of performance, usability, workload, enjoyment, perceived naturalness, precision, and speed. Based on their results, the authors implemented a tool that provides developers with an overview and suggestions for choosing interaction techniques based on the specifics of the task that needs to be performed. However, the works discussed above primarily focus on selection and manipulation —– interactions central to 3D user interfaces —– but lack the specificity required for direct application in gaming contexts. In the absence of standards and guidelines regulating VR IM design that were developed with respect to established gaming practices, as well as the benefits and hazards of active gaming, the onus is on each VR game developer studio to determine the optimal IM settings based on their own judgment and testing practices.

To address this gap, in this paper we aim to investigate the effects of different configurations of three common game mechanics (slash, pick-and-place, and shoot) by utilizing an INTERACT-compliant platform. These particular IMs commonly present in VR titles, as reflected in the top three Meta Quest Store all-time best-selling VR titles at the time of this writing (Beat Saber, Job Simulator, and Blade and Sorcery: Nomad; Meta (2026)), each of which prominently features at least one of these mechanics. More broadly speaking, while the literature on the exact popularity/adoption of different VR IMs is limited (to our knowledge), they appear across different VR genres. For example, the shoot IM is central to the shooter genre (e.g., Half-Life: Alyx, Pavlov VR). The slash IM can be encountered across action and adventure genres (e.g., Blade and Sorcery: Nomad, Asgard’s Wrath), with one of its most popular games (Beat Saber) belonging to the rhythm/fitness genre. The pick-and-place IM, due to its versatility, appears across different games and genres, from being a supportive mechanics in action, adventure, and shooter games featuring other prominent IMs, to playing a more central role in many simulation and puzzle games (e.g., Job Simulator, I Expect You To Die). Notably, multiple aforementioned genres (shooter, action, adventure, simulator, puzzle) have previously been listed among the most prominent genres on the VR market, according to literature (Yoon et al., 2024; Foxman et al., 2020). Based on the IM classification in our previous work (Vlahovic et al., 2022), this particular choice of IMs includes both tool-mediated and non-mediated mechanics, covering the interactive range from immediate peripersonal to distant extrapersonal space. All of these interaction methods are based on familiar actions and movements (grasping, repositioning, and releasing objects; swinging a sword; pulling a trigger) and typically rely on widely used input mappings that are consistent across different VR applications.

Borrowing the term from robotics literature, by pick-and-place we refer to a mechanics in which a virtual object is grasped, transported through the scene, and placed at a target location, typically with a change in orientation. Unlike the remaining two interaction mechanics (IMs) explored in our work, pick-and-place is considered a direct or non-mediated IM, i.e., one that does not involve the use of a handheld virtual tool (LaViola et al., 2017; Vlahovic et al., 2022). Instead, it is realized solely through the simple virtual hand interaction technique (Hale and Stanney, 2014; Weise et al., 2020), enabling direct manipulation of available interactable objects within the arm’s reach. Comprised of basic object acquisition, manipulation, and release subtasks, pick-and-place can be considered a foundational VR game mechanics upon which tool-based IMs are built. When implemented in the described manner, pick-and-place is an isomorphic IM, i.e., it provides a direct, 1:1 mapping between physical motion in the real world and the resulting movement in the virtual environment (LaViola et al., 2017). However, pick-and-place can also be performed remotely, usually with a straight pointer reaching out from the controller. Once the pointer collides with an object, the user may choose to select it for further manipulation. Even though the direction of handled objects remains determined by controller movement within three-dimensional space, such remote implementations of pick-and-place are considered non-isomorphic. While non-isomorphic IMs increase the available interactable space, and may even be preferred by users, isomorphic handling of virtual objects resulted in better performance during a reaching and grasping task, in addition to producing higher levels of positive affect, flow, competence, and immersion, as reported by Kartick et al. (2020). Additionally, pick-and-place IMs can be enhanced by so-called snap zones (Saravas et al., 2020; Coutinho, 2022), which facilitate accurate object placement by attracting the manipulated object to the intended position upon its release from the user’s virtual hand.

Slash is a tool-mediated IM used to interact with target objects within the area immediately surrounding the player, but extending beyond their arm’s reach accounting for the length of the handheld tool (Vlahovic et al., 2022). As demonstrated in Katzourin et al. (2006), this IM typically involves a bladed weapon bound to its corresponding controller such that its position and orientation directly reflects the movement of the player’s arm. Target objects may range from passive inanimate objects to in-game characters seeking to harm the player’s avatar. While literature features different versions of VR applications centered around this IM (Coutu et al., 2020; Szpak et al., 2020; Cui and Mousas, 2021; Gamage et al., 2021; Yamamoto and Fukuchi, 2022; Tammy Lin et al., 2023; Islam et al., 2024; Wang et al., 2024; Ellahiyoun et al., 2025), we focus on selected works featuring two popular commercial VR games, Fruit Ninja VR and Beat Saber, as they resemble the IM implementation used in our work. In Fruit Ninja, the player uses a bladed weapon to slice fruit launched into the air. Islam et al. (2024) found that playing the VR version of Fruit Ninja resulted in higher workload scores compared to its mobile version. Also comparing the two, Coutu et al. (2020) found that immersion and arousal were higher in Fruit Ninja VR. However, participants criticised the VR version’s limited field of view, which required them to turn their heads in order to take in the entire interactive area of the game. In Beat Saber, the player is armed with two lightsabers used to slice floating cubes approaching in the rhythm of the background music. Tammy Lin et al. (2023) explored the effect of different playable angles on physiological and psychological outcomes during Beat Saber gameplay. The authors noted that higher cognitive demand led to higher enjoyment in some participants, but lower in others, depending on whether they found pleasure in the increased mental effort necessary to complete the task. Ellahiyoun et al. (2025) found that difficulty level and prior game experience affected participants’ objective task performance in Beat Saber. Szpak et al. (2020) demonstrated that neither short nor prolonged episodes of Beat Saber produced debilitating or sustained aftereffects, but cautioned that further research is needed on the impact of VR gaming on user well-being.

Shoot is an IM that revolves around two distinct but interconnected elements: a handheld tool (usually a type of gun) and the output exiting its muzzle. The gun’s direction, determined by the position of the controller, dictates the direction of the output, which is activated by pressing the trigger button, corresponding to the trigger of the virtual gun. Contingent on the context, the output can take many forms, such as a laser beam or a water jet; however, most often it is a type of projectile exiting the barrel with an initial velocity and continuing through the air along a parabolic trajectory. As with the slash IM, targets may range from static objects to armed characters. While such games may offer potential benefits for training purposes (Sudiarno et al., 2024), they are primarily designed for entertainment, whereas dedicated serious applications are typically used for training armed personnel and simulating emergency situations (De Armas et al., 2020; Awada et al., 2021; Liu et al., 2023; Harris et al., 2023; Fedasyuk and Vorobiov, 2023; Buga et al., 2024; de Sousa Abreu and Silva, 2025). As a result, there is significant variation in fidelity of presentation and projectile behavior in different implementations of the shoot IM, which necessitates divergence in IM evaluation approaches depending on the use case. For shooting simulators, it may be more beneficial to focus on ensuring that the simulation depicts convincing situations, provides a realistic shooting experience, and collects measures that correspond to real-world marksmanship benchmarks. On the contrary, realism may not play a large role in participants’ enjoyment of FPS games (Daneels et al., 2018). In the context of VR games, it may therefore be more useful to focus on identifying an IM configuration that is both challenging and achievable, as well as entertaining, for a user without professional shooting training or experience. In their text-mining analysis on Meta Quest store reviews, Yoon et al. (2024) report that users frequently mention unique gameplay mechanics in the shooting genre, including time control and gesture-based interactions. The academic literature on VR FPS games emphasizes the role of input modalities, spanning both software implementations and hardware interfaces (Rachevsky et al., 2018; Heydn et al., 2019; Krompiec and Park, 2019; Luro and Sundstedt, 2019; Berna-Moya and Martinez-Plasencia, 2019; Monteiro et al., 2020; Kocur et al., 2020).

While there is a substantial amount of literature exploring the specificities of chosen IMs, a lot of these publications are performance-focused or positioned outside of the gaming context, and thus lack subjective measures central to the player experience. Moreover, many fail to include measures of workload and VRISE, despite comfort being identified as one of the key enablers of player enjoyment in VR games (Sweetser and Rogalewicz, 2020). Multiple studies utilize equipment that is not accessible to the average consumer or available across different research institutions. Others rely on commercial games as test material, which prevents researchers from systematically manipulating individual design parameters or accessing detailed interaction-level data due to games operating as black-box systems with game-specific, rather than IM-specific, measures of success. Additionally, these studies typically examine a single IM in isolation and explore only a limited subset of parameters, which constrains the scope of their findings and rarely results in actionable design recommendations.

