After a few months of work, I’ve finally released an update to my brickbattle game, Brickbattle Frenzy. The game features procedurally generated maps and is a casual brickbattle experience with 9 different gamemodes, whether free-for-all, team-based, or asymmetrical games like a zombie survival or a boss fight.
The most recent update features Brickbattle bots which are able to pathfind on those procedural maps, using advanced techs from the competitive brickbattle community such as Trowel Climb (using your trowel against a wall as a temporary platform to give yourself a much higher jump height) and much more.
The bot works with a sort of dual state machine implementation: one state machine will handle its pathfinding, seeking either new targets, game objectives or cover, while another state machine will determine how the AI should try to use tools to fight. As both state machines may require the use of tools (pathfinding may require the bot to use advanced movement techs like rocketjump), a core behavior will observe the two state machines and allow one or the other to control the bot’s usage of tools based on the situation. Credits to @prooheckcp for their implementation of the RobloxStateMachine library which is used extensively in this project. Credits to @bigobook, who worked on a similar bot in the past, for developing and helping me implement the cover-seeking functionality of this bot.
The AI has several difficulty levels, always selecting Medium by default on all public servers; however if you’re genuinely interested in testing this behavior, I encourage you to check out our free private servers where you’ll have access to a Team Editor that lets you tweak the difficulty of the AI (have fun trying to beat the hardest difficulty in a deathmatch duel!)
I’ve been interested in roblox brickbattle for as long as I’ve been around, and was super interested in the old brickbattle bot implementations from the SNCAI games, if you’ve ever seen those. This is a revival of my first brickbattle bot project which I worked on in 2020, and I’m pretty happy with how the new version turned out. If you’re at all interested in this game genre or in NPC development, I’d love to hear your thoughts!