task_games__darts__highest_scoring_dart_label

Contract

1. Domain: games
2. Scene: darts
3. Scene id: darts
4. Public task id: task_games__darts__highest_scoring_dart_label
5. Supported query_id values: single
6. Answer schema: option_letter
7. Annotation schema: point
8. Program schema: argmax(label for labeled_dart in darts, score(labeled_dart)); scene=darts; scope=highest_scoring_dart_label

Program Contract

Program: argmax(label for labeled_dart in darts, score(labeled_dart)); scene=darts; scope=highest_scoring_dart_label

Candidate set: the visible game board, pieces, tokens, cards, tiles, marked state, legal-move cues, result panels, and labeled options inside the highest_scoring_dart_label objective scope. Operands: visible scene state and prompt-bound operands named by label, labeled_dart, score, darts, and highest_scoring_dart_label. Operation: evaluate argmax over the labeled dart set using the simplified dartboard scoring rule; generation enforces a unique highest-scoring labeled dart. Output binding: answer uses the option_letter schema; generation binds a unique final answer. Annotation witnesses: annotation uses the point schema at the center of the selected dart. Query ids: single.

Reasoning Operations

Families: ranking, formula_evaluation

Generation Notes

1. The scene renders a simplified dartboard with 10 numbered sectors, one center bullseye, and four visible darts labeled A through D.
2. A dart in a numbered sector scores that number; a dart in the center bullseye scores 50.
3. Generated instances place the four labeled darts in numbered sectors and enforce one unique highest score.
4. Query ids are internal replay/sampling keys and do not define public task units.
5. Annotation is the center point of the highest-scoring labeled dart, projected from the same generated game state used for answer verification.