Project overview
Making predictions of future outcomes in an uncertain environment is a difficult but important task, as these predictions are often required for making informed choices by individuals, companies and Governments. The project developed innovative approaches to modelling information such that predictions could be developed which outperformed market predictions in a complex setting where prediction markets had been thought to use information efficiently. The project revealed that certain types of data are often under-utilised when making predictions, notably complex data which requires some pre-processing, time-related data and more novel data, which it appears market participants are slow to utilise. The project resulted in a number insights regarding the reasons for mis-pricing in prediction markets and means of correcting for likely over-reactions to certain types of information. The project resulted in several papers being published in leading international journals and formed the basis for a number of subsequent PhD theses.
