2-Year Test Shows SOTA LLMs Suck at Stock Trading
garrytan · x · 2026-08-01
A deep analysis evaluating SOTA LLMs over a 2-year stock trading period yields negative conclusions. The study reveals that:
- Poor Performance: No model comes close to beating a simple static baseline.
- Reasoning Doesn't Help: Higher reasoning capabilities do not translate into better trading decisions.
- Behavioral Bias: When losing money, models tend to trade less frequently rather than optimizing their strategies.
Related event: LLMs Fail to Beat Equal-Weight Baseline in Stock Trading(2 posts)→
More from Research
- It's not gradient descent: burny_tech corrects the diagnosis to RLHF mode collapse — burny_tech · 2026-10-16
- A beginner-friendly introduction to causal inference kicks off, Part 1 — Nasereliver · 2026-10-16
- FBI method trains for Pass@K with one sample per update at 12.3% of the cost — EmtiyazKhan · 2026-10-16
- Berkeley researchers sample humanoid robot policies directly from weight space with diffusion — berkeley_ai · 2026-10-16
- OneWM: unified causal training shows world and action modeling improve each other — IAAI233 · 2026-10-16
- Baidu's Virtual Head Attention tames the KV cache, FLOPs, quality triangle in long-context LLMs — baidu · 2026-10-16