In set theory without the axiom of regularity, we consider a game in which two players choose in turn an element of a given set, an element of this element, etc.; a player wins if its adversary cannot make any next move. Sets that are winning, i.e. have a winning strategy for a player, form a natural hierarchy with levels indexed by ordinals. We show that the class of hereditarily winning sets is an inner model containing all well-founded sets, and that all four possible relationships between the universe, the class of hereditarily winning sets, and the class of well-founded sets are consistent. We describe classes of ordinals for which it is consistent that winning sets without minimal elements are exactly in the levels indexed by ordinals of this class. For consistency results, we propose a new method for getting non-well-founded models. Finally, we establish a probability result by showing that on hereditarily finite well-founded sets the first player wins almost always.
We present a version of the Banach-Mazur game, where open sets are replaced by elements of a fixed partially ordered set. We show how to apply it in the theory of Fraisse limits and beyond, obtaining simple proofs of universality of certain objects and classes.
We propose a new, game-theoretic, approach to the idealized forcing, in terms of fusion games. This generalizes the classical approach to the Sacks and the Miller forcing. For definable ($\mathbfΠ^1_1$ on $\mathbfΣ^1_1) $σ$-ideals we show that if a $σ$-ideal is generated by closed sets, then it is generated by closed sets in all forcing extensions. We also prove an infinite-dimensional version of the Solecki dichotomy for analytic sets. Among examples, we investigate the $σ$-ideal $\E$ generated by closed null sets and $σ$-ideals connected with not piecewise continuous functions.
This work extends the present author's computational game semantics of Martin-Löf type theory to the cumulative hierarchy of universes. This extension completes game semantics of all standard types of Martin-Löf type theory for the first time in the 30 years history of modern game semantics. As a result, the powerful combinatorial reasoning of game semantics becomes available for the study of universes and types generated by them. A main challenge in achieving game semantics of universes comes from a conflict between identity types and universes: Naive game semantics of the encoding of an iden…
We prove a number of results on the determinacy of $σ$-projective sets of reals, i.e., those belonging to the smallest pointclass containing the open sets and closed under complements, countable unions, and projections. We first prove the equivalence between $σ$-projective determinacy and the determinacy of certain classes of games of variable length ${<}ω^2$ (Theorem 2.4). We then give an elementary proof of the determinacy of $σ$-projective sets from optimal large-cardinal hypotheses (Theorem 4.4). Finally, we show how to generalize the proof to obtain proofs of the determinacy of $σ$-projec…
Schmidt's game is a powerful tool for studying properties of certain sets which arise in Diophantine approximation theory, number theory, and dynamics. Recently, many new results have been proven using this game. In this paper we address determinacy and indeterminacy questions regarding Schmidt's game and its variations, as well as more general games played on complete metric spaces (e.g. fractals). We show that except for certain exceptional cases, these games are undetermined on Bernstein sets.
We establish some similarities/analogies between uncountable cardinals or powersets and the class $V$ of all sets. They concern mainly the Boolean algebras ${\cal P}(κ)$, for a regular cardinal $κ$, and ${\cal C}(V)$ (the class of subclasses of the universe $V$), endowed with some ideals, especially the ideal $[κ]^{<κ}$ for ${\cal P}(κ)$, and the ideal of sets $V$ for ${\cal C}(V)$.
We show that a statement concerning the existence of winning strategies of limited memory in an infinite two-person topological game is equivalent to a weak version of the Singular Cardinals Hypothesis.
Given a target set $A\subseteq \mathbb{R}^d$ and a real number $β\in (0,1)$, McMullen introduced the notion of $A$ being an absolutely $β$-winning set. This involves a two player game which we call the $β$-McMullen game. We consider the version of this game in which the parameter $β$ is removed, which we call the no-$β$ McMullen game. More generally, we consider the game with respect to arbitrary norms on $\mathbb{R}^d$, and even more generally with respect to general convex sets. We show that for strictly convex sets in $\mathbb{R}^d$, polytopes in $\mathbb{R}^d$, and general convex sets in $…
The Hausdorff $δ$-dimension game was introduced by Das, Fishman, Simmons and {Urba{ń}ski} and shown to characterize sets in $\mathbb{R}^d$ having Hausdorff dimension $\leq δ$. We introduce a variation of this game which also characterizes Hausdorff dimension and for which we are able to prove an unfolding result similar to the basic unfolding property for the Banach-Mazur game for category. We use this to derive a number of consequences for Hausdorff dimension. We show that under $\mathsf{AD}$ any wellordered union of sets each of which has Hausdorff dimension $\leq δ$ has dimension $\leq δ$. …
Player ONE chooses a meager set and player TWO, a nowhere dense set per inning. They play $ω$ many innings. ONE's consecutive choices must form a (weakly) increasing sequence. TWO wins if the union of the chosen nowhere dense sets covers the union of the chosen meager sets. A strategy for TWO which depends on knowing only the uncovered part of the most recently chosen meager set is said to be a remainder strategy. Theorem (among others): TWO has a winning remainder strategy for this game played on the real line with its usual topology.
