Virtual and Augmented Reality in Sports
Linas JuozėnasShare
Virtual & Augmented Reality
A simulated world can make practice more accessible, varied and engaging. Its lasting value comes from what people can do with the experience.
A player can face another serve without retrieving a ball. A learner can explore a procedure before handling real equipment. Someone working on movement after illness can interact with a task adjusted to their reach and pace.
These possibilities deserve both curiosity and careful evaluation. The useful questions are concrete: what is being practised, what does the system actually measure, and does progress carry into the activity that matters?
A higher game score, better fitness and a transferable skill answer different questions.
Timing, decisions, movement and physical forces matter in different proportions for different tasks.
Comfort, meaningful choice, suitable challenge and access determine whether a tool is usable.
VR, AR, mixed reality and the systems between them
Virtual reality (VR) presents a computer-generated environment. An immersive headset can surround the user’s view and update the scene as they move. Research also uses “VR” for screen-based and projected environments, so the word alone does not identify the equipment.
Augmented reality (AR) adds digital information to a view of the physical world. Mixed reality (MR) often describes experiences in which digital objects are spatially registered to the surroundings and can interact with them. These labels overlap across manufacturers and researchers. Extended reality (XR) is a convenient umbrella term.[1]
| Format | What the person experiences | A possible training use |
|---|---|---|
| Immersive VR | A headset presents a virtual scene that responds to tracked movement. | Practising a sport-related visual decision or an interactive movement task. |
| Screen or projection simulator | A virtual task appears on a monitor or large screen, sometimes with specialised controls. | Instrument handling or a tracked batting task. |
| AR on a phone or tablet | Digital content appears over a camera view on a handheld screen. | Inspecting a model or following an equipment demonstration. |
| AR or MR through a headset | The surroundings remain visible through optics or camera passthrough, with digital additions. | Displaying a cue near the relevant object in a controlled practice space. |
| 360-degree video | The viewer can look around a recorded scene; interaction and viewpoint changes depend on the recording and software. | Observing a situation and discussing what information matters. |
The field has developed through several routes: simulation equipment, computer graphics, motion sensing, games and wearable displays. Modern headsets are one part of that history. A study of an older simulator can still teach an important lesson, provided its results remain attached to the task and equipment actually tested.
Tracking, interaction and the task behind the display
A training environment combines a display, input devices, software and a physical space. Depending on the system, cameras, inertial sensors or external trackers estimate the position and orientation of a headset, controller, hand or piece of equipment.
“Six degrees of freedom” describes three directions of translation and three axes of rotation. It does not mean that every joint is measured. Technical interfaces such as WebXR distinguish tracked poses, reference spaces and input sources; the application must decide how those signals control the task.[2]
Interaction determines what is actually practised
Watching an opponent, pressing a button to select a response and physically intercepting a ball are different activities. Each may be useful for a particular purpose. They provide different information about whether a person can coordinate a complete sporting action.
A controller can stand in for a racket while differing in grip, mass and balance. A virtual hand can appear to grasp an object even when the real fingers have received little information about its texture or resistance. These differences should shape the training goal.
Recording has limits
Some systems store movement traces, decisions and timing. Others provide only a score. A replay can show the modelled event from another viewpoint when the relevant data were recorded; it cannot reconstruct every unmeasured detail of the person’s body or surroundings.
Before using a dashboard, ask what the sensor observes, what the software estimates, and what the score rewards. A “successful” repetition may mean that a controller entered a target zone. Whether that was achieved through the intended movement requires a separate check.
“What must I do to succeed in this simulation, and how closely does that match what I need to do outside it?”
How much realism does a skill need?
Immersion describes properties of the system, such as how fully it surrounds the senses and responds to movement. Presence is the person’s experience of being there. Learning concerns what they retain and can use. A convincing experience can support engagement without establishing either of the other outcomes.
