Gamified Learning With An AI Board Game Tournament: Course assessment

Table of Links

Abstract and 1. Introduction

2. Course design
3. Course assessment
4. Conclusion, Software and Data, Acknowledgements, and References

3. Course assessment

We did not yet perform a systematic and formal teaching assessment campaign. We base the judgements reported here on the informal interactions we had with students and the overall atmosphere during the lectures and practical sessions.

Overall, our experience with the students has been very positive. Although it was not mandatory to attend, there was high attendance during both lectures and practical sessions, compared to other courses that also did not have mandatory attendance.

Most student groups seemed to show interest in understanding how the algorithms they saw during the lectures were working and came up with their own implementations of the algorithms seen in the lectures.

All student groups managed to create a working agent that interacted with the game server using the proper protocol. Less than 15% of student groups’ agents tried to play bad moves of the game. Around 75% of student groups managed to create agents that were able to beat the random agent in some games. Some student groups’ agents would still lose when the random agent was the first player. Note that playing first indeed gives an advantage in the game. The top 15% of student groups experimented with different heuristics to estimate the quality of a board position in adversarial search algorithms. The best group implemented an iterative deepening minimax algorithm with alpha-beta pruning and transposition tables.

Some students reported that the course motivated them to continue their study in the field of computer science. As this course is given quite early in the studies, it is for us very important that it is designed to be attractive to students.