The more practical objective of our research, however, is to utilize these empirical findings to inform the creation of IM design recommendations that could benefit researchers in the field as well as VR game development professionals.

The structure of this paper is as follows. Section 2 presents the methodologies of the conducted user studies, while results are presented in Section 3. Section 4 provides a discussion of our findings, presents our recommendations, and lists relevant limitations, as well as opportunities for future research. Finally, concluding remarks are summarized in Section 5.

2 Materials and methods

In this section, we provide the methodologies of three exploratory user studies — US1_SL, US2_PP, and US3_SH — each focused on a different IM (slash, pick-and-place, and shoot, respectively). The methodologies are described collectively, as all studies were conducted in the same laboratory setting, followed the same procedures and used the same test application, although they involved different IMs, tasks, and objective measures. All study-specific differences are clearly indicated where relevant.

2.1 Materials

As described in our previous work (Vlahovic et al., 2022), a specialized INTERACT-compliant platform¹ (Matokanović, 2022) was developed to facilitate the exploration of user experience with three types of bimanual VR interaction mechanics: pick-and-place as a non-mediated mechanic, slash as an example of mechanics involving tools as extensions of peripersonal space, and shoot as an example of projectile-based interaction mechanics. While the platform itself does not constitute a contribution of this paper, it was implemented specifically for the purposes of this research and is described here as it served as the primary experimental tool. The platform consists of a desktop user interface and a VR interface, developed using the Unity game engine with the Virtual Reality Interaction Framework (VRTK) package. The desktop interface allows the administrator to define the parameter space for multiple IM-specific parameters, which could then be saved for future use. These parameters are used to generate the virtual environment which is experienced through a head-mounted display, with its dynamic behavior randomized within administrator-defined boundaries. Performance measures collected during gameplay are stored in a local folder and organized based on identification provided by the administrator.

The application’s environmental design is constrained to a limited set of brightly coloured low-poly target objects set against a celestial nocturnal backdrop. Sitzmann et al. (2018) demonstrated that environments characterized by low saliency-map entropy (i.e., those containing a small number of highly dominant salient regions rather than many equally salient ones) enable users to more rapidly direct their attention to relevant regions of a VR scene. Thus, by choosing a dark, uniform environment contrasted with several visually salient target objects, but void of superfluous detail, our goal is to avoid introducing additional variables that may confound our results by diverting the attention of participants away from the tested IMs. This aesthetic is shared across three customizable scenes (one for each implemented IM type). Individual descriptions of each scene are provided in the subsections below with screenshots of each mechanics implementation presented in Figure 1. By testing each interaction mechanic separately in a controlled, visually simple environment, we aim to produce general findings that can later be used as a baseline for more complex contexts.

FIGURE 1

Tool-mediated mechanics tested in our studies, slash and shoot, are centered around handheld virtual tools. In such cases, handheld controllers or similar input devices can provide tactile feedback that further supports the illusion of gripping a virtual object. On the other hand, the non-mediated quality of pick-and-place complicates decisions pertaining to control modality. With pick-and-place, the player’s hands are supposed to be empty unless they are actively holding an object, and then become empty again as soon as the object is released. Because of this, it may be argued that this IM is better realized through vision-based hand tracking. However, as explained by LaViola et al. (2017), hand tracking may be perceived as more uncanny in the context of pick-and-place compared to controller-based approaches, as it lacks the tactile feedback that is expected to occur when virtual objects are touched, handled, or dropped. Moreover, multiple studies (Gusai et al., 2017; Caggianese et al., 2018; Hameed et al., 2021; Kangas et al., 2022) demonstrated that, compared to commercially available hand tracking interfaces, controllers perform more reliably and lead to better task performance. Kangas et al. (2022) proposed a well-received middle-ground solution that combines the physical quality and performance of controllers with the natural grasping gestures of hand-tracking solutions, enabled by the specific capabilities of the Valve Index controller, which support a more natural experience of grasping and releasing objects. To satisfy our requirements for high tracking precision, trigger-based input beneficial for shooting interactions, passive tactile feedback, vibrational haptic feedback in response to events in the game world, and convenience for the pick-and-place IM, we have chosen to use Valve Index controllers for all three studies, along with the HTC VIVE Pro system which is compatible with both Valve Index and the platform used for testing.

2.1.1 The slash mechanics scene

The customizable scene for the evaluation of slash mechanics, inspired by Fruit Ninja VR² and Beat Saber³, places the user on a platform surrounded by a designated number of cannons positioned either radially all around the user or along a smaller circular arc, as determined by the chosen configuration. Depending on defined parameter values, each cannon will sporadically expel a cuboid box — either straight up into the air, or at an angle — which serves as a target object to be destroyed by the user. Object destruction is performed by slashing the object using an elongated handheld weapon. If the user fails to make contact — or fails to use the adequate force whilst making contact — the object will eventually fall down, influenced by gravity, and shatter upon impact with the ground.

2.1.2 The pick-and-place mechanics scene

The customizable scene for the evaluation of pick-and-place IMs was designed as a three-dimensional puzzle game in the style of a solid dissection puzzle known as the Soma cube (Figure 2), the variations of which have already been used in user studies featuring immersive technology (Cho et al., 2023; Tian et al., 2023). However, the exact puzzle presentation and puzzle mechanics used in our platform more closely resemble the implementation of the commercial game called Cubism⁴. Upon entering the scene, the user is placed next to a desk with scattered puzzle pieces. Each puzzle piece is a polycube, which consists of individual cubic units joined together in a randomized spatial arrangement, reminiscent of tetrominoes in a Tetris game if they were converted into three dimensions. When properly assembled, all provided puzzle pieces form a singular large cube. Users are also provided with a solution cube, a large cube the size of a completed puzzle, divided into smaller cubic slots. Upon picking up a puzzle piece, a set of cubic slots corresponding to its shape will light up, indicating its solution space, i.e., the designated placement of that puzzle piece within the solution cube. Even though requiring each user to decide on the placement of each puzzle piece is presented as a core challenge in games such as Cubism, in the case of our experiment, clearly highlighting the correct solution space of each puzzle piece ensures that the cognitive effort necessary to solve the puzzle is eliminated, so that participants can be fully focused on the straightforward task of positioning and orienting virtual objects. Users are then expected to slide the piece into place, fitting it inside the solution cube, and repeat this step with all other puzzle pieces until the finalized cube is assembled inside of the solution cube. As the user makes progress on the puzzle, the task of fitting complex polycubic shapes into place increases in complexity, as their efforts are hindered by existing puzzle pieces blocking the placement. Depending on the selected configuration, the task can be made easier by providing a small buffer space between the assembled puzzle pieces. Beyond its similarity to certain commercial puzzle mechanics, the Soma cube approach was selected because it provides a simple and well-defined assembly task whose geometric elements allow for precise parameter manipulation and easily quantifiable measures of success. At the same time, despite its apparent simplicity, this type of 3D puzzle requires precise positioning and orientation of individual pieces through complex six-degree-of-freedom manipulation. As such, it serves as an abstraction of object assembly tasks encountered across a variety of VR applications, ranging from games to industrial assembly simulators. Although the Soma cube does not reproduce the full complexity of such tasks, it aims to capture fundamental object manipulation behaviours that may be encountered across these contexts.

FIGURE 2

2.1.3 The shoot mechanics scene

Inspired by games such as Serious Sam VR: The Last Hope⁵ and Space Pirate Trainer⁶, the customizable scene for the evaluation of shoot mechanics places the user on a platform with a view of floating (stationary) or flying (moving) targets. Visually contrasting the night sky, these targets are sporadically spawned at random positions within the boundaries of a volume defined by a predetermined spawn angle and an acceptable range of altitudes and distances with respect to the player. The targets are destroyed immediately after being hit with a projectile expelled from the user’s handheld weapons. In addition to customizing target specifications and behaviour, the application provides a significant level of customization pertaining to ballistic properties of selected weapons, as well as enabling the use of optional visual aiming aids.

