arXiv Science⌕ Search

arXiv · 2610.26447

Comparing Experiences of Speech-Driven Game Modification in Solo Play, Paired Play, and Paired Watching Using the Fighting Game PATCH FIGHTER

Abstract

Players often have ideas about a game while playing it, but turning them into changes usually means stepping out of play to edit it. We built PATCH FIGHTER, a two-player fighting game defined by 91 types of adjustable parameters. A large language model (LLM) reads players' conversations during play and rewrites these parameters between games, so players can change the game just by talking. We conducted a within-subjects experiment with 27 pairs (54 participants) comparing three conditions, solo play, paired play, and paired watching. Participants rated enjoyment after each game and answered a questionnaire on speech and changes after each condition. Paired play was the most enjoyable condition. In paired play, participants also felt more able to say what they thought about the game than in solo play. In interviews, participants valued talking with the partner and checking changes through their own play. Paired watching was less preferred because they could not try the changes themselves. As our design cannot separate the enjoyment of editing from that of play, we outline a design that measures them separately and the evaluation of the resulting games as future work.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Maya Grace Torii, Takahito Murakami, Yoichi Ochiai. 2026-10-19. Comparing Experiences of Speech-Driven Game Modification in Solo Play, Paired Play, and Paired Watching Using the Fighting Game PATCH FIGHTER. https://arxiv.org/abs/2610.26447

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

FlexMind: Supporting Deeper Creative Thinking with LLMs

Expert ideation requires more than generating ideas. People must articulate trade-offs, develop mitigations, and compare parallel directions, yet this reasoning process is cognitively demanding. Existing ideation tools fall short by either focusing only on expanding or organizing the idea space, or by offering evaluations without supporting mitigation generation and continued reasoning over newly introduced trade-offs. We present FlexMind, an AI-augmented system that proactively scaffolds multi-threaded trade-off-mitigation chains within a structured workspace. FlexMind helps users surface concrete trade-offs, generate actionable mitigations, and navigate and expand multiple idea branches. In a controlled study (n=20), FlexMind supported deeper exploration without sacrificing breadth and produced higher-quality ideas, as judged by domain experts, compared to a baseline. A case study with six experts (n=6) further reveals how FlexMind supports idea analysis while helping uncover new solutions. These results inform implications for enhancing human-AI co-creation.

cs.HC↗

ZORO: Active Controls to Support Rule Following in Vibe Coding

Rules files (e.g., AGENTS$.$md, CLAUDE$.$md) are the primary mechanism for human-agent alignment when developers vibe code. However, they remain passive: it is not immediately apparent when rules are being used or followed, or how to improve them. To transform rules from passive text into active controls, we introduce ZORO, an interactive interface that integrates directly with a coding agent and anchors rules to every step of the coding process. After an agent generates an initial plan, ZORO enriches the plan with rules, enforces the rules during implementation by requiring the agent prove that each rule was followed, and allows users to provide in-situ feedback when they are unsatisfied with a rule application to evolve the ruleset. A technical evaluation shows that coding agents follow rules more with ZORO than without. A user study demonstrates a change in people's behavior and cognitive strategies when rules are at the forefront of vibe coding. We discuss how making rules active in agentic systems unlocks broader opportunities for human-agent alignment in coding settings and beyond.

cs.HC↗

Calibrating LLM Judges for Human and AI Conversations

Measuring how successful a conversation is remains difficult, even for humans judging spoken dialogue. We evaluate state-of-the-art LLMs as pointwise and pairwise judges of conversational success on CANDOR, finding pointwise scoring correlates moderately with human ratings, while pairwise comparison suffers from long transcripts and positional bias. Since this leaves judge scores incomparable across models, we propose a small anchor set and a calibration function that calibrates any judge onto a shared, interpretable scale. We further release the Voice Arena Goal Dataset (VA), 200 task-oriented human-AI and human-agent conversations with pairwise annotations, revealing a substantial gap between current judges and human-level discrimination. Using VA, we test whether CANDOR-fitted calibration transfers to human-AI conversations, finding it brings judges onto a shared scale despite never observing VA during fitting.

cs.HC↗