💥 Try it on Lichess! 💥
This basic Chess game is a humble tribute to famous Chess player Robert James "Bobby" Fischer, World Chess Champion.
Note that the author does not agree with Fischer's political and religious opinions, but remains a fan of his genius at Chess.
Note: requires Java 17
If you are interested in how this engine works, please have a look at those two articles by the author:
- Implementing a Chess engine from scratch
- Enhancing a Java Swing App to a clean, elegant Web App without changing the code
- If not already done, you can install OpenJDK 17 for free (see the excellent distributions of Temurin by Adoptium)
- Download the latest release JAR
- Then launch the downloaded JAR
java -jar bobby-0.1.jar
- Checkout the code
- Build it with maven
mvn package -Dmaven.test.skip=true - Launch the created JAR
java -jar target/bobby-1.0-SNAPSHOT.jar
Thanks to the great lib Chariot, it was possible to play against Bobby on Lichess.
Just browse to bobby-bot profile on Lichess.org and challenge it! If the server is running, the bot will accept the challenge and play.
However, the bot has been stopped for costs reasons.
There was a browser version of the game (limited to one single user at a time), but it was stopped for costs reasons.
- AI with multiple strength level, from random-stupid to 3-depths computation, using a minimax algorithm
- All moves implemented
- Change GUI view according to selected color
- Draw proposals
- Undo last move
- Suggest move
- Usage of 15+ famous openings
- Limit computation time to n seconds
- Save game to and load from text files with basic notation
- Load PGN file
- AI is as arrogant as the real Bobby was 😄
- Uses Java 17
- Ability to use bundled light JRE (doc here)
- Strong code coverage, incl. GUI testing
- Code style with Checkstyle, code quality with SpotBugs and SonarSource
- Pipeline with GitHub Actions
- Uses a nice modern Look & Feel FlatLaf for all platforms
- Uses free font FreeSerif in order to have a nice rendering of chess pieces
The implemented AI works uses a depth-first computation, which means that if the computation time is restricted, it may not evaluate every single possible move: it evaluates as deep as possible a first move, then a second, etc, but has no guarantee to cover every move of the first depth.