StakeBench: Evaluating Language Understanding Grounded in Market Commitment
2026-05-25 • Computation and Language
Computation and LanguageArtificial Intelligence
AI summaryⓘ
The authors created StakeBench, a tool that tests how well language models understand financial language based on real market actions instead of just opinions. They matched over half a million comments from two market platforms with actual trading data to see if models can guess people's market commitments and future moves. They found that while models can somewhat tell which side a person supports, they struggle with predicting future actions and odds changes. Bigger models or special finance training didn’t always help, and market rules affected results. The authors provide StakeBench and its data openly for others to use.
financial NLPmarket commitmentPolymarketManifoldposition sidemarket oddslanguage modelstruthful evaluationcollective odds projectionpredictive behavior
Authors
Yunhua Pei, Jingyu Hu, Yiwei Shi, Hongnan Ma, Weiru Liu, John Cartlidge
Abstract
Existing financial NLP benchmarks often rely on labels supplied by outside observers, measuring how language is perceived rather than what speakers have committed to in the market. We introduce StakeBench, an evaluation framework for language understanding grounded in market commitment. StakeBench links 560,876 comments from 2,261 resolved markets to verified position, action, and market-odds records across Polymarket and Manifold. Supervision is derived from observable market behavior. Position sides, post-comment trading actions, and market-odds trajectories replace human annotation. Four diagnostic tasks test whether models detect market commitment, identify the revealed side, anticipate future action, and perform collective odds projection. Three commitment-aware metrics measure alignment with revealed preferences rather than perceived sentiment. Validity audits and explicit interpretation boundaries help distinguish observable commitment signals from latent belief and causal market-odds impact. Across 15 LLMs and 18 topics and platform settings, models partially recover position-side signals, with Directed Accuracy from 0.506 to 0.599, but show structural failures on later tasks. Ten of the fifteen models collapse to one or two action labels in future action anticipation, and no model consistently improves on the naive odds-direction baseline in collective odds projection. Model scale is not correlated with performance, finance-domain tuning does not improve revealed-side identification, and platform incentives strongly shape higher-order results. StakeBench is packaged with evaluation code and dataset under CC-BY 4.0.