Team games, hypergraph spaces, and projective Boolean algebras

Abstract

We modify the game Fuchino, Koppelberg, and Shelah used to characterize the $\kappa$-Freese-Nation property for a given Boolean algebra $A$, replacing players I and II each with a team of $n$ players with limited information. We show that $A$ is tightly $\kappa$-filtered exactly when team II has a winning strategy for every finite team size. Case $\kappa=\aleph_0$ characterizes projective Boolean algebras and, hence, Dugundji spaces. In terms of the open-open game of Daniels, Kunen, and Zhou, this characterization is a team version of very I-favorable. We similarly characterize Cohen algebras in terms of a team version of I-favorability. If $A$ is the clopen algebra of the space of $n$-uniform hypergraphs on $\kappa^{+n}$ that avoid copies of $[n+1]^n$, then team II has a winning strategy for our modified FKS game for team size $n-1$ but not $n$. For $n\geq 3$, this algebra also answers a question of Geschke when combined with a locally ${<}\kappa$-sized characterization of tightly $\kappa$-filtered Boolean algebras that we prove. Case $\kappa=\aleph_0$ includes a locally finite characterization of projective Boolean algebras.