In one university science-simulation experiment, immersive VR produced greater reported presence but lower learning-test performance than a desktop version.[3] This was a particular educational task, not a verdict on every headset. It illustrates why visual impact and learning should be evaluated separately.
| Demand | What deserves attention | What a convincing scene cannot establish |
|---|---|---|
| Perception and timing | Are the useful cues visible at the right moment, scale and speed? | That a beautiful stadium reproduces a real opponent’s information. |
| Decision and action | Does the learner choose and carry out a response that resembles the intended task? | That selecting an answer proves full movement competence. |
| Contact and force | Are grip, resistance, impact or support forces important to the skill? | That a vibration recreates a collision, heavy object or tissue resistance. |
| Physical environment | Which demands come from temperature, oxygen availability, terrain or equipment? | That virtual snow or mountains produce cold or altitude acclimatisation. |
| Pressure and teamwork | Are communication, uncertainty and consequences represented usefully? | That virtual spectators guarantee confidence under competition pressure. |
This is task-relevant fidelity: preserving the features needed for the intended lesson. A simple display may be adequate for learning a sequence. Handling a tool may require a physical replica. A contact sport may still need extensive real practice to develop force control and interaction with another person.
Selective simplification can also be useful. An instructor might first reduce distracting background activity so a learner can identify one cue, then restore complexity. The simplification should be deliberate, and later assessment should include the demands that were temporarily removed.
Learning, retention and transfer
Practice performance is what a person can do during a session. Retention asks what remains after a delay. Transfer asks whether learning helps in a different, relevant setting. Research on learning shows that immediate improvement can be an unreliable guide to durable change.[4]
For XR training, this suggests three separate checks: performance within the programme, a later attempt with less assistance, and a meaningful task outside the simulation. The delay, task and assessment conditions should be chosen for the skill; there is no single universal test schedule.
Movement skill includes choosing and adjusting
Consider a racket player. Useful practice may involve noticing the incoming ball, deciding where to send it, organising the stroke and responding to the result. Repeating an isolated arm movement addresses only part of that problem.
A practical session can therefore have a narrow purpose: recognise one kind of serve, place the return into a broad target area, or maintain a workable recovery position. Once the task is understood, the instructor can vary speed, direction or uncertainty without changing everything at once.
Observation should lead to an attempt
A replay becomes more useful when it answers a question. Ask what information was available, what the learner chose and what they would change next. Then provide an opportunity to try the adjustment. The number of camera angles matters less than whether the learner can connect the explanation to an action.
It is also worth looking for habits that the simulation rewards but the real activity does not. An unusually short swing or memorised sequence may earn points while avoiding an important demand. That is a reason to change the exercise or scoring rule.
Successful learning need not feel effortless. Equally, confusion, nausea and repeated failure are not evidence of productive challenge. The learner should understand the task and have a plausible way to improve.
Feedback, haptics and what measurements mean
Feedback can describe an outcome, such as where a shot landed, or an aspect of movement, such as when a turn began. It can arrive visually, through sound or through touch. Research on augmented feedback shows that its usefulness depends on the task, learner and delivery; adding more channels is not automatically better.[5]
An original practical approach is to agree on one question before a short practice block. Review the most relevant information afterward, choose one adjustment, and try again. If the goal is independent performance, include attempts with guidance reduced or removed. A cue that remains essential during use should be evaluated as continuing assistance.
| Output | What it can represent | What needs separate evidence |
|---|---|---|
| Position or movement path | Tracked or estimated locations of a controller, body segment or tool. | Accuracy for the actual movement, speed, clothing and environment. |
| Joint angle or movement score | A calculation based on tracked points and a model. | Clinical meaning, acceptable variation and relevance to the intended task. |
| Vibration or touch cue | A signal that an event happened or an action is requested. | Realistic force, weight or contact mechanics. |
| Heart-rate display | A physiological signal when a suitable sensor supplies it. | Why the signal changed; it does not identify a single emotion or training benefit. |
| Gaze estimate | Where a system estimates the eyes are directed. | What the person understood, intended or felt. |
| Energy or injury-risk estimate | An algorithmic prediction based on selected inputs. | Validation for the person and purpose; it is not direct measurement of metabolism or tissue damage. |
Haptics and biofeedback describe different things
Haptics concerns touch and force sensations. A vibrating controller can mark contact without reproducing the forces of an actual racket strike. Specialised devices can add resistance, but their capabilities have to be assessed individually.
