[论文解读] LLM as a Risk Manager: LLM Semantic Filtering for Lead-Lag Trading in Prediction Markets

Sumin Kim, Minjae Kim+8 more|arXiv (Cornell University)|Feb 4, 2026

Stock Market Forecasting Methods被引用 0

一句话总结

一个两阶段框架将 Granger 因果与基于 LLM 的语义筛选结合起来，在预测市场中对 lead–lag 关系进行排序，从而提升交易表现并在 Kalshi 数据上降低下行风险。

ABSTRACT

Prediction markets provide a unique setting where event-level time series are directly tied to natural-language descriptions, yet discovering robust lead-lag relationships remains challenging due to spurious statistical correlations. We propose a hybrid two-stage causal screener to address this challenge: (i) a statistical stage that uses Granger causality to identify candidate leader-follower pairs from market-implied probability time series, and (ii) an LLM-based semantic stage that re-ranks these candidates by assessing whether the proposed direction admits a plausible economic transmission mechanism based on event descriptions. Because causal ground truth is unobserved, we evaluate the ranked pairs using a fixed, signal-triggered trading protocol that maps relationship quality into realized profit and loss (PnL). On Kalshi Economics markets, our hybrid approach consistently outperforms the statistical baseline. Across rolling evaluations, the win rate increases from 51.4% to 54.5%. Crucially, the average magnitude of losing trades decreases substantially from 649 USD to 347 USD. This reduction is driven by the LLM's ability to filter out statistically fragile links that are prone to large losses, rather than relying on rare gains. These improvements remain stable across different trading configurations, indicating that the gains are not driven by specific parameter choices. Overall, the results suggest that LLMs function as semantic risk managers on top of statistical discovery, prioritizing lead-lag relationships that generalize under changing market conditions.

研究动机与目标

• 识别事件级预测市场中稳健 lead–lag 关系的挑战动机。
• 提出一个将统计发现与语义验证相结合的两阶段框架。
• 评估基于 LLM 的重新排序是否在 Granger-only 筛选之上提升交易绩效。

提出的方法

• 阶段 1 在对数胜率变换的市场价格上应用 Granger 因果关系，以在多个滞后（p∈{1,2,3,4,5}）中识别候选的 leader–follower 对。
• 阶段 2 使用 LLM（GPT-5-nano）对事件标题/描述进行评估，以评估可能的经济传导机制并分配可信度分数，从而对这些候选进行重新排序。
• 交易协议：对每对 leader–follower，当领导者价格变动超过阈值时触发跟随交易，持有 h 天，衡量样本外盈亏。
• 最终投资组合从统计中选取前 M 对有向对（M=20）；作为对比的混合方法。
• 评估采用滚动窗口，在 Kalshi Economics 市场上进行 60 天训练和 30 天测试。

Figure 1: Two-stage framework for leader–follower pair discovery in prediction markets. Stage 1 produces a candidate set of Top K directed pairs (K=100) ranked by Granger significance, and Stage 2 applies LLM-based semantic re-ranking to select the final Top M portfolio (M=20).

实验结果

研究问题

• RQ1 LLMs 是否能够将机理上可行的 lead–lag 关系与 Granger 因果识别出的脆弱统计相关性区分开来？
• RQ2 语义重新排序是否在 Granger 基础筛选之上提升交易绩效并降低下行风险？
• RQ3 在不同的持有期和市场条件下，收益是否具备鲁棒性？

主要发现

• 基于 LLM 的语义筛选在统计基线之上实现了显著的总盈亏提升（+205%）。
• 下行风险降低，平均亏损从 $649 降至 $347（降低 46.5%）。
• 在默认设置下胜率从 51.4% 提升至 54.5%。
• 在同事件与不同事件对中，亏损减少仍然存在（40%+ 的改进）。
• 在领导者价格移动较大时（5–10 点及 10 点以上移动）语义筛选带来更大收益。
• LLM 选择的对可能包含在 Granger Top-M 之外的经济有意义的联系，体现出超越统计排序的定性收益。

Figure 2: Signal-triggered trading protocol used to evaluate ranked lead-lag relationships from Figure 1 : leader price moves trigger follower trades, with direction determined by the Granger-induced trade sign and out-of-sample PnL used to evaluate the ranked pair list.