2.2 Procedure

The study procedure for all three studies was as follows. Upon entering the laboratory premises, each participant was asked to provide informed consent for the participation in the study. Participants were warned about the possibility of VRISE and encouraged to pause or terminate the experiment if they started to feel uncomfortable. Participants filled out an online pre-study questionnaire, providing their demographic information (age, sex, dominant hand) and reporting their level of experience with VR (i.e., whether they had previously used it and, if so, how often), as well as their sentiment toward VR technology (on a scale of “1 — extremely negative” to “5 — extremely positive”). Participants were then given instructions describing how to adjust and use VR hardware and entered VR to partake in a short tutorial session to grasp the controls and the task at hand. The tutorial scenario was set to default values of the application, some of which are presented in Table 1, while a more detailed overview of default values for all parameters is provided at the link⁷. These defaults were established internally, through a small pilot study, to provide a consistent starting point for all participants and were selected to offer a reasonable and achievable challenge for the sample demographic. The same pilot study informed the specific parameters and parameter values to be tested in the scope of presented research. Firstly, we identified interaction parameters that were expected to have a noticeable impact on user performance and experience.

The slash IM is typically performed in time-sensitive contexts that require users to continuously monitor their surroundings and react quickly to incoming targets. Therefore, the target spawn angle was selected as an experimental factor, as increasing the angular range over which targets can appear is expected to increase the levels of workload and challenge, and possibly, in-game performance. Furthermore, compared to the other interaction mechanics, slash involves more physically demanding arm movements. The force-to-destroy parameter was included to investigate the preferred intensity of slashing actions required to complete the task, thereby influencing the physical effort associated with the interaction. Finally, weapon length directly affects the user’s effective reach and the precision required to successfully hit targets.

Unlike the remaining two mechanics, tasks utilizing the pick-and-place IM are typically performed in less immediately time-sensitive contexts (e.g., puzzle and simulation games rather than action-oriented genres), but require finer movements and more precise manipulation that depends on the dimensional properties of virtual objects. Thus, we focused on the scale of the puzzle elements that need to fit together, as well as on the scale of the colliders, which directly determines the precision required for successful object placement. The remote grab parameter was considered as an additional facilitative element intended to reduce the physical effort associated with reaching for and retrieving objects during the assembly process.

Like the slash IM, the shoot IM is often performed in fast-paced situations, which motivates the selection of target spawn angle as a parameter to be explored. In real-life ranged weapon use, aiming aids such as laser sights are sometimes employed to help shooters anticipate the projected path of the weapon and align their aim accordingly, a feature that is also commonly emulated in shooter games. Research on intercepting moving virtual projectiles (Bosco et al., 2012) suggests that users rely on both feedback-based (servo-control) and predictive strategies when the target remains visible. However, when visual information becomes occluded, users rely primarily on predictive mechanisms based on prior experience, with less experienced participants exhibiting errors consistent with assuming a trajectory governed by normal gravity. While shooting (i.e., generating outgoing projectiles) differs from interception of incoming objects, we assume that similar perceptual and predictive mechanisms may still be relevant. In particular, manipulating the visibility of projectiles in flight, as well as the shoot force, which directly influences projectile trajectory, may yield insights into user expectations and preferences regarding ranged weapon behaviour in games.

After identifying the parameters to be explored, we determined the specific values for each tested condition. This process was straightforward for binary parameters—for example, the remote grab option was either enabled or disabled. For parameters defined on a continuous spectrum, and due to the lack of established guidelines or prior research that could be directly mapped to our scenarios, we relied on internal empirical testing to select values that appeared realistic, while still providing sufficient variation to meaningfully influence the interaction without rendering the tasks either trivial or unusable.

Following the tutorial session, participants in each study completed either three or four groups of scenarios (depending on the study/IM), each containing two to four separate scenarios, with all conditions in each study administered using a within-subjects design. All scenarios within a particular group were identical in terms of configuration (which was set to default values), except for a single parameter value that was manipulated for that group. The information regarding all scenarios is presented in Table 1.

Scenario groups were experienced in the order presented in the table to control for potential carryover effects between groups, ensuring that comparisons across groups were consistent. Within each scenario group, the individual scenarios were randomized for each participant to prevent order effects from influencing performance or responses on the manipulated parameter. Conceived as a series of short interactive tests (International Telecommunication Union, 2018), each individual scenario was set to last 90 s. The only exception were the scenarios in US2_PP (which focused on pick-and-place mechanics), as they would be terminated earlier in case the puzzle was completed prior to the expiration of the allotted time. Participants were not able to track the time whilst being immersed in a scenario.

Even though multiple items were adapted from existing questionnaires, to further simplify the cumbersome process of repeatedly completing the post-scenario questionnaire during the course of a study session, the same 5-point scale (ranging from “1 - not at all” to “5 - very much”) was used for collecting the evaluations of each feature’s intensity. Participants were also asked to rate the overall QoE on a scale from “1 - bad” to “5 - excellent”, and to report whether they were willing to continue playing in the presented conditions (yes/no). After completing each scenario group, participants were given a post-group questionnaire, in which they were asked to pick what they considered to be the best and the worst scenario from that group.

Objective task performance measures were also collected for each tested scenario; however, due to specific characteristics of each implemented IM, they differed between studies. Measures included in this paper are listed and described in Table 2.

2.3 Participants

Due to limited access to regular VR users, participants were generally inexperienced with VR, as seen in Table 3. Despite their lack of experience, participants in all three studies expressed a generally positive attitude toward VR technology, with this item receiving the average score of 4.20 (SD = 0.66) from participants in US1_SL, 4.10 (SD = 0.66) from participants in US2_PP, and 4.27 (SD = 0.74) from participants in US3_SH.

3 Results

In this section, we summarize the results of all three studies, analyzed using non-parametric statistical tests. For each study, the Friedman test was applied to compare the results within each scenario group, and post hoc Wilcoxon signed-rank tests with Bonferroni correction were used for pairwise comparisons. Due to the large amount of data collected, only partial results are reported in the main text. Percentages of users willing to continue playing in given conditions of each scenario, more detailed descriptive statistics, as well as Wilcoxon signed-rank tests with Bonferroni correction, are provided in the Supplementary Material.

3.1 Slash interaction mechanics (US1_SL)

Participants’ preferences for the best and worst scenario in each scenario group are listed in Figure 3. The results of Friedman tests, comparing the scenarios in each scenario group with respect to collected measures, are presented in Table 4.

FIGURE 3

With regard to target spawn angle, participants clearly preferred the 90_DEG scenario, which also scored the highest in terms of QoE (M = 4.4, SD = 0.62) compared to 180_DEG (M = 3.97, SD = 0.96) and 360_DEG (M = 3.27. SD = 1.26). On average, the highest scoring scenario was also found to be the least challenging, easiest to control, and least mentally and physically demanding. Furthermore, it was found to be more fun and made the participants feel more competent compared to the other two scenarios. While neither scenario seemed to cause significant VRISE, 90_DEG still scored the lowest compared to the other two. On the opposite end of the scale, the 360_DEG scenario consistently scored the worst for all tested items. While 97% of participants were willing to continue the 90_DEG scenario, and 90% of participants were willing to continue the 180_DEG scenario, this percentage dropped to only 47% for 360_DEG. Accuracy also suffered with an increase in target spawn angle, dropping from the average of 68.50% and 56.42% for 90_DEG and 180_DEG respectively, to 33.05% for 360_DEG. Differences in accuracy between all three scenarios were found to be statistically significant.

In terms of force required to destroy a target object, 0N and 2N scenarios received relatively similar scores, both in terms of QoE — with an average score of 4.2 (SD = 0.85) for 0N and 4.3 (SD = 0.75) for 2N — and participant preferences. Even though the 0N scenario was found less physically demanding, as well as significantly less challenging and easier to control, the 2N scenario was considered more fun, which may have influenced a higher number of participants (93%) to report their willingness to continue this scenario, as opposed to 0N (80%). For the 6N scenario, which received the lowest QoE score in this scenario group (M = 3.17, SD = 1.11), participants reported experiencing significantly less fun and feeling significantly less competent compared to the other two scenarios, with only 47% of participants willing to continue playing with this configuration. This scenario was also found to be significantly more challenging, physically demanding, and difficult to control, in addition to causing significantly more pain and muscle fatigue. Task performance also differed significantly between scenarios. While participants attempted to slash the target objects with a significantly higher average force (M = 5.05N, SD = 0.76N) compared to 0N (M = 2.91N, SD = 0.84N) and 2N (M = 3.53, SD = 0.76) scenarios, their performance suffered with, on average, only 34.29% of objects successfully destroyed, compared to 81.09% and 71.36% for 0N and 2N, respectively.

The least significant results were obtained for the weapon length scenario group. Even though a higher number of participants seemed to prefer the 1U scenario over 1_3U, both scenarios received the average QoE score of 4.23, with 0_7U trailing closely behind (M = 4.1, SD = 0.76). Interestingly, a slightly higher percentage (93%) of participants was willing to continue the 1_3U scenario, compared to the other two (87% for both). Even though differences in scores between the scenarios were fairly small for this scenario group compared to other scenario groups, the shortest weapon length was reported as posing the highest challenge and presenting the most physically demanding task, also slightly affecting participants’ perceived competence. The issue was somewhat reflected in the accuracy scores, as average accuracy of 0_7U (M = 67.63%, SD = 8.18%) was slightly lower compared to 1U (M = 71.31%, SD = 10.87%) and 1_3U (M = 70.85%, SD = 8.92%) scenarios.

