This is a borderline paper. After the rebuttal and discussions, R1 has turned more positive about the paper, while R3 still thinks the paper is not ready for publication at NeurIPS. The paper studies the strategy space of real-world games from an interesting and novel perspective. That said, while the paper teases at how the results might produce more concrete impacts, it is quite hard to predict whether any of the follow-up directions would be fruitful at all, and it would be better if the authors take one of the possible directions and showcase the usefulness of the findings in this paper. After discussions among the reviewers and the meta-reviewers, we decide to accept the paper given that its novel perspectives may spark interesting discussions and new ideas in the community.