Analysis of Uncertainty in Procedural Maps in Slay the Spire

Mahsa Bazzaz, Seth Cooper · 2025

Read Analysis of Uncertainty in Procedural Maps in Slay the Spire by Mahsa Bazzaz, Seth Cooper (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-03918.txt.

Introduction

Mahsa Bazzaz and Seth Cooper analyze in 2025 how randomness shapes play in Slay the Spire, the deck-building video game whose maps are generated fresh each run. Players choose a path through a branching map, and chance governs what each branch holds.

Drawing on twenty thousand recorded runs, the authors measured the entropy, the information-theoretic spread, of the paths players took. Winning runs carried higher normalized entropy, a sign of greater risk-taking. Skilled players also showed their own late-game pattern of risk. The study gives game designers hard numbers on how dealt uncertainty drives human choices.

Catalog record

CREATORMahsa Bazzaz, Seth Cooper
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE35 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-03918

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Analysis of Uncertainty in Procedural Maps in Slay the Spire Mahsa Bazzaz, Seth Cooper Abstract . This work investigates the role of uncertainty in Slay the Spire using an information-theoretic framework. Focusing on the entropy of game paths (which are based on procedurally-generated maps) we analyze how randomness influences player decision-making and success. By examining a dataset of 20,000 game runs, we quantify the entropy of paths taken by players and relate it with their outcomes and skill levels. The results show that victorious runs are associated with higher normalized entropy, suggesting more risk-taking. Additionally, higher-skill players tend to exhibit distinct patterns of risk-taking behavior in later game stages. Conference: International Conference on the Foundations of Digital Games; April 15–18, 2025; Graz, Austria International Conference on the Foundations of Digital Games (FDG ’25), April 15–18, 2025, Graz, Austria DOI: 10.1145/3723498.3723846 ISBN: /25/04 † † cc-license: by 1. Introduction Slay the Spire ( Crit, 2019 ) is a strategic deck-building game where players navigate a procedurally generated ( Togelius et al., 2011 ) path, engaging in battles using a customizable deck of cards. Players select one of several unique characters, each with their own abilities and card sets. Success in the game depends on how well players manage their resources, plan their moves, and adapt to evolving challenges. The ultimate goal is to defeat powerful bosses and, eventually, reach the final boss. Each attempt at the game, known as a “run”, begins when the player starts the game and ends either in victory or failure. If the player fails, they must begin a new run from the start. This is because this game is a roguelike game, meaning that the game is designed with the expectation that players will fail and restart multiple times. This iterative process allows players to enhance their skills, deepen their understanding of the game, and unlock new content, such as additional cards, characters, and relics. These improvements and unlocked elements increase the likelihood of success in subsequent runs. Randomness plays a significant role in digital games, ranging from simple mechanics like rolling dice to more complex behaviors in AI systems. These elements, though often outside the player’s control, have a substantial impact on gameplay experience and outcomes. Previous studies have explored the effects of randomness in Slay the Spire, particularly in relation to cards and relics, showing that higher-skilled players manage these random elements better than lower-skilled players. However, players across all skill levels are still influenced by the inherent randomness of the cards and relics ( Andersson et al., 2023 ; Lantz et al., 2017 ) . Despite this, the randomness of an entire “run” in Slay the Spire has not been thoroughly examined. This is particularly important because players are presented with multiple paths, each offering different map locations and varying levels of uncertainty. This aspect is especially significant in Slay the Spire, as its procedurally generated maps introduce numerous possible layouts and choices — far more than would be found in a game with manually designed maps. Choosing a specific path allows players to adjust the game’s difficulty by encountering easier rooms or taking advantage of resources like money, campfires for healing, or strengthening to face elite enemies. Figure 1 shows a sample path selection in Slay the Spire. The paths are shown vertically in the original game. Information theory has been previously applied to game research to understand difficulty ( Chen et al., 2023 ) and depth ( Lantz et al., 2017 ) of games, and the evolution of the game’s system over time ( Zuparic et al., 2021 ) . However, no studies to date have analyzed the gameplay of Slay the Spire using information theory. This work aims to fill that gap by examining how randomness and entropy influence player behavior and outcomes in Slay the Spire. By applying information theory, we seek to understand how players confront and manage randomness throughout a run and how these elements related with their success. Our research questions are: RQ1 : Which runs (ended with Victory or Defeat) involve more uncertain paths? RQ2 : Which type of player (lower-level or higher-level) tends to choose more uncertain paths? Our contributions are summarized as follows: • We define the concepts of “Uncertainty” and “Risk-Taking” within the context of the game Slay the Spire, and propose an information-theoretic approach to quantify these concepts. • We analyze the uncertainty associated with all potential paths available to players using a comprehensive set of player logs. • We investigate the relationship between path uncertainty and player success in game runs, assessing whether risk-taking strategies provide an advantage in Slay the Spire. • We explore the relationship between path uncertainty and player ascension level in runs, identifying which types of players are more inclined towards risk-taking strategies. Figure 1 . Screenshot of the game showcasing the path selection in Slay the Spire. 2. Related Work 2.1. Information Theory Information theory has become an instrumental tool in understanding the dynamics, complexity, and engagement levels within games. In strategic and puzzle-based games, it provides a framework to quantify uncertainty, player choices, and the structural characteristics that define a game’s complexity and difficulty. Lantz et al. (2017) examine the concept of “depth” in strategic games through information theory, aiming to define depth as a measurable property reflecting a game’s capacity for prolonged engagement and continuous learning. They propose a metric, depth, corresponding to a game’s susceptibility to computational exploration, effectively capturing the level of challenge and variation a game presents as players encounter new strategies. This depth is often associated with games that foster enduring interest, as they offer continuous layers of strategy that emerge over time. Chen et al. (2023) apply entropy to single-player puzzle games, focusing on how uncertainty in solving puzzles relates to difficulty. They introduce entropy-based algorithms to measure a puzzle’s uncertainty, which corresponds to the number of choices or steps required for solution discovery. By calculating entropy scores for each puzzle state, they assess the difficulty level based on the choices available to players at each step, ultimately creating a metric that predicts player challenge and engagement with different puzzles. Zuparic et al. (2021) leverage Shannon entropy and information storage concepts in the analysis of archetypes in the collectible card game Hearthstone. By measuring the entropy of player choices over time, they assess the game’s adaptability and complexity as players experiment with new strategies. Their study shows that higher entropy values correlate with phases of player exploration, particularly during periods of significant game updates or shifts in available strategies. This dynamic analysis provides insights into the game’s “meta” and its evolution in response to player behavior and system changes. More recently, Shen and Sturtevant (2024) expanded on their prior work ( Chen et al., 2023 ) , by introducing generalized entropy and solution information to assess puzzle difficulty. They propose novel metrics, Minimum Solution Information (MSI) and Total Solution Information (TSI), to quantify the information required for players to solve puzzles, considering their knowledge and policies. These methods enhance earlier work by providing a mathematically grounded approach for evaluating player engagement and difficulty in puzzle games such as The Witness. Their findings highlight how entropy-based measures can predict

Read the complete text