3.2 Pick-and-place interaction mechanics (US2_PP)

Participants’ preferences for the best and worst scenario in each scenario group are listed in Figure 4. The results of Friedman tests, comparing the scenarios in each scenario group with respect to collected measures, are presented in Table 5.

FIGURE 4

The first observed parameter was related to the scale of the three-dimensional puzzle. It is worth noting that puzzle pieces — when scaled to fit the solution cube of a size determined by this parameter — were comprised of individual cubes with an edge length of approximately 3 cm (0_1U), 10 cm (0_4U), and 20 cm (0_7U). Each puzzle piece was comprised of multiple individual cubes, but no more than three in each dimension (i.e., each piece was 1-3 cubes wide, 1-3 cubes high, and 1-3 cubes deep). Participants appeared to dislike the largest puzzle scale (0_7U), with 20 out of 30 participants choosing it as the worst option in the puzzle scale scenario group. Moreover, this scenario received the worst average QoE score (M = 4.00, SD = 0.83) out of the three scenarios. While slightly more participants rated the 0_1U scenario as the best scenario compared to 0_4U, it was considered the worst by a third of all participants in this study. Considering that 0_4U (M = 4.50, SD = 0.68) also received a higher QoE score compared to 0_1U (M = 4.23, SD = 0.86), the medium puzzle scenario appears to have garnered a more universally positive response. With both 0_1U and 0_7U scoring higher in terms of challenge and all measures of workload, participants felt most competent following the 0_4U scenario, while also experiencing slightly more fun. Participants’ subjective experience was reflected in collected objective metrics, as the 0_4U scenario was successfully completed by 80% of participants, significantly exceeding the success rate of both 0_1U (53%) and 0_7U (27%). However, despite fairly low success rates, the majority of participants — 87% and 93% for 0_7U and 0_1U, respectively — expressed their willingness to continue playing both of the lower rated scenarios. Only one participant was not willing to continue playing 0_4U.

On the subject of collider scale, the most conclusive results were obtained for 0_2U as the least preferred option. Chosen for the worst scenario in the group by 73% of participants, this scenario received a significantly worse mean QoE score (M = 3.60, SD = 1.28) compared to both 0_5U (M = 4.40, SD = 0.72) and 1U (M = 4.30, SD = 0.75). Due to the implications of smaller colliders on placement precision, the 0_2U scenario significantly lowered participants’ perceived competence, in addition to making the task significantly more challenging, mentally demanding, and difficult to control compared to the other two. This was evidenced by objective measures, as the percentage of successful participants for this scenario (40%) was considerably smaller compared to 0_5U (87%) and 1U (93%). Moreover, the results of both temporal measures (average duration and average duration of successful completion) further indicate that participants were struggling with the placement of puzzle pieces in the smallest collider scale scenario. Remarkably, the percentage of participants willing to continue this scenario (83%) was comparable to the percentage of participants willing to continue the other two (86% for both), which is somewhat surprising considering that, in addition to its negative effects on player performance, the 0_2U scenario also received the lowest mean fun score in the group (although there were no statistically significant differences between scenarios in this case).

Even though both scenarios in the remote grab group received identical average QoE scores (4.50), a significant majority of participants preferred the REM_GRAB scenario, as seen in Figure 4. However, this preference was not evident from the average scores of examined QoE features, as REM_GRAB was deemed to be more challenging, mentally demanding, and more difficult to control, making participants feel less competent. Whereas all participants successfully completed the puzzle during the NO_GRAB scenario, only 70% of participants managed to do the same for the REM_GRAB scenario. Moreover, on average, completing the NO_GRAB scenario required significantly less time compared to REM_GRAB. These results may be somewhat unexpected, as the remote grab option was supposed to serve as an aid to participants. However, as evidenced by participants’ comments, enabling this option made it easier to accidently grab the wrong puzzle piece, or shift assembled pieces out of place. Still, the scenario received a significantly higher fun score compared to NO_GRAB, which is likely the main reason for it being chosen as a clear favourite as participants praised its novelty and magical qualities. Furthermore, despite the increased workload of picking up and placing the right pieces that came with the REM_GRAB scenario, multiple participants expressed their satisfaction with the ability to grab puzzle pieces from a distance, further stressing the importance of introducing accessibility improvements to VR games.

As seen in Figure 4, smaller scale offsets were preferred over the 1_OFF scenario, which left no buffer space between puzzle pieces assembled in the solution cube. The preference was stronger for the 0_8_OFF scenario, although its QoE score (M = 4.43, SD = 0.77), was only slightly higher compared to 0_9_OFF (M = 4.33, SD = 0.71) and 1_OFF (M = 4.27, SD = 0.87). Despite unremarkable differences between scenarios, the 1_OFF scenario consistently stood out from the other two, as it made participants feel less competent by significantly increasing the challenge and task difficulty compared to other scenarios. This is due to significant maneuvering required to place a new puzzle piece among existing puzzle pieces, densely packed inside of the solution cube. Moreover, this scenario’s scores for mental and physical demand were also slightly higher compared to the others, while its average fun score was slightly lower. The 1_OFF scenario also had the lowest success rate (63%), compared to 0_8OFF (100%) and 0_9OFF (93%). In general, however, participants were still mostly willing to continue playing either scenario in the group, with 87% reporting their willingness to continue 1_OFF, and 90% for the other two scenarios.

3.3 Shoot interaction mechanics (US3_SH)

Participants’ preferences for the best and worst scenario in each scenario group are listed in Figure 5. The results of Friedman tests, comparing the scenarios in each scenario group with respect to collected measures, are presented in Table 6.

FIGURE 5

Exploring the use of visual shooting aids, the highest number of participants preferred the use of laser. Even though participants avoided listing LAS_TRAJ as the best scenario in the group, its QoE score (M = 4.40, SD = 0.81) was close to those of LAS (M = 4.47, SD = 0.90) and NO_AID (M = 4.43, SD = 0.77), while TRAJ received the worst average QoE score (M = 4.17, SD = 0.91). While reported preferences and QoE ratings for this group of scenarios are mixed, individual ratings of different QoE features show significant differences between scenarios. Overall, scenarios that did not involve the use of laser (TRAJ and NO_AID) resulted in lower perceived competence, along with increases in challenge, pain and muscle fatigue, and all measured categories of workload compared to the other two scenarios. As expected, the NO_AID scenario was deemed the most demanding. Even though the LAS_TRAJ involved the use of two aiming aids in conjunction, participants found it more challenging than LAS, which only used laser. This may have been somewhat distracting to participants as using both aiming aids at once highlighted the offset between the straight-line laser beam and the slightly curved trajectory of the projectile as it succumbs to gravity. The results of subjective measures addressing competence, challenge, and workload were confirmed by task performance metrics, with the LAS scenario resulting in the accuracy of 80.21%, followed by LAS_TRAJ at 71.72%, TRAJ at 61.42% and NO_AID at 51.43%. However, even though it produced the worst performance of the scenario group, the NO_AID scenario (M = 4.43, SD = 0.86) was comparable with LAS (M = 4.40, SD = 0.97) and slightly higher from TRAJ (M = 4.33, SD = 0.88) in terms of fun, with LAS_TRAJ obtaining the lowest score in the group (M = 4.07, SD = 1.17).

As with visual shooting aids, results for target spawn angle were also mixed. Although participants were more likely to choose a wider spawn angle for their favourite scenario, the QoE rating of the 90_DEG scenario (M = 4.43, SD = 0.82) was higher than the 360_DEG scenario (M = 4.20, SD = 0.96), although both were exceeded by 180_DEG (M = 4.7, SD = 0.6). Likewise, participants were slightly more inclined to go on with playing the 90_DEG scenario, with 90% of participants stating their willingness to continue compared to 87% for 180_DEG and 77% for 360_DEG. This is likely due to the increased challenge, mental and physical workload, and task control difficulty of wider target spawn angles. Furthermore, increases in target spawn angle also increased the overall sense of physical discomfort, although there were no statistically significant differences for this feature. However, when it comes to fun, participants seemed to enjoy the 180_DEG scenario (M = 4.53, SD = 0.73) more than 90_DEG (M = 4.17, SD = 0.95) and 360_DEG (M = 4.23, SD = 0.90).

