I’ve upgraded the bot’s intelligence. Instead of only being aware of the enemy when they are in its immediate vision, the bot now remembers where it previously saw the enemy, keeps track of how many opportunities to move the enemy had since then, and precisely calculates all the squares the enemy could be on using a flood fill algorithm, based on how many moves the bot thinks the enemy had. (The eye icons represent what the bot thinks are the possible positions of my player based on where it last saw my player. (Sometimes the player can “escape” the eyes by having more moves than what the bot thinks the player has.) (Otherwise, if the bot is right about how many moves the player has, the player should not be able to move to a square without an eye icon.))