Agentic Time Machine as an Infrastructure for Future-Event Forecasting

2026-06-18

AI Summary

This work addresses the longstanding trade-off in future event prediction evaluation between efficiency and environmental fidelity: retrospective evaluations lack dynamic realism, while real-time assessments suffer from prohibitively long cycles. To reconcile these constraints, we propose Agentic Time Machine—the first prediction sandbox system that simultaneously achieves high efficiency and high realism. Our approach reconstructs historical web states through time-truncated webpage snapshots and employs a multi-agent framework comprising planner, solver, and aggregator modules to enable parallel evidence collection and collaborative reasoning. Built upon large language models, our architecture significantly outperforms strong baselines on both FutureX-Past and Polymarket benchmarks, securing the top average rank for four consecutive weeks on the official FutureX live leaderboard and ranking first overall across eight weeks.

Abstract

Forecasting future events is a critical challenge for large language model (LLM) agents, spanning domains from elections and monetary policy to financial markets. However, evaluating progress on this task presents a fundamental trade-off between efficiency and environment fidelity. While live evaluation benchmarks suffer from an inherently slow feedback loop, existing retrospective replays typically restrict agents to static, pre-frozen databases that sacrifice the environmental realism of actual deployments. To tackle this issue, we introduce Agentic Time Machine (TM), an infrastructure that approximately reconstructs the web state at any chosen past time by filtering post-cutoff content. Leveraging this evaluation infrastructure, we further propose a planner-solver-aggregator multi-agent framework that breaks each question into diverse analytical angles, gathers evidence in parallel, and combines the results into a single forecast. Experiments show that offline scores under TM correlate strongly with live FutureX scores, validating that TM offers a fast and reliable sandbox for forecasting-agent evaluation. On FutureX-Past and Polymarket evaluated under TM, our framework achieves the highest score among strong closed-book, tool-augmented, and self-consistency baselines. On the official FutureX live leaderboard, our system achieves the best average rank over four consecutive weeks, including 1st place in May Week 1. As of June 17, it also ranks 1st on FutureX's official eight-week overall leaderboard.

Problem

• future-event forecasting
• evaluation infrastructure
• environment fidelity
• LLM agents
• forecasting benchmark

Innovation

• Agentic Time Machine
• future-event forecasting
• multi-agent framework
• offline evaluation infrastructure
• web state reconstruction