task_games__backgammon__pip_count_value¶
Contract¶
- Domain:
games - Scene id:
backgammon - Public task id:
task_games__backgammon__pip_count_value - Supported
query_idvalues:single - Answer schema:
integer_value - Annotation schema:
bbox_set - Program schema:
sum(map(active_player_occupied_points, checker_count * distance_to_bear_off)); scene=backgammon; scope=pip_count_value
Program Contract¶
Program: sum(map(active_player_occupied_points, checker_count * distance_to_bear_off)); scene=backgammon; scope=pip_count_value
Candidate set: the visible game board, pieces, tokens, cards, tiles, marked state, legal-move cues, result panels, and labeled options inside the pip_count_value objective scope.
Operands: visible scene state and prompt-bound operands named by map, active_player_occupied_points, checker_count, distance_to_bear_off, backgammon, pip_count_value.
Operation: evaluate sum 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 witnesses: annotation uses the bbox_set schema; the prompt/annotation contract defines the minimal visual witnesses.
Query ids: single.
Reasoning Operations¶
Families: aggregation, topology, formula_evaluation
Generation Notes¶
- Query ids are internal replay/sampling keys and do not define public task units.
- The generated board is a sparse exact-answer race position with answer support
1..8. - Pip distance follows standard Backgammon bearing-off distance: black checkers on point
pcontributepeach, while white checkers on pointpcontribute25 - peach. - Pip-count renders omit dice and add
Dlabels beside active-player stacks to show the distance-to-bear-off value used in the sum. - Annotation is projected from active-player occupied point bboxes that contribute to the pip-count sum, not individual checker bboxes.