FAQ

Q: What programming language TAG supports?

A: Currently TAG only supports Java.

Q: Where can I find the code for the project?

A: The latest code for TAG is available on Github.

Q: How do I get started with TAG?

A: The Documentation and Resources on this site are a good starting point. To get a better idea of the project you may read “Design and Implementation of TAG: A Tabletop Games Framework”. If you just want to jump into the framework you may check out the “Running a game” page on the wiki.

Q: What can I use TAG for?

A: TAG is primary built as a common interface for AI agents to play various Tabletop games, but we would like to supports more use-cases.

• General Game AI: Adding a new agent should work with any of the available games. Each implemented game provide a Forward Model to support Statistical Forward Planning Agents.
• Game Design: Adding a new game supports all the available agents in the framework. Some of the games uses parameters instead of hard-coded values, which may be changed to find different variants of a game. You may use TAG as a prototyping tool.
• Human interface: The provided games all have an implemented GUI, which allows observing the agents playing the game. Human gameplay is also possible, but due to the large action spaces some games may be a bit difficult to play as a human.

If you are interested in any other use-cases not listed here feel free to get in touch with us.

Q: Can I play games on TAG?

A: TAG is a research tool above all, and is not designed to compete in any way with commercial (or non-commercial) software or websites for playing these games. It will support single-person play against AI agents for some games, but this is mostly for help with debugging and testing (although see the Human Interface research use-case above). There is no multi-player support at present (multiple humans that is, we do support multiple AIs).