CardCoach
An AI sparring partner for popular two-player card games like Lost Cities, Jaipur, and Splendor that adapts difficulty inside a single match and explains where your line went wrong, so hobbyists can sharpen between in-person game nights.
Two-player tabletop card game hobbyists who want a smart opponent and a coach between in-person sessions
- Adaptive Difficulty Coach: opponent scales strength inside a session so beginners don't get crushed and intermediates stay challenged
- Strategy Explain Mode: post-game walkthrough pointing out which decision lines lost EV and what the better play would have been
- Cross-Game Profile: tracks your rating and tendencies across Lost Cities, Jaipur, Splendor and similar decks
- Variant Sparring: lets you drill specific deck configurations or endgame states you want to study
Recent advances in dynamic resource allocation for Monte Carlo search over games with hidden information directly translate into a stronger, less predictable two-player card-game opponent — the exact hobby upgrade tabletop fans have asked for.
Board Game Arena (millions of users, hundreds of games including these titles) and dedicated Splendor/Lost Cities apps demonstrate steady demand for online play of these exact two-player card games, but the specific 'coach with line analysis' niche is unproven.Play board games online from your browser • Board Game Arena ↗Splendor AI Game | AI Board Games ↗
Existing players (BGA, aiboard.games, official Splendor app, Git hobby ports like bevy_jaipur) offer static Easy/Medium/Hard bots but none combine in-match adaptive difficulty with post-game line explanation — the coach layer is open.Splendor AI Game | AI Board Games ↗GitHub - dmackdev/bevy_jaipur ↗
BGA's working subscription model proves hobbyists will pay for premium online tabletop access, but a single-developer 'coach' niche is small; willingness to pay for an analytical layer (vs. just an opponent) is unproven.Board Game Arena: Subscription Plans ↗Board Game Arena Review: Is BGA Worth Your Time and Money? ↗
These specific card games have decade-long tails on BoardGameGeek and print through multiple editions.
Open MCTS/ISMCTS implementations already exist for Splendor (academic Splendor-AI) and Jaipur; the harder build is the replay/explanation UI and adaptive-difficulty controller, which are well-scoped but non-trivial.Splendor-AI documentation ↗Splendor AI Game | AI Board Games ↗