This is an implementation of the Ultimate Tic Tac Toe game. The game is written in Java using JavaFX and FontAwesomeFX by Jens Deters.
- Open the project as a Maven project. Dependencies (including JavaFX) are resolved automatically via
pom.xml. - Run
Launcher(insrc/dk/easv/Launcher.java) — click the green play arrow next to itsmainmethod.
Any JDK 21+ works (no special JavaFX distribution needed).
On the start screen you choose player types (human or AI) and select bots from the dropdown. The bot move delay slider controls how fast bot-vs-bot games play out.
During gameplay, the board highlights playable sections in gold. Players place moves as diamonds (green, player 0) or trash cans (red, player 1). Won macro-board sections show a large icon for the winning player.
When a player wins, the result is shown as an overlay on the board.
Here is a nice explanation of the rules of the game https://www.thegamegal.com/2018/09/01/ultimate-tic-tac-toe/
The game can be played as either human or bot, and any combination can be used: human-human, human-bot, bot-human or bot-bot.
The bots must follow the IBot interface in the BLL package. The game uses reflection for loading class files from the Bots folder under BLL. To implement your own bot, create a class that implements IBot and place it in the dk.easv.bll.bot package.
When the game starts it creates a list of the bot names in the project root folder (bots.txt). This is for usage with online tournament tools.
The bots provided with the game are simple examples — see src/dk/easv/bll/bot/README.md for descriptions.
A REST client bot that plays against a remote server hosting stronger MCTS bots. The REST server is part of the UTTT challenge at SEA Business Academy and is not publicly available. If you'd like access to test your bot against the server, feel free to reach out.
Change SERVER_URL in TeacherBotRESTv2.java to point to your server. Change DEFAULT_SLUG to select the bot strength:
teacher— Basic MCTS (strength 2.0)r16linrave— MCTS + RAVE (strength 3.0)r19bitreuse— Bitboard + tree reuse + RAVE (strength 3.5)r20multiroll— R19 + multi-rollout (strength 4.0)
Network/connection errors are clearly reported instead of showing as "false move".
Configurable delay (0–2500ms, default 1000ms) for bot-vs-bot games so you can watch the game progress.
Run bulk bot-vs-bot games across all CPU cores. Games are split evenly: half with bot A as player 0, half swapped. The simulation slider goes from 1 to 200 games.
The stats window is split into two resizable sections:
Current Games — live mini-board cards showing every active game. Each card displays the full 9x9 board with colored cells (green = player 0, red = player 1, blue = tie). Won macro-board sections turn solid. The current player's name shows a ▶ arrow. Cards appear when games start and disappear when they finish.
Finished Games — two columns (one per player) with W/L/T stats and mini-board cards showing the final board state of each completed game.
Other features:
- Progress bar — animated indeterminate bar while simulation is running
- Warning banner — reminder about REST bot server load (closeable)
When both bots have the same name, they're automatically suffixed (#1, #2) to distinguish them in stats.
During bot-vs-bot games, the UI disables the board to prevent accidental clicks while the AI is computing.