This issue of the ICGA Journal is a double issue numbers 3 and 4, concluding the 2017 volume of the journal. In this issue we have included two new research articles and a collection of reports for the 2017 ICGA events. We regret the publication delays that occurred in 2016/2017 but are glad to report we are on a good track to publish four issues in a more timely manner for 2018.
One of the main aims of the annual ICGA events is to facilitate the exchange of new ideas and techniques for computer game researchers and developers. This year’s events in Leiden, consisting of the Fifteenth International Conference on Advances in Computer Games (ACG2017), the World Computer Chess Championship, and the Computer Olympiad are a continuation of this tradition, with a few changes to the format. We have included daily reports on ACG2017, the results of WCCC 2017, and several Computer Olympiad reports for your enjoyment.
The first contribution in this issue by Cameron Browne on the analysis of the puzzle game “Gloop” is a good example of this exchange of ideas. The game’s inventor Fred Horn introduced the game to Browne at the previous year’s (2016) conference, which motivated him to thoroughly investigate and take on one of the long standing challenges – more than a decade, in fact – put forth by Horn. The article’s verdict on this so-called “perfect packing” challenge was arrived at through computer search, followed by careful mathematic formulation and proof.
The second contribution by Nobuo Araki et al. presents an improvement on their previous work on the technique of simulation adjusting, which aims to develop an overall better simulation policy for Monte Carlo methods. The authors are the team behind the Go program MC_ark, which has participated in the UEC Cup in Japan.
We have all witnessed many ground-breaking developments in our field this past year, the most exciting of which is Deepmind’s AlphaGo Zero, and recently, AlphaZero. The former has continually defied many of our expectations since its previous versions were announced, while the latter has opened the door for all kinds of game A.I. development. It is now possible, with enough computing resources, of course, to create programs that exceed expert human play with close to no expert domain knowledge I imagine this will inspire many new topics of research, and encourage all of our readers to share any new findings with the community.
With the above in mind, readers are welcome to join the 2018 ICGA events, with the conference taking place in Taiwan and Stockholm jointly. I look forward to seeing many of you in the coming year.