Biofeedback uses information about a physiological process to help someone learn to influence it. A movement trace is ordinary movement feedback unless a physiological signal is also involved. A heart-rate number alone does not establish that a biofeedback intervention improves sporting performance.
For decisions about pain, injury or rehabilitation, a colourful movement score should remain one piece of information. It cannot replace the person’s account of symptoms, relevant examination or a suitable functional assessment.
What studies show about transfer
Real-world assessment is possible, and several studies provide concrete examples. Their value comes from the detail: who trained, which system they used, what the comparison group did and how performance was tested.
| Study | System and finding | What the result does not establish |
|---|---|---|
| Table tennis · Michalski and colleagues, 2019[6] | 57 participants were analysed. The immersive VR group improved real table-tennis test performance more than a no-training group; an expert assessor was unaware of group assignment. | The study did not compare VR with an equivalent programme of coached real practice or establish elite competition benefits. |
| Baseball · Gray, 2017[7] | 80 high-school players entered four groups. Adaptive practice with a projected environment and tracked real bat improved real batting measures; the study also included delayed testing. | The intervention was screen-based. Results differed across measures and comparators, so they do not establish that every VR programme outperforms real batting practice. |
| Laparoscopic surgery · Seymour and colleagues, 2002[8] | 16 surgical residents were randomised. Task-specific simulator training to a proficiency criterion improved aspects of subsequent supervised operating-room performance. | This was a specialised laparoscopic simulator, not a modern entertainment headset. It cannot establish effectiveness for a different sport or clinical task. |
Read the comparison as carefully as the headline
Improvement compared with no extra practice answers a useful first question. Comparison with an equal amount of high-quality conventional practice answers a more demanding one. If the XR group receives additional time, attention or coaching, those additions may help explain the result.
Also ask whether the test resembles the practice too closely. A new score on the same familiar scenario can show adaptation to that scenario. An unfamiliar but relevant task provides a stronger challenge to claims of broad transfer.
The most useful evidence for a local decision concerns a comparable population, a clear learning goal and an outcome worth improving. Small promising studies can justify further evaluation. They should not be converted into promises of guaranteed mastery, injury prevention or professional success.
Where augmented reality can help
AR can place a label, target or instruction near the thing a learner is using. An illustrative example is a learner inspecting a stationary piece of exercise equipment: a cue could identify an adjustment point while an instructor explains why it matters.
The format is useful when it reduces the effort of finding relevant information. It becomes less useful when the learner must read long paragraphs, chase moving symbols or divide attention between an overlay and an important physical event.
Alignment is part of accuracy
A spatially anchored cue depends on tracking, calibration and rendering. Microsoft’s technical guidance describes how instability and alignment problems can make digital objects appear to move or sit incorrectly in the world.[9] In practical terms, an arrow that looks precise can still indicate the wrong place.
Camera passthrough and optical see-through displays also present the surroundings differently. Visibility, field of view, image delay and occlusion need to be assessed on the actual system. An overlay is not a reliable detector of every obstacle or hidden hazard.
Choose between learning support and continuing assistance
If an overlay is intended to teach a sequence, eventually test the sequence without it. If the intention is to provide a continuing reference during work, evaluate whether people use it accurately, notice when it fails and can obtain help.
These are both legitimate goals. Confusing them can lead to declaring someone independently competent when the system has supplied each answer. For moving, loaded or hazardous tasks, the information display must also be considered within the full equipment and supervision arrangement.
Exergaming and physical fitness
Exergaming combines game interaction with physical activity. Reaching, stepping, squatting and moving the arms can require real effort even when the targets are virtual. How much effort depends on the game, settings, movement choices and person.
A 2025 pilot crossover study compared one VR boxing interval protocol with a conventional circuit in 30 healthy medical students. The VR protocol produced higher mean heart rate and perceived exertion; several motivational measures did not differ significantly.[10] These were acute responses. They do not prove greater long-term fitness gains or continued participation.
