A new round-robin format defines the next Forecasting Dojo meetup. Participants spend ten minutes thinking aloud to decompose questions, allowing subsequent predictors to build on previous reasoning. This iterative approach aims to refine human predictive accuracy. Practitioners can use these techniques to reduce bias and improve risk assessment in AI safety forecasting.