Multi-agent trading system tests LLM news judgment and tracks attacks
Start here
Researchers built a Jev-compatible multi-agent trading system to test whether the system tracks or follows web news when it lags the market. In a hash-commitment test, a reader based on fixed-version LLM option probabilities correlated with abnormal returns in announcement windows for web news (weekly Spearman 0.119), and event types matched issuers' 8-K filings (AUC 0.833). A weekly confirmation test found no ranking information beyond code-only features; in an attack scenario, decision cards traced 88% of successful attacks to planted entries.
Generated by AI from reports · Updated 2 days ago
Attention and developments
Attention, hourly
Now 24, -2% in 24 hours·Peak 26 (Nov 10 23:00)
Related research
This event's paper
- Track or Follow? A Jev-Compatible Multi-Agent Trading System When Web News Lags the Market
arXiv · 2026-11-07 · Deep reading
Progress timeline
- arXivJev-compatible multi-agent trading system: when online news lags the market, track or follow?
Researchers built a Jev-compatible multi-agent trading system; in a hash-commitment test, a reader based on fixed-version LLM option probabilities judged online news in a way correlated with abnormal returns in announcement windows (weekly Spearman 0.119), and event types matched issuers' 8-K filings (AUC 0.833). A weekly confirmation test found no ranking information beyond code features alone, and under attack the decision card traced 88% of successful attacks to planted entries.
Related discussion
Attention trend
New sources per day
- Nov 10: 1 new sources (1 outlets, 0 accounts)
- Nov 11: 0 new sources (0 outlets, 0 accounts)
- Nov 12: 0 new sources (0 outlets, 0 accounts)
- Nov 13: 0 new sources (0 outlets, 0 accounts)