Low-rank matrices, tournaments, and symmetric designs

Abstract: Let $\mathbf{a} = (a_{i})<em>{i \geq 1}$ be a sequence in a field $\mathbb{F}$, and $$f \colon \mathbb{F} \times \mathbb{F} \to \mathbb{F}$$ be a function such that $f(a</em>{i},a_{i}) \neq 0$ for all $i \geq 1$. For any tournament $T$ over $[n]$, consider the $n \times n$ symmetric matrix $M_{T}(f, \mathbf{a})$ with zero diagonal whose $(i,j)$th entry (for $i &lt; j$) is $f(a_{i},a_{j})$ if $i \to j$ in $T$, and $f(a_{j},a_{i})$ if $j \to i$ in $T$. It is known (cf. Balachandran et al., Linear Algebra Appl. 658 (2023), 310-318) that if $T$ is a uniformly random tournament over $[n]$, then $$\operatorname{rank}(M_{T}(f,\mathbf{a})) \geq (\frac{1}{2}-o(1))n$$ with high probability when $\operatorname{char}(\mathbb{F}) \neq 2$ and $f$ is a linear function. In this paper, we investigate the other extremal question: how low can the ranks of such matrices be? We work with sequences $\mathbf{a}$ that take only two distinct values, so the rank of any such $n \times n$ matrix is at least $n/2$. First, we show that the rank of any such matrix depends on whether an associated bipartite graph has certain eigenvalues of high multiplicity. Using this, we show that if $f$ is linear, then there are $n \times n$ real matrices $M_{T}(f;\mathbf{a})$ of rank at most $\frac{n}{2} + O(1)$. For rational matrices, we show that for each $\varepsilon &gt; 0$ we can find a sequence $\mathbf{a}(\varepsilon)$ for which there are $n \times n$ matrices $M_{T}(f;\mathbf{a}(\varepsilon))$ of rank at most $(\frac{1}{2} + \varepsilon)n + O(1)$. These matrices are constructed from symmetric designs, and we also use them to produce bisection-closed families of size greater than $\lfloor 3n/2 \rfloor - 2$ for $n \leq 15$, which improves the previously best known bound (cf. Balachandran et al., Electron J. Combin. 26 (2019), #P2.40).