Preference Optimization Drives Monoculture in LLM Prediction Markets
This paper identifies a critical failure mode for LLM-based prediction markets, showing that alignment techniques like DPO undermine the independence assumption essential for market accuracy.
LLM agents fine-tuned with Direct Preference Optimization exhibit high pairwise error correlations (ρ=0.70), reducing the effective forecasting power of ten agents to that of approximately 1.4 independent forecasters. This monoculture effect is causally driven by preference optimization and persists across scales, with cross-model diversity being the most effective mitigation.
Prediction markets rest on the independence of participant errors. As LLM agents become active traders on platforms like Kalshi and Polymarket, we ask: does this independence hold when the crowd is composed of LLMs? We find it does not. LLM agents fine-tuned with Direct Preference Optimization (DPO) share a convergent output distribution, producing pairwise error correlations of $ρ= 0.70$ and reducing ten agents to the effective forecasting power of ${\approx}1.4$ independent forecasters $N_{\text{eff}}$. This is not a scaling problem: $N_{\text{eff}}$ remains flat from $N=5$ to $N=40$, and the 10-agent market (67.6%) fails to match a single standalone agent (70.2%). Two controlled ablations isolate preference optimization as the causal driver, replicated across labs and scales ($Δρ= +0.24$ to $+0.46$ on identical-SFT controls at 8B and 70B). Among mitigations tested, cross-model diversity achieves the largest correlation reduction ($ρ$ from 0.68 to 0.40). As LLMs become more aligned, markets built from them become more monocultural.