With regard to task performance metrics, the mean accuracy of expelled shots/projectiles for all three scenarios was relatively close, ranging from 60.80% for 360_DEG to 65.48%. It is necessary to point out, though, that the number of shots fired decreased with wider spawn angles, dropping from 174.17 for 90_DEG down to 103.93 for 360_DEG, which means that, overall, significantly more targets were successfully hit during the 90_DEG scenario.

As opposed to mixed results for visual shooting aids and target spawn angle, results for the examination of shoot force were very conclusive, with a large majority of participants preferring the 80N scenario. This scenario received the highest QoE score (M = 4.47, SD = 0.68), with 40N receiving the average score of 4.23 (SD = 0.77), and 20N receiving the average score of 3.63 (SD = 1.00). Furthermore, 97% of participants were willing to continue playing the 80N scenario, compared to 80% for 40N and only 50% for 20N. A higher shoot force also helped participants experience more fun and perceive themselves as more competent, while the lowest shoot force significantly increased the overall challenge of the scenario, as well as its mental demand and task control difficulty. With regard to VRISE, there were no significant differences between scenarios. Significant differences were found for task performance, as 80N resulted in the highest accuracy (64.53%) with 192.2 shots fired on average, although this performance was closely followed by the 40N scenario with the average of 186.27 shots fired and the mean accuracy of 62.73%. The increased workload of the 20N impaired the participants’ performance, with the average of 171.97 shots fired at the average accuracy of 51.92%.

4 Discussion

Analyzing all three IMs in parallel, it is evident there is a lot of obvious overlap between measures of player experience and different measures of workload, which is to be expected. General response to the challenge item had a tendency to increase along with the increases in some or all measures (e.g., certain tasks were considered mentally, but not physically demanding and vice versa) of workload, while competence — for the most part — increased in the opposite direction. Aforementioned subjective measures were generally in line with the objective measures of task performance. Another measured aspect of the player experience — fun — varied between scenarios, but was less consistent with the other measures, which will be further discussed later in this section.

With regard to VRISE, even though significant differences between scenarios were noted, this effect was not observed for every type of mechanics. Furthermore, measured levels of experienced VRISE were generally quite modest. This is likely due to particular study design choices, as well as specific characteristics of the tested platform. With regard to study design, the duration of each scenario was short and immediately followed by a short break. Moreover, the intensity of the material was varied, with constant switching between configurations that were less demanding, and those that were more demanding, meaning that even the most intense parts of the VR experience were very temporally limited. Additionally, because the application used to test these mechanics was not an actual game, it lacked the actual pacing and dynamics of an actual VR game, and omitted physically demanding secondary game mechanics such as physically interacting with another in-game entity, avoiding enemy attacks, or moving with respect to a specified rhythm. This likely reduced its physical demands, mitigating the risk of extensive VRISE.

Overall, it can be observed that there is no one-size-fits all approach to implementing each of the tested mechanics. However, in a significant number of cases, participants were inclined toward less demanding scenarios. Nonetheless, we did note some parameters for which participants preferred scenarios that were higher in terms of workload. Therefore it can be noted that, in the context of our studies, tested parameters may be roughly classified into three groups based on the comparative level of perceived workload experienced during the most fun scenario, as presented in Table 7. The first group consists of parameters for which added workload was deemed entertaining, and thus, the most demanding scenario was preferred. The second group consists of parameters for which some added challenge was deemed entertaining, but extensive workload started to detract from the experience. In such cases, the option with the moderate workload was generally preferred. The third group consists of parameters for which added workload was deemed draining (i.e., increased workload leads to a less fun experience), and participants favoured the least challenging option. The complex relationship between positive aspects of player experience and workload has previously been explored by Tammy Lin et al. (2023) in the context of the slash IM. The authors reflected on the variability of perceived enjoyment with respect to perceived difficulty, finding it was influenced by the individual traits of each participant.

With respect to specific findings, for the slash IM, our results show that an increased play angle leads to higher workload and lower fun, which is consistent with the results of Tammy Lin et al. (2023). Overall, the weapon length of 1U, which is consistent with the default saber length in Beat Saber (1 m), appeared to be preferred over both shorter and longer swords. Previous work by Yoo et al. (2018) suggests that VR gamers prefer light to moderate physical exertion over no exertion. While our results indicate that the scenario requiring the highest force to destroy the target object was the most disliked by participants, the least physically demanding scenario (0N) was surprisingly well tolerated in comparison, although the moderate scenario was deemed the most entertaining.

For the pick-and-place IMs, our initial expectation was that allowing users to obtain objects from a distance would facilitate their workload. However, this scenario was deemed comparatively more challenging, likely due to the relative proximity of objects to be acquired, along with the increased precision requirements of remote aiming. While the results of Kartick et al. (2020) suggest that isometric manipulation of objects leads to higher flow and positive affect, our findings suggest that — despite its comparatively higher workload — the non-isometric remote grab scenario resulted in higher levels of fun and was generally preferred by participants. This finding can potentially be attributed to non-isometric manipulation extending beyond real-life constraints (LaViola et al., 2017), and thus providing users with a more “magical” experience.

With respect to scaling, participants preferred larger colliders and more buffer space between designated puzzle placement positions. While puzzles in the two smaller sizes were well tolerated, our findings indicate that having to manipulate larger objects results in a worse experience during assembly tasks. In physical tasks involving object manipulation, users have to modify their grasp depending on object size, shape, weight, and material (Lee and Jung, 2014; Feix et al., 2014; Bergmann Tiest and Kappers, 2014; Paulun et al., 2016; Bergmann Tiest and Kappers, 2019). However, in controller-based manipulation of virtual objects, as implemented in our platform, all objects are virtually weightless and the grasp remains the same regardless of material and object size. In our case, participants performed all grasping actions by gripping the cylindrical handle of the Valve Index controller in what resembles a typical power grip. Because participants manipulated all objects using the same controller and grip, differences in performance cannot be attributed to changes in grasp strategy. Nevertheless, rotating larger objects may be more cumbersome than rotating smaller ones, as the same angular controller movement produces larger linear displacements at the extremities of a large object. In the context of a task requiring precise positioning and orientation of puzzle pieces, this may help explain why participants preferred smaller puzzle pieces and more forgiving scale-related parameters. Moreover, based on the results of Flusberg and Boroditsky (2011), it is possible that the increased motor demands of controlling larger objects during a precision task could have increased the cognitive load necessary for motor planning, further hindering participants’ performance and experience.

In the context of shoot mechanics, as expected, the use of a laser sight improved participants’ objective performance. While less pronounced compared to the laser sight, the setting with a visible trajectory improved accuracy relative to the scenario with an invisible bullet trajectory, presumably due to the activation of different strategies for anticipating projectile motion Bosco et al. (2012). However, our results suggest that these visual aids should be used separately. Participants preferred the highest shooting force, which resulted in a more straightforward trajectory and thus reduced the additional load associated with anticipating projectile movement along a visibly parabolic path. With regard to target spawn angle, participants in US3_SH were more willing to tolerate (and even preferred) wider angles compared to, for example, participants in US1_SL. These preferences are reminiscent of a comment made by a participant in the study by Yoo et al. (2017), who preferred a VR archery game over shooting at a real-life range specifically because it allowed targets to be positioned across the full 360° surrounding the player. However, this discrepancy between the findings in US1_SL and US3_SH is likely due to the greater urgency associated with targets succumbing to gravity if not slashed in time in US1_SL. It is likely that a more time-sensitive shooting context (for example, one in which targets were programmed to attack the player) would have yielded different results.

4.1 Recommendations for the implementation of VR IMs

Approaching each individual parameter, our results indicate certain findings regarding user preferences that could be utilized in the formulation of recommendations for the implementation of IMs. First and foremost, it is evident that there is no one universal solution equally embraced by all participants, which motivates the primary recommendation (labelled as R.G1) that can be generalized to all mechanics and parameters.

Following the specification of this general-purpose recommendation, Table 8 presents a set of IM-specific recommendations related to the implementation of slash (R.SLX), pick-and-place (R.PPX), and shoot (R.SHX) mechanics. Several of these guidelines are broadly consistent with the Seven Principles of Universal Design proposed by Story (1998). Specifically, R.G1 promotes Flexibility in Use (Principle 2) by enabling interaction mechanics to be adapted to users’ preferences and abilities. Recommendations addressing the implementation of the slash interaction mechanic (R.SL1 and R. SL2) emphasize Low Physical Effort (Principle 6). Recommendations for implementing the pick-and-place interaction mechanic (R.PP1–R.PP4) underscore the importance of providing appropriate Size and Space for Approach and Use (Principle 7), while also supporting Tolerance for Error (Principle 5) by minimizing the occurrence of accidental errors during object positioning and assembly. Finally, recommendations R. SH1 and R. SH3 for implementing the shoot interaction mechanic enhance Perceptible Information (Principle 4) and support Tolerance for Error (Principle 5) by increasing the visibility of the projectile trajectory and providing visual aiming aids.

