# CoToan: Executable Rules and a Representation Contract for Reproducible Game AI Evaluation

DevFeed: [CoToan: Executable Rules and a Representation Contract for Reproducible Game AI Evaluation](<https://devfeed.tech/articles/cotoan-executable-rules-and-a-representation-contract-for-reproducible-game-ai-evaluation-108178.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.39353>)

Author: Ha-Thanh Nguyen

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Algorithms](<https://devfeed.tech/topics/algorithms.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>)

Tags: [actions](<https://devfeed.tech/tags/actions.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-evaluation](<https://devfeed.tech/tags/ai-evaluation.md>), [arithmetic](<https://devfeed.tech/tags/arithmetic.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [capture](<https://devfeed.tech/tags/capture.md>), [chess](<https://devfeed.tech/tags/chess.md>), [configuration](<https://devfeed.tech/tags/configuration.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [diagnostics](<https://devfeed.tech/tags/diagnostics.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [exhaustive](<https://devfeed.tech/tags/exhaustive.md>), [experimental](<https://devfeed.tech/tags/experimental.md>), [game](<https://devfeed.tech/tags/game.md>), [horizon](<https://devfeed.tech/tags/horizon.md>), [identity](<https://devfeed.tech/tags/identity.md>), [implementation](<https://devfeed.tech/tags/implementation.md>), [language](<https://devfeed.tech/tags/language.md>), [legal](<https://devfeed.tech/tags/legal.md>), [masks](<https://devfeed.tech/tags/masks.md>), [rules](<https://devfeed.tech/tags/rules.md>)

## AI overview

CoToan links Vietnamese Mathematical Chess rules in GDL and OpenSpiel through an executable representation contract. It defines actions by the resulting state transition rather than by arithmetic witnesses, and includes remaining time and victory settings in task identity. Cross-representation tests found agreement on legal actions, successor states, and terminal outcomes across sampled and constructed positions, alongside exhaustive checks of 32 operator masks and 512 opposing-material subsets. Diagnostics found that uniform sampling of witnesses changes endpoint-action probabilities, and omitting time or victory settings can merge states with different terminal behavior. The abstract's evidence supports treating these representation choices as part of the evaluation task for reproducible game AI.

## Source excerpt

arXiv:2610.39353v1 Announce Type: new Abstract: Game AI results are comparable only when different implementations expose the same decision problem. Yet seemingly minor representation choices can silently change that problem by altering action probabilities or collapsing states with different outcomes. We study this issue through CoToan, an executable specification and representation contract for Vietnamese Mathematical Chess connecting the Game Description Language (GDL) with OpenSpiel. The game provides a useful case because a single capture may have multiple arithmetic justifications, while outcomes can depend on scoring rules and a finite playing horizon. CoToan therefore represents actions by their state transitions rather than by arithmetic witnesses, and treats time and victory configuration as part of task identity. Cross-representation tests show agreement in legal actions, successor states, and terminal outcomes across hundreds of sampled and constructed positions, complemented by exhaustive checks of all 32 operator masks and 512 opposing-material subsets. Controlled diagnostics further show that uniform witness-level sampling alters endpoint-action probabilities, while omitting remaining time or victory settings can merge states with different terminal behavior. These results show that action semantics, time, and victory configuration are not merely implementation details. They are part of the experimental task itself and must be represented explicitly for reproducible game AI.