Judge the activity by the movement it produces
The talk test offers a simple way to describe relative aerobic intensity: during moderate activity, talking is generally possible but singing is difficult; vigorous activity usually permits only a few words before pausing for breath.[11] Individual circumstances can affect its usefulness, and game difficulty levels need not correspond to physiological intensity.
For general health, WHO guidance includes aerobic activity and muscle strengthening, with appropriate balance-focused activity for older adults.[12] An active game may contribute to part of this pattern. Its contribution depends on the actual duration, effort and movement demands.
Virtual objects do not supply physical mass
A bodyweight squat can challenge the body. A graphic of a heavy barbell adds no external mass unless physical equipment supplies resistance. Similarly, waving a controller at virtual targets does not reproduce every demand of punching a bag or controlling contact with another person.
Choose activities that are comfortable enough to repeat and enjoyable enough to return to. Track whether sessions are actually happening and how they fit alongside walking, strength work, sport or daily movement. Calorie estimates and leaderboard positions are optional information; neither is a complete assessment of a person’s health.
Use exercise equipment only in a system designed for that combination. Improvised headset use with heavy free weights, a moving treadmill or added controller weights can change both visibility and physical demands in ways the game was not designed to manage.
Rehabilitation, individual goals and participation
Rehabilitation aims to improve functioning in the person’s own life. WHO emphasises goals, preferences and the interaction between a person and their environment.[13] A virtual reaching task might be useful within that work, but the important outcome could be handling a cup, dressing or taking part in a valued activity.
What the updated stroke evidence says
The 2025 Cochrane review included 190 trials involving 7,188 people after stroke, with searches to September 2023. Compared with alternative therapy, VR may slightly improve arm function and balance and probably improves activity limitation. Adding it to usual care also adds therapy time, which matters when interpreting benefit. Confidence varied by outcome; many trials were small. Very few used immersive headsets, so the findings largely concern other kinds of virtual and interactive systems.[14]
Choose the task with the rehabilitation team
The clinician and participant can identify a relevant movement, suitable level of assistance and a way to assess everyday function. Vision, sensation, understanding, fatigue, pain, balance and available support may affect the choice of system.
For example, an adjustable target could make a reaching task possible from a supported sitting position. The target’s height, distance and pace should follow the agreed goal. A larger score achieved through an unintended strategy might require a task change; a lower score can still accompany a more useful movement.
Remote access may reduce travel for some people, but the practical arrangement still matters: setup help, a suitable space, an accessible way to ask questions and a plan for symptoms or technical failure. A household should not be expected to recreate clinical supervision merely because an app runs at home.
Progress should include what the person reports and what they can do in relevant activities. A comfortable, worthwhile experience can support participation even when a leaderboard would be an inappropriate measure of success.
Professional simulation and supervised competence
Professional training can use simulated situations to rehearse particular decisions, sequences or interactions. The design question is which part of the job can be represented well enough to practise, and how that practice connects to the required real-world assessment.
Illustrative uses include discussing equipment checks, rehearsing communication between team members, identifying an item in a complex workspace or walking through a procedure before using live equipment. Military, emergency-service and industrial settings all contain tasks of this kind; each has its own operational requirements.
A scenario needs a learning purpose
An instructor might pause an exercise after a communication failure and ask which information was missing. Another attempt could focus on stating the request clearly and confirming understanding. The value comes from the opportunity to examine and practise the interaction.
For practical skills, the physical controls and their consequences also matter. A virtual button can teach where a control is located, while handling a real connector may depend on grip, resistance and tactile cues. Those parts may require physical equipment or an appropriate task trainer.
Competence remains tied to the task
Passing a simulation is evidence about performance in that simulation. Decisions about independent work, clinical privileges or certification require the relevant training and assessment process. A historical surgical transfer study provides a model of testing a specified skill in supervised practice, rather than treating simulation scores as universal competence.[8]
Human-factors guidance for medical devices emphasises intended users, uses and environments.[15] Applied as a design principle here, this means testing the actual people, tasks and setting. An expert demonstrator in a quiet room does not reveal every difficulty that a novice, tired worker or person using assistive equipment may encounter.
