Overview
Led by Professor Peter Bossaerts, Faculty of Economics
When students can use AI to write much of their code, looking at the code itself no longer shows what they understand. This project tests a different approach in Algorithmic Trading Robot Design, an award-winning economics course taught like a robotics class. Students build Python trading robots that trade in a controlled market. The robots are assessed on how they behave against hidden test cases, designed to see whether a strategy holds up when markets do something unexpected.
The team will build an AI tool that analyses each robot’s performance, explains where and why the strategy failed, and gives feedback the student can act on before resubmitting. The project will look at whether this approach supports students’ understanding, and whether the feedback improves their learning. The team will produce practical guidance on behaviour-based assessment for any course that involves programming.
