task_games__cards__blackjack_best_hand_label

Contract

1. Domain: games
2. Scene id: cards
3. Public task id: task_games__cards__blackjack_best_hand_label
4. Supported query_id values: single
5. Answer schema: string_label
6. Annotation schema: bbox_set
7. Program schema: label(arg_extreme(hands, metric=blackjack_score(hand), direction=highest_valid)); scene=cards; scope=blackjack_best_hand_label

Program Contract

Program: label(arg_extreme(hands, metric=blackjack_score(hand), direction=highest_valid)); scene=cards; scope=blackjack_best_hand_label

Candidate set: the visible game board, pieces, tokens, cards, tiles, marked state, legal-move cues, result panels, and labeled options inside the blackjack_best_hand_label objective scope. Operands: visible scene state and prompt-bound operands named by arg_extreme, hands, metric, blackjack_score, hand, direction, highest_valid, cards, blackjack_best_hand_label. Operation: evaluate label 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 string 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: ranking, formula_evaluation

Generation Notes

1. Prompt wording comes from src/trace_tasks/resources/prompts/games/cards/games_cards_v1.json.
2. Annotation is projected from the same generated game state used for answer verification.