Comfort, cybersickness and physical safety
A virtual scene can remove a simulated consequence while leaving real physical demands. People still move, tire, lose balance and share space with objects and other people. Planning the room and the activity is part of planning the session.
Make the physical setup workable
Follow the device and application instructions. Use an appropriate clear space, check overhead and floor hazards, manage cables, and keep bystanders and pets outside the movement area. Set up the boundary system correctly, while recognising that a digital boundary cannot guarantee collision prevention. Check fit, visibility and controls before starting.[16]
Recognise cybersickness early
Cybersickness can include nausea, dizziness, eyestrain, headache and disorientation during or after an immersive experience. Research shows that content contributes to symptoms alongside technical characteristics.[17] Visually moving through a scene while the body remains still can be challenging for some people.
A faster display or more expensive headset does not guarantee comfort. Movement style, duration, fit and the individual’s response all deserve attention. The useful outcome is a tolerable experience, not proving that someone can endure discomfort.
Remove the headset, rest somewhere secure and allow symptoms to resolve. Avoid driving or other activities requiring balance and coordination while affected. Persistent or concerning symptoms warrant medical advice. Manufacturer guidance supports gradual familiarisation and regular breaks; no single session length is suitable for everyone.[18]
Keep the physical workload visible
A game may encourage another attempt after movement quality or concentration has declined. Agree in advance that pausing is ordinary. Review discomfort and fatigue alongside the task score, and keep an accessible way to stop the experience.
People with relevant health concerns should use guidance appropriate to their condition and the product. Device and application age guidance also matters when children take part. A headset is an optional tool; difficulty tolerating it is a reason to adapt the activity.
Accessibility, privacy and fair use
Design for the people who will use it
W3C’s XR accessibility work identifies needs such as alternative controls, adaptable interaction, multimodal information and customisation.[19] Useful questions include whether a task can be done seated, whether reach and height can be adjusted, and whether instructions are available through suitable text, sound or captions.
Also consider one-handed use, colour and contrast, pace, motion settings and the effort of putting the device on. Test these features with the intended participants. An option in a settings menu is only useful if people can find and operate it.
Offer another route to the same meaningful learning objective when immersive use is unsuitable. A learner should not be judged less capable merely because the interface requires a movement or sensory ability unrelated to the skill being assessed.
Movement data can be identifying
In a large research dataset, trained models identified VR users from patterns of head and hand motion across recordings.[20] That does not mean every app can instantly identify everyone. It does show why removing a display name may not make movement records anonymous.
Depending on the system, recordings may also contain voice, gaze estimates or information about a room. Before collecting them, explain what is recorded, why it is useful, who receives it and how long it is kept. Consider whether a small set of task outcomes would serve the purpose.
Use the right data safeguards
Where the GDPR applies, identifiable information is personal data. Health data and biometric data processed to uniquely identify someone have additional protections under Article 9; not every motion trace automatically belongs to a special category. Processing requires an appropriate lawful basis and the applicable safeguards, including purpose limitation and data minimisation.[21]
Consent also requires real choice. EDPB guidance explains why employment power imbalances can make freely given consent difficult.[22] Clubs, workplaces and clinics should make clear whether data will affect selection, care or performance review, and provide a way to question an interpretation. An estimated gaze direction or movement score is not a measure of a person’s character, effort or worth.
A five-step framework for adding XR to training
The following framework is an original practical synthesis. It can help a learner, coach or organisation make a modest, reviewable trial. It is not a clinical treatment protocol or a substitute for professional qualification.
-
Choose one meaningful outcome.
Write it in ordinary language: recognise a particular situation, carry out a sequence, improve a real movement task or take part in enjoyable activity more often. Decide how you will recognise a useful change. “Use VR three times” describes attendance; it does not yet describe the benefit.
-
Check the match and establish a baseline.
Try the relevant real task under repeatable conditions, with suitable supervision where needed. Then examine whether the simulation preserves the important information and actions. Check comfort, accessibility, space, data settings and the time needed for setup. Include familiarisation so difficulty operating the interface does not dominate the trial.
