task_games__ludo_board__winning_roll_value

Public taxonomy: games -> ludo_board -> task_games__ludo_board__winning_roll_value.

Program Contract

Program: exact_finish_roll(token_position, finish_cell); scene=ludo_board; scope=winning_roll_value

Candidate set: the visible game board, pieces, tokens, cards, tiles, marked state, legal-move cues, result panels, and labeled options inside the winning_roll_value objective scope. Operands: visible scene state and prompt-bound operands named by token_position, finish_cell, ludo_board, winning_roll_value. Operation: evaluate exact_finish_roll over the candidate set using the visible game state, rules, legal moves, comparisons, counts, simulations, or option-selection constraints encoded in the program expression; generation enforces a unique final answer. Output binding: answer uses the integer schema; generation binds a unique final answer. Annotation schema: point. Annotation witnesses: annotation marks the queried token center. Query ids: single.

The finish cell is the final colored home-lane cell near the center for the queried color. The answer is the exact single die roll needed to move from the queried token's current home-lane cell to that finish cell.

Reasoning Operations

Families: topology, state_update, formula_evaluation

Generator

• Implementation: src/trace_tasks/tasks/games/ludo_board/winning_roll_value.py
• Config: src/trace_tasks/resources/configs/domains/games/ludo_board.yaml
• Prompt bundle: src/trace_tasks/resources/prompts/games/ludo_board/games_ludo_board_v1.json