[Paper Review] Strange Beta: An Assistance System for Indoor Rock Climbing Route Setting Using Chaotic Variations and Machine Learning

Caleb Phillips, Lee A. Becker+1 more|arXiv (Cornell University)|Oct 3, 2011

Music and Audio Processing27 references3 citations

TL;DR

This paper introduces Strange Beta, an AI-assistance system that uses chaotic variation and machine learning to generate novel, stylistically consistent indoor rock climbing routes. By modeling climbing movements as symbolic sequences and applying Variable Order Markov Models (VOMMs) with chaotic perturbations, the system produces variations that expert climbers rate as equal or superior to traditionally set routes in a blinded study.

ABSTRACT

This paper applies machine learning and the mathematics of chaos to the task of designing indoor rock-climbing routes. Chaotic variation has been used to great advantage on music and dance, but the challenges here are quite different, beginning with the representation. We present a formalized system for transcribing rock climbing problems, then describe a variation generator that is designed to support human route-setters in designing new and interesting climbing problems. This variation generator, termed Strange Beta, combines chaos and machine learning, using the former to introduce novelty and the latter to smooth transitions in a manner that is consistent with the style of the climbs This entails parsing the domain-specific natural language that rock climbers use to describe routes and movement and then learning the patterns in the results. We validated this approach with a pilot study in a small university rock climbing gym, followed by a large blinded study in a commercial climbing gym, in cooperation with experienced climbers and expert route setters. The results show that {\sc Strange Beta} can help a human setter produce routes that are at least as good as, and in some cases better than, those produced in the traditional manner.

Motivation & Objective

• To develop a computer-assisted system that supports human route setters in creating novel, engaging indoor climbing problems.
• To formalize the representation of climbing routes using a domain-specific grammar and natural language parsing.
• To apply chaotic dynamics to generate stylistically coherent variations that maintain route integrity while introducing novelty.
• To validate the system’s effectiveness through controlled studies with expert climbers and route setters in real climbing environments.

Proposed method

• The system uses a formal grammar, Phoenix, to represent climbing routes as symbolic sequences of actions, holds, and movement types.
• It parses natural language descriptions of climbs into structured symbolic representations using a context-free grammar and linguistic rules.
• A Variable Order Markov Model (VOMM) is trained on a corpus of existing climbs to learn stylistic patterns and smooth transitions in generated sequences.
• Chaotic variation is introduced via perturbations of initial conditions in a strange attractor (e.g., Lorenz system), generating new routes that deviate meaningfully but stylistically from the original.
• The variation generator uses nearest-neighbor mapping from chaotic trajectory coordinates to symbolic sequences to ensure stylistic consistency.
• A user interface allows human setters to explore and refine generated route variations interactively.

Experimental results

Research questions

• RQ1 Can chaotic variation combined with machine learning generate climbing route variations that are perceived as more interesting or novel by experienced climbers?
• RQ2 To what extent can a machine learning model trained on real climbing data produce sequences that maintain stylistic coherence and avoid awkward transitions?
• RQ3 Does the use of a formal symbolic representation improve the reliability and consistency of route generation compared to freeform input?
• RQ4 Can computer-aided route setting produce climbs that are preferred over traditionally set routes in a blinded, controlled study?

Key findings

• In a blinded study at a commercial climbing gym, climbers preferred routes generated by Strange Beta over traditionally set routes, with a mean preference score of 0.638 on a 1–1 scale.
• The system’s variation generator produced routes with improved stylistic consistency, as evidenced by low Kendall’s τ correlation (0.173) for awkward transitions in the generated sequences.
• The VOMM model successfully smoothed transitions in generated routes, reducing awkwardness compared to raw chaotic perturbations.
• The parsing framework correctly mapped 90% of natural language climb descriptions into the formal Phoenix grammar, enabling reliable symbolic representation.
• The internal consistency of the evaluation instrument (Cronbach’s α = 0.708) confirmed reliable measurement of climber preferences across 14 Likert-scale items.
• No significant correlation was found between ice cream preference and climbing ability, validating the lack of bias in the study’s demographic data.