-
Practise with a clear purpose.
Choose manageable task difficulty and one main focus. Combine attempts with explanation or feedback that the learner can use. Keep enough real practice to address physical demands that the simulation omits. Record changes to the task and any discomfort, alongside a small number of relevant outcomes.
-
Test what remains and what transfers.
Repeat the agreed assessment after practice, then at a suitable later point. Where independent skill is the goal, reduce prompts. Include an unfamiliar but relevant variation when appropriate. Keep test conditions comparable and record other training that could explain a change.
-
Review the benefit and the burden.
Ask whether the useful outcome improved enough to justify the time, cost and effort. Consider the learner’s experience, assistance needed and alternative ways to practise. Continue, adapt or stop based on that review. Record software or equipment changes before comparing future results.
An example: adding virtual practice to table tennis
Imagine a recreational player who wants more consistent forehand returns. The player and coach first agree on a repeatable real-table drill and note both successful returns and one relevant movement feature. They then try virtual practice as an addition to their existing sessions.
During the trial, they check whether the virtual task encourages useful timing and positioning. They keep some real-table practice because a real ball, racket and opponent provide information the simulation may represent differently. Later, they repeat the agreed drill and try a modest variation.
If only the game score improves, the next question is which demand failed to carry over. If real performance improves, they consider how much may have come from the extra practice overall. Either result can inform the next decision without requiring a claim that one medium is universally best.
Goal · task and settings · practice completed · feedback used · comfort and fatigue · result on the meaningful task · next adjustment.
Future tools, better questions and lasting value
Several directions are worth evaluating: more stable tracking, more useful physical interfaces, accessible controls, shared remote practice and software that adjusts tasks to a learner. Each development should be judged through the problem it solves.
Adaptive difficulty should serve learning
A programme can change speed, target size or task variation in response to performance. The important question is whether the adjustment creates suitable practice. A system optimising session length or a game score may be optimising something different from the learner’s goal.
AI-generated explanations and movement interpretations require evaluation as well. NIST’s AI Risk Management Framework emphasises validity, reliability and the context of use.[23] For training, a practical application of those principles is to ask whether feedback is accurate for the intended users, whether uncertainty is communicated and whether an instructor can correct it.
Shared practice needs more than shared graphics
A remote environment may let participants meet around the same task. Useful design questions include how they communicate, whether timing is adequate, what each person can see and what happens when a connection fails. Recording another person’s session also requires an agreed purpose and suitable data handling.
Choose sustainable tools
For a club, school or clinic, the purchase price is only one part of the commitment. Account for software access, staff time, cleaning, replacement parts, physical space and support. Ask whether useful records can be exported and what happens if the service changes.
For an individual, a shared trial or supervised demonstration can reveal whether the experience suits the goal before a larger commitment. A modest system with a well-designed task may be entirely adequate. A technically impressive system still has to fit the person’s life.
The most promising future is one in which people gain more ways to practise and participate. That includes the freedom to combine virtual tools with ordinary equipment, other people and the physical world they ultimately want to navigate.
Common questions
Can VR really improve a physical skill?
Some task-specific studies have demonstrated improvement on real-world tests. The important details are the skill, system, participants and comparison. Test the outcome you care about; improvement in a familiar virtual task is not sufficient evidence of broad transfer.
Is a more immersive headset always better for learning?
No single level of immersion suits every task. The useful information, actions, physical interface and teaching approach matter. Choose a system that represents the demands of your goal and that you can use comfortably.
What is the difference between AR and mixed reality?
AR adds digital content to a view of the real world. Mixed reality often emphasises spatial anchoring and interaction between digital and physical elements. Usage varies, so ask how the display and tracking actually work.
Can an active VR game count as exercise?
Yes, when it produces physical activity. Its aerobic or strengthening contribution depends on the movements, effort and duration. A game can be an enjoyable part of a broader routine without covering every fitness goal.
Does a vibrating controller reproduce real resistance?
Vibration can signal an event or provide a cue. It does not by itself reproduce the mass, force or impact of a real object. Specialised resistance devices need to be assessed for the particular task.