4.2 Limitations and future work

While the findings presented in this chapter provide useful information on user preferences pertaining to the implementation of common VR IMs, it is necessary to point out that they are based on the opinions of a limited sample of users. In addition, participants in all three studies were young and healthy, with no mobility or other limitations that would have significantly affected their experience. Thus, their experiences with workload and VRISE are not indicative of the issues that may be encountered by users of a different demographic. The majority of participants were also quite inexperienced with VR, which may have influenced their opinions and preferences, especially when considering the possible novelty effect. While background information regarding general expertise with gaming was not collected for these studies, choosing participants with this factor in mind could have provided useful context for the interpretation of results.

Despite our efforts to mitigate order effects, it is important to note that even brief exposure to certain mechanics may induce learning effects (as demonstrated in the context of shooting by Sudiarno et al. (2024)), which may have influenced our results. However, it is also necessary to highlight the duration of each scenario, which may have been too short to produce significant effects on measured features. Our results may have been influenced by our choice of input modality. For example, when considering our findings in the context of designing realistic shooting simulators, it is important to note that the controllers used in our study lacked the realistic gun-mimicking shape, weight, and recoil capabilities provided by dedicated force-feedback controllers (further explored by Berna-Moya and Martinez-Plasencia (2019)). Moreover, while the application used as test material is based on existing games, its implementation is simplified and lacks highly relevant elements of commercial game design that could have made a significant impact on user experience, such as the inclusion of secondary game mechanics, visual progress indicators, narrative, a more sophisticated artistic design, etc.

While multiple popular VR games on the market utilize minimal backdrops with contrasting detail (e.g., Beat Saber, Superhot VR, Space Pirate Trainer), it is expected that commercial solutions may present the user with significantly more visual clutter compared to the test platform. Thus, while allowing for a highly controlled experiment, its minimalistic visual design may come at a cost of limited generalizability of our results. Moreover, it is also necessary to point out that the chosen visual environment, in which the user is situated on a platform surrounded by the night sky, was found uncomfortable by two participants with acrophobic tendencies. Even though participants in question decided to move forward with their participation in the study, we consider this issue an oversight on our part. Therefore, we advise acrophobia screening for potential participants wishing to enroll in any future studies utilizing the application in its current form.

With regard to observed parameters and objective measures, the application itself provides numerous additional parameters (e.g., frequency with which targets are spawned, target size, target velocity and direction of movement, haptic feedback, etc.) and parameter values, and collects detailed measures that have not been explored in this paper (e.g., specific data pertaining to the lifetime, size, position and orientation of each individual target, forces exhibited by the player, whether left or right hand was used, etc.). The inclusion of these additional parameters and measures into our methodology may provide useful insights into the user experience and will be considered in future work. Furthermore, the conclusions reached in this section may have been different if the implementations of existing mechanics, parameters, and measures were further refined or extended to a wider range of configurations. Certain concepts explored in this research, like challenge and workload, are likely intricately linked, and adding a more diverse range of measures could have provided a more comprehensive view of user preferences. Lastly, it needs to be stressed that subjective measures used herein — while inspired by existing questionnaires — have not been standardized. With all this in mind, it is worth reiterating that the primary goal of these exploratory studies was to shed light on understudied elements of VR IMs. Additional studies approaching this topic are necessary to corroborate the findings presented in this paper. Further advancements in this field of research could also be applied to AI-driven performance prediction and the personalization of gaming experiences, representing a promising direction for the future of gaming (Lopez-Gordo et al., 2021).

Finally, the proposed recommendations should be interpreted in the context of each application and its respective mechanics, objectives, and gameplay structure. For example, recommendation R.G1 emphasizes providing users with multiple options so that the IM can be adapted to their preferences and abilities. However, depending on the game, players may also compete against time constraints, AI-controlled elements and opponents, or other players. In such cases, certain IM configurations may reduce or increase the level of challenge, or otherwise affect objective measures of performance, which may influence game progression or outcomes to a significant degree. Developers should therefore consider the impact of configurable interaction options on gameplay balance and fairness and, where appropriate, adapt game rules, scoring systems, matchmaking, or game modes accordingly. The design of such game balancing strategies falls outside the scope of this work, but warrants targeted investigation.

5 Conclusion

Enabled by 6DoF tracking and the immersive output capabilities of contemporary commercial systems, VR games usually offer interaction opportunities that go beyond traditional touch- or button-based controls of more established platforms such as desktop, console or mobile. VR gamers are often encouraged to perform fast-paced, high-fidelity actions requiring gross motor movements. While such mechanics have been shown to increase game enjoyment, if they are not properly calibrated to the player’s preferences and capabilities, they may also lead to increased workload and the onset of VRISE. In the absence of specific guidelines for implementing interaction mechanics, game development teams must rely solely on internal practices and in-house testing to determine the optimal configuration of their chosen IM. To address this issue, in this paper we focused on the systematic evaluation of the effect of various IM parameters on player experience and task performance, building upon our previous work and existing academic sources. Bridging the gap between academia and industry, we summarized our findings in a set of empirically grounded recommendations aimed at VR developers and researchers alike.

First, we presented the results of three user studies involving a total of 90 participants. Considering that the research objective of investigating VR IMs requires a more in-depth approach, instead of using commercial games as test material, we chose to utilize a custom application specifically designed for this purpose, which provided the means to evaluate different configurations of established VR game mechanics — slash (US1_SL), pick-and-place (US2_PP), and shoot (US3_SH) mechanics. Subjective measures pertaining to player experience, workload, and VRISE were supplemented with IM-specific objective measures provided by the test application, ranging from accuracy and force exerted by players to duration needed for completing the task. After analyzing collected results, we found that acceptance of particular parameter values varied between participants, indicating that an ideal solution to IM implementation likely does not exist. However, we identified certain trends regarding participant preferences, which then served as a foundation for the proposed set of practical recommendations for the implementation of VR IMs.

As for future work, the application used as test material provides the means for future experiments by offering a broad range of additional parameters and measures that have not yet been explored in the scope of this paper or otherwise. Moreover, future research endeavours could focus on exploring additional IMs, parameters, and measures. Finally, future studies may be conducted using a larger sample of more experienced VR users, or participants belonging to different demographics.

References

• 1

  AwadaM.ZhuR.Becerik-GerberB.LucasG.SouthersE. (2021). An integrated emotional and physiological assessment for vr-based active shooter incident experiments. Adv. Eng. Inf.47, 101227. 10.1016/j.aei.2020.101227

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  BelyeaC.RummelG.LanhamN.KrulK. (2025). Upper extremity and musculoskeletal injuries related to videogame and augmented reality game usage: a narrative review article. Curr. Rev. Musculoskelet. Med.18, 201–206. 10.1007/s12178-025-09956-9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  Bergmann TiestW. M.KappersA. M. (2014). “The influence of material cues on early grasping force,” in International Conference on Human Haptic Sensing and Touch Enabled Computer Applications (Springer), 393–399.

  • Google Scholar

• 4

  Bergmann TiestW. M.KappersA. M. (2019). The influence of visual and haptic material information on early grasping force. R. Soc. Open Science6, 181563. 10.1098/rsos.181563

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 5

  Berna-MoyaJ. L.Martinez-PlasenciaD. (2019). “Exploring the effects of replicating shape, weight and recoil effects on vr shooting controllers,” in IFIP Conference on Human-Computer Interaction (Springer), 763–782. 10.1007/978-3-030-29381-9_45

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  BoscoG.Delle MonacheS.LacquanitiF. (2012). Catching what we can’t see: manual interception of occluded fly-ball trajectories. PLoS One7, e49381. 10.1371/journal.pone.0049381

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  BowmanD. A. (2002). “Principles for the design of performance-oriented interaction techniques,” in Handbook of Virtual Environments (Boca Raton, FL: CRC Press), 317–340.

