Main Street: GDD -- AI Strategy and Hint System

[SorraTheOrc]

Summary

Write the AI and player assistance section of Main Street's Game Design Document covering AI strategy design for auto-play, hint systems, and any tutorial/onboarding guidance.

User Story

As a game designer, I want Main Street's AI behaviour and player assistance systems documented so that we can deliver smart hints and compelling auto-play from an early milestone.

Prerequisites

• Main Street: GDD Core Rules and Mechanics (CG-0MM4RC1K81JU4U5D) completed

Sections to Cover

1. AI Strategy Overview -- What does the AI need to do in Main Street? (Auto-play for testing/demo, hint generation, difficulty simulation)
2. Strategy Tiers -- Define 2-3 AI strategy levels:
   • Random/Naive -- Makes valid moves randomly (baseline, useful for Monte Carlo testing)
   • Heuristic/Greedy -- Follows simple priority rules (e.g. always craft if possible, prefer high-value actions)
   • Lookahead/Smart (optional) -- Considers future consequences of moves
3. Hint System -- How are hints generated? Single best move? Multiple suggestions? Progressive hints (vague to specific)?
4. Move Evaluation Heuristics -- What makes a move 'good' in Main Street? Priority ordering of actions. Scoring function for comparing moves.
5. Tutorial / Onboarding -- Is there a tutorial? How does Main Street teach the player its mechanics? Guided first game? Tooltip-based learning?
6. Difficulty Adjustment (if applicable) -- Does the AI assist in difficulty? Dynamic difficulty? Selectable difficulty levels that change deal generation or available content?

Expected Output

A formal GDD section covering Main Street's AI design with enough detail for an engineer to implement the strategy classes using the engine's existing AI abstractions.

Acceptance Criteria

1. At least 2 AI strategy tiers are fully specified with decision logic
2. Hint system design is documented with player-facing behaviour
3. Move evaluation heuristics are defined and prioritised
4. Tutorial/onboarding approach is documented
5. Design references existing engine AI abstractions (AiStrategyBase, AiPlayer, pickRandom, pickBest)
6. The document is reviewed and approved by the producer