Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2410.12464

Exploring LLM Cryptocurrency Trading Through Fact-Subjectivity Aware Reasoning

Qian Wang·Yuchen Gao·Zhenheng Tang·Bingqiao Luo·Nuo Chen·Bingsheng He
View original source ↗Read source document ↗Export metadata ↓

Abstract

While many studies show that more advanced LLMs excel in tasks such as mathematics and coding, we observe that in cryptocurrency trading, stronger LLMs sometimes underperform compared to weaker ones. To investigate this counterintuitive phenomenon, we examine how LLMs reason when making trading decisions. Our findings reveal that (1) stronger LLMs show a preference for factual information over subjectivity; (2) separating the reasoning process into factual and subjective components leads to higher profits. Building on these insights, we propose a multi-agent framework, FS-ReasoningAgent, which enables LLMs to recognize and learn from both factual and subjective reasoning. Extensive experiments demonstrate that this fine-grained reasoning approach enhances LLM trading performance in cryptocurrency markets, yielding profit improvements of 7\% in BTC, 2\% in ETH, and 10\% in SOL. Additionally, an ablation study reveals that relying on subjective news generates higher returns in bull markets, while focusing on factual information yields better results in bear markets. Code is available at https://github.com/Persdre/FS-ReasoningAgent.

Explore related subjects

cs.MA

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Qian Wang, Yuchen Gao, Zhenheng Tang, Bingqiao Luo, Nuo Chen, Bingsheng He. 2025-03-12. Exploring LLM Cryptocurrency Trading Through Fact-Subjectivity Aware Reasoning. https://arxiv.org/abs/2410.12464

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Highway Congestion Reduction through Reinforcement Learning Based Eulerian Headway Control

Connected automated vehicles (CAVs) equipped with adaptive cruise control (ACC) create new opportunities for highway congestion mitigation. Traditional practice relies on Eulerian variable speed limits (VSL) which regulate traffic through roadside signs, but suffer from infrequent updates and limited driver compliance. Recent research explored Lagrangian strategies that directly control individual vehicles, offering high reactivity and compliance, yet in realistic multi-lane settings they depend on drivers' latent lane-change intentions, making robust vehicle-level decisions difficult. Hence, we propose an Eulerian control system optimized through reinforcement learning, that (i) leverages ACC for reactivity and compliance, and (ii) obviates dependence on latent driver intentions by regulating aggregate density near bottlenecks, crucially via headway commands rather than speed commands. We evaluate two variants of our system, time-headway and distance-headway control, in large-scale simulations across a range of traffic conditions. Both variants outperform baselines, improving traffic flow by up to 10.6% over human traffic and 6.7% over traditional VSL. To strengthen evaluation, we propose a novel boundary-aware speed metric addressing a recognized flaw in simulation studies with dynamic vehicle entry and exit. The empirical results, together with our emphasis on deployable system design, suggest a path towards practical, safe, and scalable highway congestion mitigation.

cs.MA↗
arXiv2026

The Bystander Effect in Multi-Agent Reasoning: Quantifying Cognitive Loafing in Collaborative Interactions

Multi-agent systems (MAS) assume that collaborating inherently improves Large Language Model (LLM) reasoning. We challenge this by demonstrating that simulated social pressure triggers an algorithmic ``Bystander Effect,'' inducing severe cognitive loafing. By evaluating 22,500 deterministic trajectories across 3 dataset contexts (GAIA, SWE-bench, Multi-Challenge) with 3 state-of-the-art (SOTA) models, we semantically audit internal reasoning traces against external outputs. We formalize the \textit{Interaction Depth Limit} ($D_L$), the exact plurality threshold where an agent's logical sovereignty collapses into social compliance. Crucially, we uncover the \textit{Sovereignty Gap}: models frequently compute the correct derivation internally but suffer ``Alignment Hallucinations'' -- actively subjugating empirical evidence to sycophantically appease a simulated swarm. We prove that multi-agent social load is strictly non-commutative; the "brand" identity of the ``Lead Anchor'' auditor disproportionately dictates the swarm's integrity. These findings expose architectural vulnerabilities, proving that unstructured multi-agent topologies can degrade independent reasoning.

cs.MA↗
arXiv2026

CONCAT: Consensus- and Confidence-Driven Ad Hoc Teaming for Efficient LLM-Based Multi-Agent Systems

Although large language model (LLM) based multi-agent systems (MAS) show their capability to solve complex tasks and achieve higher performance over single agent systems, they lead to huge computational overheads because of heavy communication between agents. Previous research has made efforts to train a sparse multi-agent graph or fine-tune a planner to orchestrate the workflow better. However, such extra training processes introduce computational costs and limit MAS to specific domains, therefore compromising their generalizability. In this paper, we propose CONCAT, a training-free multi-agent collaboration framework based on CONsensus and Confidence-driven Ad hoc Teaming to efficiently organize agent interactions. Specifically, agents are clustered based on their initial answers, and leaders of each cluster are selected based on the agents' confidence. Then, a heuristic function based on the Theory of Mind is designed to predict the collaboration benefits between every two leaders according to their answers and confidence. Finally, an ad hoc multi-agent network is organized after evicting a percentage of communications based on the predicted benefits. Experiments across three LLMs and three benchmarks show that CONCAT achieves up to 2.02x higher efficiency (accuracy/latency ratio) than LLM-Debate and outperforms training-aware methods such as AgentDropout, while reducing average latency by 50.1% on Qwen2.5-14B-Instruct, without any task-specific training.

cs.MA↗