King's College London

It is well known that the computation of solutions to decision and enumeration problems in argumentation can be very hard. In this work, we analyse some of the results of the 2017 International Competition on Computational Models of Argumentation. Our analysis shifts the focus from the performance of individual solvers to how well/badly they can collectively tackle different classes of abstract argumentation frameworks. In so doing, we were able to identify the instances that were particularly difficult for all/most solvers and look into their particular structural properties.