A set is nontypical in the Russell sense, if it belongs to a countable ordinal definable set. The class HNT of all hereditarily nontypical sets satisfies all axioms of ZF and the double inclusion HOD $\subseteq$ HNT $\subseteq$ V holds. Solving a problem recently proposed by Tzouvaras, a generic extension L$[a,x]$ of L, by two reals $a,x$, is presented in which L=HOD $\subsetneqq$ L$[a]$=HNT $\subsetneqq$ V=L$[a,x]$, so that HNT is a model of ZFC strictly between HOD and the universe.
It is shown that Matet's characterization of the Ramsey property relative to a selective co-ideal $\mathcal{H}$, in terms of games of Kastanas, still holds if we consider semiselectivity instead of selectivity. Moreover, we prove that a co-ideal $\mathcal{H}$ is semiselective if and only if Matet's game-theoretic characterization of the $\mathcal{H}$-Ramsey property holds. This lifts Kastanas's characterization of the classical Ramsey property to its optimal setting, from the point of view of the local Ramsey theory and gives a game-theoretic counterpart to a theorem of Farah \cite{far}, asser…
We prove a game theoretic dichotomy for $G_{δσ}$ sets of block sequences in vector spaces that extends, on the one hand, the block Ramsey theorem of W. T. Gowers proved for analytic sets of block sequences and, on the other hand, M. Davis' proof of $G_{δσ}$ determinacy.
We show the equivalence between the existence of winning strategies for $G_{δσ}$ (also called $Σ^{0}_{3}$) games in Cantor or Baire space, and the existence of functions generalized-recursive in a higher type-2 functional. (Such recursions are associated with certain transfinite computational models.) We show, inter alia, that the set of indices of convergent recursions in this sense is a complete $\Game Σ_{3}^{0}$ set: as paraphrase, the listing of those games at this level that are won by player I, essentially has the same information as the `halting problem' for this notion of recursion. Mo…
The notion of Aronszajn-null sets generalizes the notion of Lebesgue measure zero on the real line and the Euclidean space to infinite dimensional Banach spaces. We present a game-theoretic approach to Aronszajn-null sets, establish its basic properties, and discuss the ensuing open problems.
Recent work in set theory indicates that there are many different notions of 'set', each captured by a different collection of axioms, as proposed by J. Hamkins in [Ham11]. In this paper we strive to give one class theory that allows for a simultaneous consideration of all set theoretical universes and the relationships between them, eliminating the need for recourse 'outside the theory' when carrying out constructions like forcing etc. We also explore multiversal category theory, showing that we are finally free of questions about 'largeness' at each stage of the categorification process when…
We construct a univalent universe in the sense of Voevodsky in some suitable model categories for homotopy types (obtained from Grothendieck's theory of test categories). In practice, this means for instance that, appart from the homotopy theory of simplicial sets, intensional type theory with the univalent axiom can be interpreted in the homotopy theory of cubical sets (with connections or not), or of Joyal's cellular sets.
This paper deals with different concepts for characterizing the size of mathematical objects. A game theoretic investigation and generalization of two size concepts, which can both be formulated in topological terms, is provided: the so called "Baire category" and the "$σ$-category". This is mainly done by means of (generalized) Banach-Mazur games using the Axiom of determinacy (the inconsistency of AC and AD is reflected in the beginning and a weaker form of AC is chosen for proofs). Analogue versions of Cantor-Bendixson and Heine-Borel are proofed as well as some definability results. Such s…
Some decidable winning conditions of arbitrarily high finite Borel complexity for games on finite graphs or on pushdown graphs have been recently presented by O. Serre in [ Games with Winning Conditions of High Borel Complexity, in the Proceedings of the International Conference ICALP 2004, LNCS, Volume 3142, p. 1150-1162 ]. We answer in this paper several questions which were raised by Serre in the above cited paper. We first show that, for every positive integer n, the class C_n(A), which arises in the definition of decidable winning conditions, is included in the class of non-ambiguous cont…