Can VR replace a coach or rehabilitation professional?
A tool can provide practice and useful information. Coaching and rehabilitation also involve choosing goals, interpreting difficulties, adapting tasks and assessing meaningful progress. The amount of professional support should follow the person and task.
How long should a first session last?
Begin conservatively, follow the product guidance and pause regularly. Stop if symptoms develop. There is no single duration that establishes comfort or suitability for every person and application.
What should I do if I feel dizzy or nauseated?
Stop the experience and rest somewhere secure. Wait until symptoms have resolved before activities requiring balance or coordination. Seek medical advice for persistent or concerning symptoms. You can also use another training format.
What if I need seated use or different controls?
Check the actual application’s options and try them during familiarisation. Adjustable reach, pace and alternative inputs can matter more than a general accessibility label. Ask for another way to pursue the learning objective if the interface is unsuitable.
What is the clearest sign that the training is useful?
A change in the outcome you chose at the beginning: a retained skill, better performance on a relevant task, or worthwhile participation. Consider that change alongside comfort, time, cost and the learner’s experience.
Sources and further reading
Sources checked 6 September 2026. This educational guide draws on original experiments, research reviews, technical documentation and official guidance. These sources serve different purposes: a device manual informs use, a technical specification explains an interface, and a training study tests an intervention.
Selected studies illustrate how transfer can be assessed; they are not an exhaustive comparison of available systems. Older simulator findings remain identified by their actual setup. The practical framework and examples are original illustrations, and no product is endorsed.
Clinical applications need assessment appropriate to the person and condition. Professional competence and certification follow the relevant training requirements. Evidence, software and product instructions can change; use the linked sources to examine the details behind a claim.
- Microsoft. What is mixed reality?. Microsoft Learn. 2023.Technical terminology and the physical–digital spectrum; vendor terminology is not universal. Return to text ↑
- W3C Immersive Web Working Group. WebXR Device API. Technical specification; accessed 6 September 2026.Documents tracked poses, spaces and input in WebXR. An interface specification does not validate a training programme. Return to text ↑
- Makransky G, Terkildsen TS, Mayer RE. Adding immersive virtual reality to a science lab simulation causes more presence but less learning. Learning and Instruction. 2019;60:225–236. doi:10.1016/j.learninstruc.2017.12.007.One educational experiment; its result does not establish that immersive training is generally inferior. Return to text ↑
- Soderstrom NC, Bjork RA. Learning versus performance: an integrative review. Perspectives on Psychological Science. 2015;10(2):176–199. doi:10.1177/1745691615569000.Distinguishes practice performance from durable learning assessed through retention and transfer. Return to text ↑
- Sigrist R, Rauter G, Riener R, Wolf P. Augmented visual, auditory, haptic, and multimodal feedback in motor learning: A review. Psychonomic Bulletin & Review. 2013;20:21–53. doi:10.3758/s13423-012-0333-8.Feedback design depends on the task, learner and modality; it is not a universal prescription for feedback frequency. Return to text ↑
- Michalski SC et al. Getting your game on: Using virtual reality to improve real table tennis skills. PLOS ONE. 2019;14(9):e0222351. doi:10.1371/journal.pone.0222351.57 analysed participants; immersive training versus no training, with real table-tennis assessments. Return to text ↑
- Gray R. Transfer of Training from Virtual to Real Baseball Batting. Frontiers in Psychology. 2017;8:2183. doi:10.3389/fpsyg.2017.02183.80 high-school players; a projected simulator with a tracked real bat. Findings concern this programme and its comparators. Return to text ↑
- Seymour NE et al. Virtual reality training improves operating room performance: results of a randomized, double-blinded study. Annals of Surgery. 2002;236(4):458–463.16 surgical residents; task-specific laparoscopic simulation followed by supervised operating-room assessment. Return to text ↑
- Microsoft. Hologram stability. Microsoft Learn. 2021.Platform-specific engineering guidance illustrating the importance of tracking, calibration and rendering stability. Return to text ↑
