Two remarks on even and oddtown problems

TL;DR

This paper investigates extensions of even and odd town problems, providing new bounds and stability results for k-wise eventowns and defect versions of ℓ-oddtown problems, expanding understanding of these combinatorial set systems.

Contribution

It introduces a unique extremal configuration for k-wise eventowns with k ≥ 3 and improves bounds for the defect version of ℓ-oddtown problems, extending prior results.

Findings

Unique extremal configuration for k-wise eventowns when k ≥ 3

Stability results for k-wise eventowns

Improved bounds for defect ℓ-oddtown problems

Abstract

A family $\mathcal A$ of subsets of an $n$-element set is called an eventown (resp. oddtown) if all its sets have even (resp. odd) size and all pairwise intersections have even size. Using tools from linear algebra, it was shown by Berlekamp and Graver that the maximum size of an eventown is $2^{\left\lfloor n/2\right\rfloor}$. On the other hand (somewhat surprisingly), it was proven by Berlekamp, that oddtowns have size at most $n$. Over the last four decades, many extensions of this even/oddtown problem have been studied. In this paper we present new results on two such extensions. First, extending a result of Vu, we show that a $k$-wise eventown (i.e., intersections of $k$ sets are even) has for $k \geq 3$ a unique extremal configuration and obtain a stability result for this problem. Next we improve some known bounds for the defect version of an $\ell$-oddtown problem. In this problem we consider sets of size $\not\equiv 0 \pmod \ell$ where $\ell$ is a prime number $\ell$ (not necessarily $2$) and allow a few pairwise intersections to also have size $\not\equiv 0 \pmod \ell$.