LLM-as-a-Prophet: Understanding Predictive Intelligence with Prophet Arena

TL;DR

This paper explores the potential of large language models to serve as predictive tools for real-world events, introducing a new benchmark and revealing both strengths and limitations in their forecasting abilities.

Contribution

It introduces Prophet Arena, a comprehensive benchmark for evaluating LLMs' forecasting capabilities across diverse tasks and stages.

Findings

LLMs show promising forecasting accuracy and confidence.

Key limitations include event recall errors and data misunderstanding.

Market-based information aggregation outperforms LLMs near event resolution.

Abstract

Forecasting is not only a fundamental intellectual pursuit but also is of significant importance to societal systems such as finance and economics. With the rapid advances of large language models (LLMs) trained on Internet-scale data, it raises the promise of employing LLMs to forecast real-world future events, an emerging paradigm we call "LLM-as-a-Prophet". This paper systematically investigates such predictive intelligence of LLMs. To this end, we build Prophet Arena, a general evaluation benchmark that continuously collects live forecasting tasks and decomposes each task into distinct pipeline stages, in order to support our controlled and large-scale experimentation. Our comprehensive evaluation reveals that many LLMs already exhibit impressive forecasting capabilities, reflected in, e.g., their small calibration errors, consistent prediction confidence and promising market returns. However, we also uncover key bottlenecks towards achieving superior predictive intelligence via LLM-as-a-Prophet, such as LLMs' inaccurate event recalls, misunderstanding of data sources and slower information aggregation compared to markets when resolution nears.