Blog
Analysis that comes out of self-play game analysis, and details on how the model works.
Simulating a Homeworlds Prerelease
What to expect from the Homeworlds Prerelease meta. Which leaders to expect, which leaders to avoid, and which are the best cards to pull for each archetype.
Understanding Matchups with Reinforcement Learning
How the SwooLab model learns to play, and what 132,000 simulated games across twelve Spark of Rebellion decks say about the Set 1 matchup table and the metagame it points to.