- Merola P et al. Virtual Reality High-Intensity Interval Training Exergaming Compared to Traditional High-Intensity Circuit Training Among Medical Students: Pilot Crossover Study. JMIR Serious Games. 2025;13:e63461. doi:10.2196/63461.Acute responses in 30 healthy medical students; does not establish long-term fitness gains or adherence. Return to text ↑
- Centers for Disease Control and Prevention. How to Measure Physical Activity Intensity. Physical Activity Basics. 2024; accessed September 2026.Official guidance on relative intensity, including the talk test. Return to text ↑
- World Health Organization. WHO guidelines on physical activity and sedentary behaviour. 2020 guidelines; current recommendations checked September 2026.Population guidance on physical activity and sedentary behaviour; does not assess particular VR games. Return to text ↑
- World Health Organization. Rehabilitation. Fact sheet; 22 April 2024.Rehabilitation supports functioning and participation and is tailored to the person’s goals, preferences and circumstances. Return to text ↑
- Laver KE, Lange B, George S, Deutsch JE, Saposnik G, Chapman M, Crotty M. Virtual reality for stroke rehabilitation. Cochrane Database of Systematic Reviews. 2025;(6):CD008349. doi:10.1002/14651858.CD008349.pub5.190 trials; searches to September 2023. Most systems were nonimmersive; certainty and effects vary by outcome. Return to text ↑
- US Food and Drug Administration. Applying Human Factors and Usability Engineering to Medical Devices. Guidance issued February 2016.Medical-device guidance on intended users, uses and environments. Used here to explain a design principle, not to classify every fitness product as a medical device. Return to text ↑
- Meta. Health and Safety Warnings: Meta Quest 3. Device guidance; accessed 6 September 2026.Manufacturer guidance on environment, fit, use and discomfort. Consult instructions for the actual device and application. Return to text ↑
- Saredakis D, Szpak A, Birckhead B, Keage HAD, Rizzo A, Loetscher T. Factors Associated With Virtual Reality Sickness in Head-Mounted Displays: A Systematic Review and Meta-Analysis. Frontiers in Human Neuroscience. 2020;14:96. doi:10.3389/fnhum.2020.00096.Examines symptoms and contributing factors; content matters alongside technical characteristics. Return to text ↑
- Apple. Safely use your Apple Vision Pro. User guide; accessed 6 September 2026.Manufacturer guidance on gradual familiarisation, breaks, motion sickness and waiting for symptoms to resolve. Return to text ↑
- W3C Accessible Platform Architectures Working Group. XR Accessibility User Requirements. Working Group Note. 25 August 2021.User needs for multimodal interaction and customisation. This Note is not a product certification or a conformance standard. Return to text ↑
- Nair V et al. Unique Identification of 50,000+ Virtual Reality Users from Head & Hand Motion Data. 32nd USENIX Security Symposium. 2023.Primary identification research using Beat Saber motion recordings; results depend on the dataset and trained model. Return to text ↑
- European Parliament and Council of the European Union. General Data Protection Regulation — Regulation (EU) 2016/679. 27 April 2016; selected provisions checked.Articles 4–9 and related provisions inform the discussion of personal data, processing principles, lawful bases and special categories. Dataset-specific assessment remains necessary. Return to text ↑
- European Data Protection Board. Guidelines 05/2020 on consent under Regulation 2016/679. Adopted 4 May 2020; version 1.1.Official GDPR interpretation, including employment power imbalances and genuine choice. Data-processing consent and consent to an intervention are distinct. Return to text ↑
- National Institute of Standards and Technology. Artificial Intelligence Risk Management Framework (AI RMF 1.0). NIST AI 100-1. 2023.Voluntary cross-sector framework covering trustworthy AI and context of use. It is not evidence that a particular fitness algorithm is effective. Return to text ↑
A virtual world is useful when it opens a real possibility.
That possibility may be a clearer decision, a more confident movement, a procedure practised with care or an activity someone enjoys enough to repeat. Begin with the person and the goal, choose the features that help, and return regularly to the outcome that matters.
Good training gives people more capability and more choice. Virtual and augmented reality can contribute to both when their role is clear and their results are assessed thoughtfully.