Very interesting post! As I was reading Phase 2, I was wondering if I could exploit the easy AI’s predictability and goad it into making bad moves… but then Phase 3 came and ruined it :)
For the closing remarks: I would love to read more posts like this.