DraftFM：用于万智牌零日 drafting 的基础模型
DraftFM: A FoundationModel for Day-Zero Drafting in Magic: The Gathering
AI总结：
该研究针对万智牌 drafting 的零日 regime，提出离散选择策略 DraftFM，用 160 万参数网络拟合 1.49 亿个人类 pick，在预留系列预测达 50.8%-60.4% top-1 一致性，还为未发布的《霍比特人》系列生成卡牌排名，与专家评审表现相当。
AI中文摘要：
万智牌（Magic: The Gathering）新系列的 drafting 工作在任何 pick 被观察之前就已开始：完整的卡牌列表是公开的，但监督式 pick 模型训练所需的 drafting 日志尚未存在。我们直接研究这一零日 regime（阶段）。DraftFM 是一种离散选择策略，仅对当前卡包中可用的卡牌进行评分，条件是已 drafting 的池和 drafting 状态。每张卡牌作为其公开卡牌记录、结构化特征及固定文本嵌入的冻结 775 维函数输入，模型中不包含任何卡牌身份、系列身份或使用统计数据，因此未见过的卡牌与熟悉的卡牌会被相同机制评分。在来自 29 个系列的 1.49 亿个人类 pick 上拟合的 160 万参数网络，对完全预留的 3 个系列的保留 pick 进行预测，达到了 50.8%、60.4% 和 56.7% 的 top-1 一致性，而首轮 pick 的随机概率约为 7%。在所有 32 个已观测系列上重新拟合后，相同架构为当时未发布的《霍比特人》（The Hobbit）系列生成了卡牌排名，该排名带有完整的加密来源，并在该系列可在 MTG Arena 上 drafting 前约 36 小时发布。该 sealed（密封）排名与 6 位独立专家评审的一致性，大致相当于这些评审之间的一致性。无论结果如何，针对实际结果的评估将在后续说明中呈现。
英文摘要：
Drafting a new Magic: The Gathering expansion begins before any pick from it has been observed: the complete card list is public, but the draft logs that supervised pick models train on do not yet exist. We study this day-zero regime directly. DraftFM is a discrete-choice policy that scores exactly the cards available in the current pack, conditioned on the drafted pool and the state of the draft. Every card enters as a frozen 775-dimensional function of its public card record, structured features and a fixed text embedding, with no card identities, set identities, or usage statistics anywhere in the model, so an unseen card is scored by the same machinery as a familiar one. A 1.6-million-parameter network fitted on 149 million human picks from 29 expansions predicts held-out picks in three expansions withheld in their entirety, reaching 50.8%, 60.4%, and 56.7% top-1 agreement, where uniform chance at the opening pick is about 7%. Refitted on all 32 observed expansions, the same architecture produced a card ranking for the then-unreleased set The Hobbit, sealed with its complete cryptographic provenance and published roughly 36 hours before the set became draftable on MTG Arena. The sealed ranking agrees with six independent expert reviewers roughly as much as those reviewers agree with one another. Evaluation against realized outcomes is committed to a follow-on note, whatever it shows.