  • Google Scholar
  • View reference in article

• 8

  BowmanD. A.JohnsonD. B.HodgesL. F. (1999). “Testbed evaluation of virtual environment interaction techniques,” in Proceedings of the ACM symposium on Virtual reality software and technology, 26–33. 10.1145/323663.323667

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  BugaA.DeckerD. D.RobinsonB. T.CrabtreeC. D.StonerJ. T.ArceL. F.et al (2024). The virtra v-100 is a test-retest reliable shooting simulator for measuring accuracy/precision, decision-making, and reaction time in civilians, police/swat, and military personnel. J. Strength and Cond. Res.38, 1714–1723. 10.1519/JSC.0000000000004875

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 10

  CaggianeseG.GalloL.NeroniP. (2018). “The vive controllers vs. leap motion for interactions in virtual environments: a comparative evaluation,” in Intelligent Interactive Multimedia Systems and Services (Springer), 24–33. 10.1007/978-3-319-92231-7_3

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  ChoH.YuanB.HartJ. D.ChangZ.CaoJ.ChangE.et al (2023). “Time travellers: an asynchronous cross reality collaborative system,” in 2023 IEEE International Symposium on Mixed and Augmented Reality Adjunct (ISMAR-Adjunct), Sydney, Australia (IEEE), 848–853. 10.1109/ISMAR-ADJUNCT60411.2023.00186

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  CobbS. V.NicholsS.RamseyA.WilsonJ. R. (1999). Virtual reality-induced symptoms and effects (vrise). Presence Teleoperators and Virtual Environ.8, 169–186. 10.1162/105474699566152

  • CrossRef
  • Google Scholar

• 13

  CoutinhoC. (2022). Snap zones. In Unity® Virtual Reality Development with VRTK4: A No-Coding Approach to Developing Immersive VR Experiences, Games, and Apps (Springer). 225–252.

  • Google Scholar
  • View reference in article

• 14

  CoutuY.ChangY.ZhangW.ŞengünS.LcR. (2020). “Immersiveness and usability in vr: a comparative study of monstrum and fruit ninja,” in Game User Experience and Player-Centered Design (Springer), 437–448. 10.1007/978-3-030-37643-7_20

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  CsikszentmihalyiM. (1990). Flow: The Psychology of Optimal Experience. New York: Harper and Row.

  • Google Scholar
  • View reference in article

• 16

  CuiD.MousasC. (2021). “Evaluating wearable tactile feedback patterns during a virtual reality fighting game,” in 2021 IEEE International Symposium on Mixed and Augmented Reality Adjunct (ISMAR-Adjunct), Bari, Italy (IEEE), 328–333. 10.1109/ISMAR-Adjunct54149.2021.00075

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  DaneelsR.MallietS.KoemanJ.RibbensW. (2018). The enjoyment of shooting games: exploring the role of perceived realism. Comput. Hum. Behav.86, 330–336. 10.1016/j.chb.2018.04.053

  • CrossRef
  • Google Scholar

• 18

  De ArmasC.ToriR.NettoA. V. (2020). Use of virtual reality simulators for training programs in the areas of security and defense: a systematic review. Multimedia Tools Appl.79, 3495–3515. 10.1007/s11042-019-08141-8

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  de Sousa AbreuI.SilvaL. F. M. S. (2025). “BellatorVR: a virtual reality application to support firearms training for military police,” in Proceedings of the Seminário Integrado de Software e Hardware (SEMISH) (Porto Alegre, Brazil: Sociedade Brasileira de Computação (SBC)), 477–488. 10.5753/semish.2025.9061

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  EllahiyounK.PrettyE. J.GuareseR.TakacM.FayekH.ZambettaF. (2025). Effects of task difficulty and music expertise in virtual reality: observations of cognitive load and task accuracy in a rhythm exergame. arXiv Preprint arXiv:2507.06691. 10.48550/arXiv.2507.06691

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  FabricatoreC. (2007). “Gameplay and game mechanics design: a key to quality in videogames,” in Proceedings of the OECD-CERI Expert Meeting on Videogames and Education (Chile: Santiago de Chile), 29–31.

  • Google Scholar
  • View reference in article

• 22

  FedasyukD.VorobiovV. (2023). “Virtual reality training simulator for weapons shooting,” in 2023 IEEE 5th International Conference on Advanced Information and Communication Technologies (AICT), Lviv, Ukraine (IEEE), 57–61. 10.1109/AICT61584.2023.10452699

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  FeixT.BullockI. M.DollarA. M. (2014). Analysis of human grasping behavior: object characteristics and grasp type. IEEE Transactions Haptics7, 311–323. 10.1109/TOH.2014.2326871

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  FlusbergS. J.BoroditskyL. (2011). Are things that are hard to physically move also hard to imagine moving?Psychonomic Bulletin and Review18, 158–164. 10.3758/s13423-010-0024-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 25

  FoxmanM.LeithA. P.BeyeaD.KlebigB.ChenV. H. H.RatanR. (2020). “Virtual reality genres: comparing preferences in immersive experiences and games,” in Extended Abstracts of the 2020 Annual Symposium on Computer-Human Interaction in Play, 237–241.

  • Google Scholar

• 26

  GamageN. M.IshtaweeraD.WeigelM.WithanaA. (2021). “So predictable! continuous 3d hand trajectory prediction in virtual reality,” in 34th Annual ACM Symposium on User Interface Software and Technology, 332–343. 10.1145/3472749.3474753

  • CrossRef
  • Google Scholar

• 27

  GusaiE.BassanoC.SolariF.ChessaM. (2017). “Interaction in an immersive collaborative virtual reality environment: a comparison between leap motion and htc controllers,” in International Conference on Image Analysis and Processing, 290–300. 10.1007/978-3-319-70742-6_27

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  HaleK. S.StanneyK. M. (2014). Handbook of Virtual Environments: Design, Implementation, and Applications. Boca Raton, FL: CRC Press.

  • Google Scholar
  • View reference in article

• 29

  HameedA.PerkisA.MöllerS. (2021). “Evaluating hand-tracking interaction for performing motor-tasks in vr learning environments,” in 2021 13th International Conference on Quality of Multimedia Experience (QoMEX), Montreal, QC, Canada (IEEE), 219–224. 10.1109/QoMEX51781.2021.9465407

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  HarrisD.WilsonM.VineS. (2020). Development and validation of a simulation workload measure: the simulation task load index (Sim-tlx). Virtual Real.24, 557–566. 10.1007/s10055-019-00422-9

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  HarrisD. J.ArthurT.KearseJ.OloniluaM.HassanE. K.De BurghT. C.et al (2023). Exploring the role of virtual reality in military decision training. Front. Virtual Real.4, 1165030. 10.3389/frvir.2023.1165030

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  HeydnK. A. M.DietrichM. P.BarkowskyM.WinterfeldtG.von MammenS.NüchterA. (2019). “The golden bullet: a comparative study for target acquisition, pointing and shooting,” in 2019 11th International Conference on Virtual Worlds and Games for Serious Applications (VS-Games), Vienna, Austria (IEEE), 1–8. 10.1109/VS-Games.2019.8864589

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  IjsselsteijnW. A.De KortY. A. W.PoelsK. (2013). The Game Experience Questionnaire. Eindhoven, The Netherlands: Technische Universiteit Eindhoven, 3–9.

  • Google Scholar
  • View reference in article

• 34

  International Telecommunication Union (2018). Recommendation ITU-T P.809: Subjective Evaluation of Speech Quality with a Crowdsourcing Approach. Geneva, Switzerland: ITU-T.

  • Google Scholar
  • View reference in article

• 35

  IslamF.BrightZ.ChengM.ShiC. (2024). The effect of the metaverse environment and task difficulty on emotional response and cognitive workload. Int. Symposium Hum. Factors Ergonomics Health Care13, 101–105. 10.1177/2327857924131044

  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  KangasJ.KumarS. K.MehtonenH.JärnstedtJ.RaisamoR. (2022). Trade-off between task accuracy, task completion time and naturalness for direct object manipulation in virtual reality. Multimodal Technol. Interact.6, 6. 10.3390/mti6010006

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  KartickP.QuevedoA. J. U.GualdronD. R. (2020). “Design of virtual reality reach and grasp modes factoring upper limb ergonomics,” in 2020 IEEE Conference on Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW), Atlanta, GA, USA (IEEE), 798–799. 10.1109/VRW50115.2020.00250

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  KatzourinM.IgnatoffD.QuirkL.LaViolaJ. J.JenkinsO. C. (2006). Swordplay: innovating game development through vr. IEEE Comput. Graph. Appl.26, 15–19. 10.1109/MCG.2006.137

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  KocurM.DechantM. J.LankesM.WolffC.MandrykR. (2020). “Eye caramba: Gaze-based assistance for virtual reality aiming and throwing tasks in games,” in ACM Symposium on Eye Tracking Research and Applications, 1–6. 10.1145/3379156.3391841

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  KrompiecP.ParkK. (2019). Enhanced player interaction using motion controllers for first-person shooting games in virtual reality. IEEE Access7, 124548–124557. 10.1109/ACCESS.2019.2937937

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  LaViolaJr., J. J.KruijffE.McMahanR. P.BowmanD. A.PoupyrevI. P. (2017). 3D User Interfaces: Theory and Practice. 2 edn. Boston, MA: Addison-Wesley Professional.

