# Demystifying Multi-Agent Debate: The Role of Confidence and Diversity

arXiv: https://arxiv.org/abs/2601.19921 · HTML: https://mlsanity.com/paper/2601.19921/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- Builds on 28 of its 51 references that are in the map.
- First posted on arXiv 9 Jan 2026.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)
- LLM agents and tool use / [Social simulation and multi-agent debate](https://mlsanity.com/agents/social-simulation-and-multi-agent-debate/)

## Builds on (most cited first, 28 of 28)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [The Curious Case of Neural Text Degeneration](https://mlsanity.com/paper/1904.09751/) (2019)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge](https://mlsanity.com/paper/1811.00937/) (2018)
- [Improving Factuality and Reasoning in Language Models through Multiagent Debate](https://mlsanity.com/paper/2305.14325/) (2023)
- [Large Language Models are not Fair Evaluators](https://mlsanity.com/paper/2305.17926/) (2023)
- [Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate](https://mlsanity.com/paper/2305.19118/) (2023)
- [Large Language Models Can Self-Improve](https://mlsanity.com/paper/2210.11610/) (2022)
- [Understanding the Effects of RLHF on LLM Generalisation and Diversity](https://mlsanity.com/paper/2310.06452/) (2023)
- [LUQ: Long-text Uncertainty Quantification for LLMs](https://mlsanity.com/paper/2403.20279/) (2024)
- [Debate or Vote: Which Yields Better Decisions in Multi-Agent Large Language Models?](https://mlsanity.com/paper/2508.17536/) (2025)
- [Rewarding the Unlikely: Lifting GRPO Beyond Distribution Sharpening](https://mlsanity.com/paper/2506.02355/) (2025)
- [Talk Isn't Always Cheap: Understanding Failure Modes in Multi-Agent Debate](https://mlsanity.com/paper/2509.05396/) (2025)
- [Rewarding Doubt: A Reinforcement Learning Approach to Calibrated Confidence Expression of Large Language Models](https://mlsanity.com/paper/2503.02623/) (2025)
- [The Majority is not always right: RL training for solution aggregation](https://mlsanity.com/paper/2509.06870/) (2025)
- [Can LLM Agents Really Debate? A Controlled Study of Multi-Agent Debate in Logical Reasoning](https://mlsanity.com/paper/2511.07784/) (2025)
- [Atomic Calibration of LLMs in Long-Form Generations](https://mlsanity.com/paper/2410.13246/) (2024)
- [Revisiting Multi-Agent Debate as Test-Time Scaling: A Systematic Study of Conditional Effectiveness](https://mlsanity.com/paper/2505.22960/) (2025)
- [LoGU: Long-form Generation with Uncertainty Expressions](https://mlsanity.com/paper/2410.14309/) (2024)
- [Enhancing Multi-Agent Debate System Performance via Confidence Expression](https://mlsanity.com/paper/2509.14034/) (2025)
- [LoVeC: Reinforcement Learning for Better Verbalized Confidence in Long-Form Generations](https://mlsanity.com/paper/2505.23912/) (2025)
- [DebUnc: Improving Large Language Model Agent Communication With Uncertainty Metrics](https://mlsanity.com/paper/2407.06426/) (2024)
- [Confidence Estimation for LLMs in Multi-turn Interactions](https://mlsanity.com/paper/2601.02179/) (2026)
- [UNCLE: Benchmarking Uncertainty Expressions in Long-Form Generation](https://mlsanity.com/paper/2505.16922/) (2025)
- [Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging](https://mlsanity.com/paper/2510.17426/) (2025)

## Built on by (most cited first, 18 of 18)

- [Closing the Loop to Discover Psychological Theories with an Automated Cognitive Scientist](https://mlsanity.com/paper/2606.26448/) (2026)
- [Epistemic Gain, Aleatoric Cost: Uncertainty Decomposition in Multi-Agent Debate for Math Reasoning](https://mlsanity.com/paper/2603.01221/) (2026)
- [Emergent Risks in Generative Multi-Agent Systems](https://mlsanity.com/paper/2603.27771/) (2026)
- [Hear Both Sides: Efficient Multi-Agent Debate via Diversity-Aware Message Retention](https://mlsanity.com/paper/2603.20640/) (2026)
- [The Reasoning Trap: An Information-Theoretic Bound on Closed-System Multi-Step LLM Reasoning](https://mlsanity.com/paper/2605.01704/) (2026)
- [A Systematic Evaluation of Black-Box Uncertainty Estimation Methods for Large Language Models](https://mlsanity.com/paper/2606.19868/) (2026)
- [To Isolate or to Score? Model-Adaptive Assessment for Cost-Efficient Multi-Agent RAG](https://mlsanity.com/paper/2606.25191/) (2026)
- [One Axis, No Brake: Self-Knowledge Limits the Filtering of Harmful Peer Conformity in LLMs](https://mlsanity.com/paper/2609.18998/) (2026)
- [Self-Organizing Agent Teams Learn to Reason Together](https://mlsanity.com/paper/2609.22682/) (2026)
- [The impact of multi-agent debate protocols on debate quality: a controlled case study](https://mlsanity.com/paper/2603.28813/) (2026)
- [Group Perspective Matters: Regulating Debate Relationships Can Mitigate Blind Conformity in Multi-Agent Debate](https://mlsanity.com/paper/2608.03648/) (2026)
- [A Theory of Post-hoc Debate Judgement](https://mlsanity.com/paper/2608.19002/) (2026)
- [Unanimity Without Persuasion: A Single Round of Debate Erases the Disagreement That Verification Needs](https://mlsanity.com/paper/2609.26145/) (2026)
- [Do We Need Complex Topology Control? Distinct-Peer Random Routing Improves Cost-Efficiency in Sparse Multi-Agent Debate](https://mlsanity.com/paper/2609.27150/) (2026)
- [Black-Box Auditing of Epistemic Reliability in Multi-Agent Debate Distillation](https://mlsanity.com/paper/2609.32361/) (2026)
- [Measuring Collapse and Correction in Homogeneous-Panel LLM Debate](https://mlsanity.com/paper/2609.35279/) (2026)
- [It Takes Workflows to Evolve Better Workflows](https://mlsanity.com/paper/2610.01026/) (2026)
- [Right Answers, Wrong States: Hidden Information Failures in Multi-Agent Collaboration](https://mlsanity.com/paper/2610.01244/) (2026)

## Abstract (arXiv)

Multi-agent debate (MAD) is widely used to improve large language model (LLM) performance through test-time scaling, yet recent work shows that vanilla MAD often underperforms simple majority vote despite higher computational cost. Studies show that, under homogeneous agents and uniform belief updates, debate preserves expected correctness and therefore cannot reliably improve outcomes. Drawing on findings from human deliberation and collective decision-making, we identify two key mechanisms missing from vanilla MAD: (i) diversity of initial viewpoints and (ii) explicit, calibrated confidence communication. We propose two lightweight interventions. First, a diversity-aware initialisation that selects a more diverse pool of candidate answers, increasing the likelihood that a correct hypothesis is present at the start of debate. Second, a confidence-modulated debate protocol in which agents express calibrated confidence and condition their updates on others' confidence. We show theoretically that diversity-aware initialisation improves the prior probability of MAD success without changing the underlying update dynamics, while confidence-modulated updates enable debate to systematically drift to the correct hypothesis. Empirically, across six reasoning-oriented QA benchmarks, our methods consistently outperform vanilla MAD and majority vote. Our results connect human deliberation with LLM-based debate and demonstrate that simple, principled modifications can substantially enhance debate effectiveness.