The POLLOT research project studies political lotteries, which randomly select individuals to take a political decision. Recently, lotteries have put citizens together to find solutions to today’s contentious, polarizing issues, such as disaffection, abortion reform, and climate change. What explains their varied success? Lotteries equalise opportunity of access to all. Random chance can strengthen those otherwise in the minority. A wider variety of opinions can also improve debate and reduce polarization. The ideal, but unfeasible experiment would randomly vary group characteristics, assign real political power, and observe effects over time. Instead, this project draws lessons from past experiments during democratisation in Europe – whereby legislators were randomly assigned to groups within legislatures – and experiments with lotteries in online citizens’ assemblies today. I will present some initial results from the Danish constituent assembly, and from a recent pilot experiment with online citizens’ assemblies deliberating on climate change mitigation.