Existential threats to humanity can only be mitigated if they can be predicted. Complex statistical models, such as those used in climate science since the 1960s, can provide a valulable insight into the future, but it is difficult to make use of qualitative data and models tend to be limited in the amount of data from which they can make predictions. Human intuition, when combined with statistical modelling, can improve predictions, as is the case with superforecasters. This project demonstrates a novel approach to training this aspect of human intuition using historical data in a game-like environment.
(Built during HackX Cambridge 2022)