Constructing All Birthday 3 Games as Digraphs
Alexander Clow, Alfie Davies, Neil Anderson McKay
TL;DR
This work determines the minimal vertex count needed to realise all Day-3 digraph placement values, proving $F(3)=8$ and thus that every Day-3 value has a representation on at most $8$ vertices. The authors blend exhaustive enumeration of digraphs on up to $7$ vertices with large-scale random searches on $8$-vertex graphs to certify all Day-3 values, and supply a combinatorial lower bound via a symmetric Day-3 value $\mathcal{Z}$ to establish $f(\mathcal{Z})\ge 8$. They also derive improved bounds for higher birthdays, showing $F(4)\in[11,2087]$ and $F(5)\le 10^{187.63}$ (with a substantial lower bound $F(5)\ge 2.58\times 10^{12}$), by leveraging anti-chain bounds and $D(n)$-based counts of coloured digraphs. Together, these results deepen understanding of the vertex-complexity needed to realize all game values born by days $4$ and $5$ and motivate future work on universal digraph-placement representations and the structure of Day-3 values.
Abstract
Recently, Clow and McKay proved that the Digraph Placement ruleset is universal for normal play: for all normal play combinatorial games $X$, there is a Digraph Placement game $G$ with $G=X$. Clow and McKay also showed that the 22 game values born by day 2 correspond to Digraph Placement games with at most 4 vertices. This bound is best possible. We extend this work using a combination of exhaustive and random searches to demonstrate all 1474 values born by day 3 correspond to Digraph Placement games on at most 8 vertices. We provide a combinatorial proof that this bound is best possible. We conclude by giving improved bounds on the number of vertices required to construct all game values born by days 4 and 5.
