StarCraft Competition Insights and Observations
Discover valuable insights from the StarCraft competition including bot submissions, basic rules, maps, cooperation with AIIDE, and key observations on bot performance on different map types.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
E N D
Presentation Transcript
Tobias Mahlmann and Mike Preuss CIG 2012 StarCraft competition September 14, 2012 09-04-2025 1
Submissions Botname Race Contributor Nova Terran Alberto Uriarte Drexel University none Skynet Protoss Andrew Smith SCAIL Protoss Jay Young The University of Birmingham AIUR Protoss Florian Richoux University of Nantes BroodwarBotQ Protoss Gabriel Synnaeve INRIA, Coll ge de France Xelnaga Protoss Ho-Chul Cho Sejong University BTHAI Terran Johan Hagelb ck Blekinge Institute of Technology Adjutant Terran Nicholas Bowen University of Central Florida IceBot Terran Wang Zhe Ritsumeikan University UalbertaBot Protoss David Churchill University of Alberta 09-04-2025 2
Basic Rules o 90 rounds of round robin on 6 maps (= 15 times each bot against each on every map) o games played in highest possible speed, set by tournament software (from Alberta people) o bots ordered according to overall winning rate o crashing or timeout means win for opponent o hour timeout (=85716 frames) means draw o due to technical problems, read/write feature constraint to games on a single computer (of 6) 09-04-2025 3
Maps o3 3-player maps: Athena-II Neo Moon Glaive Tears of the Moon o 3 6-player maps: Legacy River of Light The Huntress 1.1 09-04-2025 4
Cooperation with AIIDE People o tournament software provided by University of Alberta team (Michael Buro, David Churchill, Jason Lorenz) o client-server system, clients can enter any time (2 per game needed) o starcraft/bot crashes handled automatically o note: very similar setup to AIIDE 12, only our maps were not known in advance 09-04-2025 5
3 Player Maps (Left), 6 Player Maps (Right) 09-04-2025 6
Some Observations o map size has an effect! o Nova against BBQ: good on 6-player, bad on 3-player o SCAIL and Adjutant better on 6-player, BBQ worse on 6-player o UAlbertaBot better on 6-player (due to scouting?) o Nova consistently good against UAlbertaBot o Xelnaga only bot to beat Skynet, but only on 3-player maps 09-04-2025 7
Averaged over all 6 Maps 09-04-2025 8
Overall Win Rate History 09-04-2025 9
The Medalists 3rd: AIUR, Florian Richoux 2nd: UAlbertaBot, David Churchill 1st: Skynet, Andrew Smith and, honorary mention to: Adjutant, Nicholas Bowen (1sttime, best terran) 09-04-2025 10
Summary things (many) bots can t do: o handle islands o play zerg o play maps of different sizes (!!!) o play against random race o finish off beaten opponent o adapt to opponent strategy o cooperate (no teamplay) 09-04-2025 11
The Future still many problems, but: o huge step in average bot performance (see BTHAI as baseline ) o we need to change basic approach: not re- implement everything, but use available stuff o finally come up with useable modules o bots need to become more flexible o very interesting: step into team AI overall: we are getting to the interesting phase now! use tournament software for research! 09-04-2025 12
