English

Hanf numbers for poset games

Combinatorics 2026-01-07 v3 Logic

Abstract

Given two well partial orders (P;P)(P;\leq_P) and (T;T)(T;\leq_T), each with a minimum element, we study the following question: which player has a winning strategy for Chomp on the poset (P×T;P×T)(P\times T;\leq_{P\times T})? Here, (P×T;P×T)(P\times T;\leq_{P\times T}) denotes the poset obtained as the Cartesian product of PP and TT, equipped with the corresponding lexicographic order. The answer to this very natural question depends strongly on the specific choice of (P;P)(P;\leq_P) and (T;T)(T;\leq_T). For this reason, we restrict our attention to classes of posets given by powers of a fixed poset: {(Pσ;Pσ)σOrd}\{(P^\sigma;\leq_{P^\sigma})\mid \sigma\in\mathrm{Ord}\}. A fundamental fact about these classes of structures is that, if the second player does not have a winning strategy for all the posets in {(Pσ;Pσ)σOrd}\{(P^\sigma;\leq_{P^\sigma})\mid \sigma\in\mathrm{Ord}\}, there exists an ordinal ξ\xi such that the second player has a winning strategy on (Pξ;Pξ)(P^{\xi};\leq_{P^{\xi}}) but not on (Pγ;Pγ)(P^{\gamma};\leq_{P^{\gamma}}) for all γξ+1\gamma\geq\xi+1. Determining the corresponding ordinal for this Hanf number-style property constitutes the main objective of this work. Inspired by results of Garc\'ia-Marco and Knauer, we focus on classes of posets with a purely algebraic definition. These posets arise from submonoids (with respect to the natural sum, or Hessenberg sum) of ordinals of the form ωσ\omega^\sigma and are generated by sets of ordinals. In the process, we provide a test to determine whether a finite set Γ\Gamma of ordinals indeed yields well partial orders, and, using set-theoretic techniques, we establish an upper bound for the ordinal ξ\xi: if Γω\Gamma\subset\omega, then ξ<ω1\xi<\omega_1, and otherwise ξ<Γ+\xi<|\bigcup\Gamma|^+.

Cite

@article{arxiv.2504.07317,
  title  = {Hanf numbers for poset games},
  author = {Fabián Rivero Herrera},
  journal= {arXiv preprint arXiv:2504.07317},
  year   = {2026}
}

Comments

We make minor corrections and present the results within a more general framework

R2 v1 2026-06-28T22:52:59.757Z