Research

We publish research on prediction market microstructure, execution, and quantitative trading strategies. Our work bridges academic rigor with practical market infrastructure.

arXiv Prediction Markets LLM February 2026

LLM as a Risk Manager: LLM Semantic Filtering for Lead-Lag Trading in Prediction Markets

Sumin Kim, Minjae Kim, Jihoon Kwon, Yoon Kim, Nicole Kagan, Joo Won Lee, Oscar Levy (River Markets), Alejandro Lopez-Lira, Yongjae Lee, Chanyeol Choi

We combine statistical and semantic approaches to identify trading relationships in prediction markets. Using Granger causality to identify candidate leader-follower pairs, we then apply LLM assessment to filter for relationships with plausible economic transmission mechanisms. Tested on Kalshi Economics markets, the hybrid strategy improved win rates from 51.4% to 54.5% and reduced average losses from $649 to $347.

Read on arXiv
arXiv Prediction Markets LLM February 2026

Forecasting Future Language: Context Design for Mention Markets

Sumin Kim, Jihoon Kwon, Yoon Kim, Nicole Kagan, Raffi Khatchadourian, Wonbin Ahn, Alejandro Lopez-Lira, Jaewon Lee, Yoontae Hwang, Oscar Levy (River Markets), Yongjae Lee, Chanyeol Choi

We examine how to design input context for prediction markets that forecast whether companies will mention specific keywords during earnings calls. We introduce Market-Conditioned Prompting (MCP), which treats market-implied probability as a starting point and instructs language models to update this prior using textual evidence.

Read on arXiv

Get Early Access

Enter your email and we'll reach out when your account is ready.