# Confident at the moment of action: belief miscalibration in LLM play under hidden information

arXiv: https://arxiv.org/abs/2608.24691 · HTML: https://mlsanity.com/paper/2608.24691/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Reinforcement learning for LLM agents, one of 16 directions in LLM agents and tool use.
- Builds on 9 of its 12 references that are in the map.
- First posted on arXiv 25 Aug 2026.

## Placed in

- LLM agents and tool use / [Reinforcement learning for LLM agents](https://mlsanity.com/agents/reinforcement-learning-for-llm-agents/)

## Builds on (most cited first, 9 of 9)

- [Language Models (Mostly) Know What They Know](https://mlsanity.com/paper/2207.05221/) (2022)
- [Taming Overconfidence in LLMs: Reward Calibration in RLHF](https://mlsanity.com/paper/2410.09724/) (2024)
- [ConfTuner: Training Large Language Models to Express Their Confidence Verbally](https://mlsanity.com/paper/2508.18847/) (2025)
- [Calibrating Verbalized Confidence with Self-Generated Distractors](https://mlsanity.com/paper/2509.25532/) (2025)
- [WOLF: Werewolf-based Observations for LLM Deception and Falsehoods](https://mlsanity.com/paper/2512.09187/) (2025)
- [Agent-BRACE: Decoupling Beliefs from Actions in Long-Horizon Tasks via Verbalized State Uncertainty](https://mlsanity.com/paper/2605.11436/) (2026)
- [BayesBench: Evaluating LLM Belief Trajectories Under Multi-Turn Evidence Accumulation](https://mlsanity.com/paper/2606.30850/) (2026)
- [Same Answer, Different Confidence: Protocol Sensitivity in LLM Confidence Calibration](https://mlsanity.com/paper/2605.27752/) (2026)
- [ConfidenceBench: Evaluating Confidence Calibration in Large Language Models](https://mlsanity.com/paper/2607.20526/) (2026)

## Built on by (most cited first, 1 of 1)

- [Replication Without Persistence in Hosted LLMs: Measurement Sensitivity in Action-Time Belief Evaluation](https://mlsanity.com/paper/2609.22478/) (2026)

## Abstract (arXiv)

Agentic systems increasingly gate actions on a model's own stated confidence, which assumes confidence tracks correctness at the moment of acting. We test this in a hidden-information chess variant where royal status can be secretly, repeatedly relocated between pieces, and where an agent's stated probability distribution over the opponent's hidden royal piece -- elicited every turn, separately from the move it chooses -- is scored against ground truth recoverable after the game. Across two independent batches, captures made at high stated confidence (≥ 0.5) about the hidden piece's location were correct in 1 of 62 cases. The calibration deficit is concentrated almost entirely in these events: 99.3% of it in the original batch, 98.7% in the replication. The same pattern, in weaker form, orders consistently (point estimates only; most pairwise gaps are not statistically distinguishable at this sample size) across four further model configurations spanning a second provider -- reported as scope for the finding, not as evidence that capability predicts calibration: a same-model comparison at a fixed external leaderboard score shows a deliberation-budget change alone moves the metric by nearly as much as a large cross-model gap. In a separate seat, conventional evaluation axes -- legality, cost, latency, completion rate -- can dissociate entirely from belief quality, with the configuration winning on every conventional axis producing the worst belief quality tested. A model exhibiting this pattern can still win the game its belief was about, which is why outcome-only evaluation would not detect it.