TimeSeek: Temporal Reliability of Agentic Forecasters

1. 1Evaluate LLM forecasters at multiple temporal checkpoints
2. 2Analyze performance on high-uncertainty and low-uncertainty markets
3. 3Compare model performance with and without web search
4. 4Identify key factors influencing model competitiveness over time

AI researchers and engineers working on LLMs and forecasting models can benefit from this study to improve their models' reliability over time, and data scientists can apply these findings to develop more accurate forecasting systems

💡 LLM forecasters are most competitive early in a market's life and on high-uncertainty markets

📊 TimeSeek: Evaluating LLM forecasters' reliability over time in prediction markets