← CHI 2026 map

CHI '26 · Honorable mention · full-paper review · confidence medium-high

Show Me How to Play: Exploring Self-Modeling for Onboarding in Virtual Reality Exergames

Sukran Karaosmanoglu , Silas Ueberschaer , Sebastian Cmentowski , Frank Steinicke

DOI PDF Program page Page JSON

This is a well-scoped CHI paper with a clear interaction contribution: it turns onboarding into self-observation by having the player watch their own avatar perform the movement. The idea is counterintuitive but grounded, and the study design is strong enough to support the main performance and experience claims without overreaching.

Axes Lens

Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.

Review Summary

This paper’s strongest contribution is conceptual and interactional rather than theoretical: it takes a familiar learning principle, observational learning, and re-implements it in VR exergame onboarding through a self-model or ghost-avatar tutorial. The design choice to temporarily reduce agency is the key departure from common onboarding intuition, because the tutorial does not merely explain or scaffold while preserving control; it intentionally suspends control so the player can watch their own embodied avatar execute the movement. That makes the technique memorable and plausibly well matched to VR’s embodied affordances. The empirical support is also solid for the scope claimed. A between-participants study with 60 participants across GHOST, T&E, and BASE is an appropriate validation mode for comparing short-term onboarding effects, and the paper reports performance and subjective benefits such as ease of control and progress feedback. At the same time, the paper is careful enough in its limitations to avoid implying broad generality: it acknowledges that the evidence comes from short-term gameplay, that retention and long-term transfer were not measured, and that safety outcomes were not assessed. So the paper reads as a strong CHI contribution because it combines a novel interaction technique with a focused controlled evaluation, but its claims should be read as task-specific and immediate-outcome-specific rather than as a general solution for all VR onboarding problems.

What Changed

Canon before

Prior CHI work on onboarding and tutorials in VR exergames typically relies on direct instruction, trial-and-error, or demonstration by an external model; this paper shifts the tutorial target to the player’s own avatar and uses temporary loss of agency as part of the instructional mechanism.

Departure from common sense

The onboarding design deliberately reduces the player’s control at the start: the avatar’s arms become semi-transparent, the player experiences a loss of agency, and then the avatar performs the tutorial movements. That is counterintuitive relative to common onboarding assumptions that tutorials should preserve control while adding guidance.

Actual novelty

The paper’s novelty is the application of a self-model concept as an instructional tutorial for VR exergames, specifically using the player’s own avatar as the demonstrator and temporarily removing agency via a ghost metaphor. The contribution is not just a new tutorial variant, but a new way to structure observational learning in embodied VR onboarding.

Evidence

The paper grounds its claims in a between-participants study with 60 participants split evenly across three conditions, comparing the self-model tutorial (GHOST), trial-and-error tutorial (T&E), and a baseline without tutorial (BASE). The study evaluates short-term performance and self-report outcomes, including punch accuracy, ease of control, progress feedback, and embodiment-related measures. The limitations section explicitly notes that results are based on short-term gameplay and do not assess retention, long-term transfer, or safety outcomes.

Limits

Method limits

The evidence is limited to a single between-participants lab study with N=60 and short-term gameplay outcomes. The paper does not establish long-term learning retention, transfer, or safety effects, and the measured outcomes are tied to the specific task and tutorial implementation used in the study.

Deployment limits

The approach is demonstrated for a specific VR exergame onboarding scenario and should not be assumed to generalize to all exergames, all VR tasks, or contexts where temporarily removing agency could be confusing, uncomfortable, or unsafe.

Boundary conditions

The reported benefits are bounded by the study’s short duration, the specific boxing-style movement task, and the tutorial’s reliance on embodied observation of the player’s own avatar. Generalization beyond this task class and beyond immediate performance/ease-of-use outcomes remains untested.

Manual check: none

Position in field

This sits at the intersection of VR onboarding, exergame tutorial design, and observational learning. Its main field contribution is to reframe tutorial design around self-modeling rather than external demonstration, offering a concrete interaction technique with empirical support.

Abstract

Exergames combine motivating game elements with bodily movement to encourage physical activity. However, onboarding players to perform correct movements remains a challenge, especially in virtual reality (VR) environments where safety and performance are critical. Drawing inspiration from sports training and learning sciences, we contrast two onboarding approaches: (i) trial-and-error and (ii) observational learning via a novel self-model tutorial. In this tutorial, players temporarily lose agency and observe their own avatar performing the movements, leveraging VR’s unique affordances for embodied experiences. To explore which of these two approaches yields a better performance and player experience, we conducted a between-participants study (N=60), comparing them against a baseline condition without a tutorial. Our findings show that the self-model tutorial not only improves players' performance but also increases the perceived ease of control and progress feedback. We discuss tradeoffs and implications for the design of future onboarding experiences in VR exergames.