Skip to content

KCL 6CCS3PRJ - AI opponent for the card game Ulti

License

Notifications You must be signed in to change notification settings

almasen/Ulti-mate

Repository files navigation

Ulti-mate

KCL 6CCS3PRJ - AI opponent for the card game Ulti

Test Coverage

About

Ulti-mate is a proof of concept implementation of a set of smart bidding algorithms for the complex trick-taking card game Ulti designed as part of my BSc thesis at King's College London under the supervision of Dr Agi Kurucz.

The supported list of smart bidding algorithms is as follows:

  • Simple
  • Simple of Hearts
  • Betli
  • Rebetli
  • Open Betli
  • Durchmars
  • Redurchmars
  • Open Durchmars

CLI

Ulti-mate provides an extensive CLI with various experimentation options as well as detailed reasoning behind computed decisions.

Commands

  • npm start to run a simulation
  • npm start -- [args] to pass CLI arguments
  • npm start -- -h to see CLI help
  • npm test to run TS module tests (concurrently)

CLI args

  • -h for CLI commands help
  • -R, --max-risk to specify max risk in percentages
  • -H, --custom-hand <card ids...> to run the simulation with a custom hand
  • --no-opponents to skip calculation of opponents' hands
  • --no-reasoning to skip logging of detailed reasoning behind decisions

CLI examples

  • npm start -- --no-opponents runs simulation with only player's hand
  • npm start -- -R 70 the bot should only play games where the winning chance is >=70%
  • npm start -- -H 0 1 2 3 4 5 6 7 8 9 sets player's hand to cards whose ids are 0-9

Local usage & development

To run Ulti-mate simulations or develop locally.

Prerequisites

  • node.js ^14.x
  • min 2 GB free memory

Set up

  1. npm i to install project dependencies
  2. npm run build to compile project
  3. Run your choice of CLI command

License

MIT © Daniel Almasi

About

KCL 6CCS3PRJ - AI opponent for the card game Ulti

Resources

License

Stars

Watchers

Forks