  • Google Scholar
  • View reference in article

• 42

  LeeK.-S.JungM.-C. (2014). Common patterns of voluntary grasp types according to object shape, size, and direction. Int. J. Industrial Ergonomics44, 761–768. 10.1016/j.ergon.2014.08.005

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  LiuR.Becerik-GerberB.LucasG. M. (2023). Effectiveness of vr-based training on improving occupants’ response and preparedness for active shooter incidents. Saf. Sci.164, 106175. 10.1016/j.ssci.2023.106175

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  Lopez-GordoM. A.KohlmorgenN.MorillasC.PelayoF. (2021). Performance prediction at single-action level to a first-person shooter video game. Virtual Real.25, 681–693. 10.1007/s10055-020-00482-2

  • CrossRef
  • Google Scholar

• 45

  LuroF. L.SundstedtV. (2019). “A comparative study of eye tracking and hand controller for aiming tasks in virtual reality,” in Proceedings of the 11th ACM Symposium on Eye Tracking Research and Applications, 1–9. 10.1145/3314111.3318153

  • CrossRef
  • Google Scholar
  • View reference in article

• 46

  MatokanovićM. (2022). Quality of Experience Assessment of Interaction Mechanics in Virtual Reality. Zagreb, Croatia: University of Zagreb. Master’s thesis.

  • Google Scholar
  • View reference in article

• 47

  Meta (2026). Best-Selling of all Time VR Games on Meta Quest. Meta Quest Store.

  • Google Scholar
  • View reference in article

• 48

  MonteiroD.LiangH.-N.WangJ.ChenH.BaghaeiN. (2020). “An in-depth exploration of the effect of 2d/3d views and controller types on first person shooter games in virtual reality,” in 2020 IEEE International Symposium on Mixed and Augmented Reality (ISMAR), Porto de Galinhas, Brazil (IEEE), 713–724. 10.1109/ISMAR50242.2020.00102

  • CrossRef
  • Google Scholar

• 49

  MüsselerJ.von Salm-HoogstraetenS.BöffelC. (2022). Perspective taking and avatar-self merging. Front. Psychol.13, 714464. 10.3389/fpsyg.2022.714464

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 50

  PallaviciniF.PepeA. (2020). Virtual reality games and the role of body involvement in enhancing positive emotions and decreasing anxiety: within-subjects pilot study. JMIR Serious Games8, e15635. 10.2196/15635

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 51

  PaulunV. C.GegenfurtnerK. R.GoodaleM. A.FlemingR. W. (2016). Effects of material properties and object orientation on precision grip kinematics. Exp. Brain Research234, 2253–2265. 10.1007/s00221-016-4631-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 52

  PoelsK.de KortY. A.IjsselsteijnW. A. (2007). D3.3: game experience questionnaire: development of a self-report measure to assess the psychological impact of digital games

  • Google Scholar

• 53

  PoupyrevI.WeghorstS.BillinghurstM.IchikawaT. (1997). “A framework and testbed for studying manipulation techniques for immersive vr,” in Proceedings of the ACM symposium on Virtual reality software and technology, 21–28. 10.1145/261135.261141

  • CrossRef
  • Google Scholar

• 54

  RachevskyD. C.de SouzaV. C.NedelL. (2018). “Visualization and interaction in immersive virtual reality games: a user evaluation study,” in 2018 20th Symposium on Virtual and Augmented Reality (SVR) (IEEE), 89–98. 10.1109/SVR.2018.00024

  • CrossRef
  • Google Scholar

• 55

  RogersK.FunkeJ.FrommelJ.StammS.WeberM. (2019). “Exploring interaction fidelity in virtual reality: object manipulation and whole-body movements,” in Proceedings of the 2019 CHI conference on human factors in computing systems, 1–14. 10.1145/3290605.3300644

  • CrossRef
  • Google Scholar

• 56

  SaravasV.PahalchukE.MolchanA.KampovY. (2020). “The development of educational application with virtual reality placing objects system using snap zone technology,” in 2020 IEEE 11th International Conference on Dependable Systems, Services and Technologies (DESSERT) (IEEE), 416–421. 10.1109/DESSERT50317.2020.9125023

  • CrossRef
  • Google Scholar

• 57

  SitzmannV.SerranoA.PavelA.AgrawalaM.GutierrezD.MasiaB.et al (2018). Saliency in vr: how do people explore virtual environments?IEEE Transactions Visualization Computer Graphics24, 1633–1642. 10.1109/TVCG.2018.2793599

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 58

  SlaterM.SpanlangB.Sanchez-VivesM. V.BlankeO. (2010). First person experience of body transfer in virtual reality. PloS One5, e10564. 10.1371/journal.pone.0010564

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 59

  StoryM. F. (1998). Maximizing usability: the principles of universal design. Assist. Technology10, 4–12. 10.1080/10400435.1998.10131955

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 60

  SudiarnoA.DewiR. S.WidyaningrumR.AkbarR. A.SudiantoY.PrastyabudiW. A.et al (2024). Analysis of human performance and potential application of virtual reality (vr) shooting games as a shooting training simulator for military personnel. Int. J. Technol.15, 87–98. 10.14716/ijtech.v15i1.5303

  • CrossRef
  • Google Scholar

• 61

  SweetserP.RogalewiczZ. (2020). “Affording enjoyment in vr games: possibilities, pitfalls, and perfection,” in Proceedings of the 32nd Australian Conference on Human-Computer Interaction, 55–64. 10.1145/3441000.3441050

  • CrossRef
  • Google Scholar

• 62

  SzpakA.MichalskiS. C.LoetscherT. (2020). Exergaming with beat saber: an investigation of virtual reality aftereffects. J. Med. Internet Res.22, e19840. 10.2196/19840

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 63

  Tammy LinJ.WuD.BowmanN. (2023). Beat saber as virtual reality exercising in 360 degrees: a moderated mediation model of vr playable angles on physiological and psychological outcomes. Media Psychol.26, 414–435. 10.1080/15213269.2022.2154806

  • CrossRef
  • Google Scholar

• 64

  TianH.LeeG. A.BaiH.BillinghurstM. (2023). Using virtual replicas to improve mixed reality remote collaboration. IEEE Trans. Vis. Comput. Graph.29, 2785–2795. 10.1109/TVCG.2023.3247113

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 65

  VlahovićS. (2024). Assessing the Quality of Experience of Virtual Reality Games. [Ph.D. thesis], University of Zagreb Faculty of Electrical Engineering and Computing, Zagreb, Croatia.

  • Google Scholar

• 66

  VlahovicS.SuznjevicM.Skorin-KapovL. (2022). A framework for the classification and evaluation of game mechanics for virtual reality games. Electronics11, 2946. 10.3390/electronics11182946

  • CrossRef
  • Google Scholar

• 67

  WangH.CheX.ChangE.QuC.LuoY.WeiZ. (2024). How to set safety boundary in virtual reality: a dynamic approach based on user motion prediction. Comput. Animat. Virtual Worlds35, e2210. 10.1002/cav.2210

  • CrossRef
  • Google Scholar

• 68

  WeiseM.ZenderR.LuckeU. (2020). How can i grab that? Solving issues of interaction in vr by choosing suitable selection and manipulation techniques. i-Com19, 67–85. 10.1515/icom-2020-0011

  • CrossRef
  • Google Scholar

• 69

  YamamotoT.FukuchiK. (2022). “Experiencing high-speed slash action in virtual reality environment,” in ICAT-EGVE (Posters and Demos), 41–42.

  • Google Scholar

• 70

  YooS.AckadC.HeywoodT.KayJ. (2017). “Evaluating the actual and perceived exertion provided by virtual reality games,” in Proceedings of the 2017 CHI Conference Extended Abstracts on Human Factors in Computing Systems, 3050–3057.

  • Google Scholar

• 71

  YooS.CarterM.KayJ. (2018). “Vrmove: design framework for balancing enjoyment, movement and exertion in vr games,” in Proceedings of the 2018 Annual Symposium on Computer-Human Interaction in Play Companion Extended Abstracts, 295–307. 10.1145/3270316.3272054

  • CrossRef
  • Google Scholar

• 72

  YoonD.HanS.ParkI.ChungT. (2024). Analyzing vr game user experience by genre: a text-mining approach on meta quest store reviews. Electronics13, 3913. 10.3390/electronics13193913

  • CrossRef
  • Google Scholar