Very I-favorable spaces (Q554400): Difference between revisions

The classes of I-favorable and very I-favorable spaces were introduced in the 1990's in [\textit{P. Daniels, K. Kunen} and \textit{H. Zhou}, Fundam. Math. 145, No.~3, 205--220 (1994; Zbl 0811.54008)]. The former class was given by a game-theoretical definition: a space \(X\) is called I-favorable if player I has a winning strategy for a so-called open-open game for \(X\); this notion can be described alternatively by a condition on a certain set of countable families of open sets; the notion of very I-favorable space is given by the same condition on an adequate subset of the previous set of countable families of open sets. The main results in this work are the characterizations, by categorical-theoretic concepts, of two subclasses of the very I-favorable spaces: the subclass of Hausdorff I-very favorable spaces and the subclass of completely regular I-very favorable spaces with respect to the co-zero sets. In both cases, a space is in the subclass iff it can be embedded in a projective limit of an adequate inverse system of spaces such that the restriction of the canonical projections of the limit space are onto. As a consequence, if the image of a compact \(T_2\) very I-favorable space under a continuous map is perfectly \(\kappa\)-normal, then \(X\) is \(\kappa\)-metrizable: this entails previous results in the literature on \(\kappa\)-metrizable compact \(T_2\) spaces.