On the Hardness of $4$-to-$1$ Games with Perfect Completeness

Abstract: We prove that for all $\varepsilon>0$, there exists a positive integer $k$ such that given a $4$-to-$1$ game $Ψ$ with alphabet size at most $k$, it is $\mathbf{NP}$-hard to distinguish between the case that $\mathrm{val}(Ψ)=1$ and the case that $\mathrm{val}(Ψ)\leq \varepsilon$. This confirms the $4$-to-$1$ Games Conjecture from [Khot, \textit{CCC 2002}]. Previously, the best known result, due to [Dinur, Khot, Kindler, Minzer, Safra], established the almost-perfect completeness version (but applied to the stricter problem of $2$-to-$1$ games). Using results from the literature, we get the following implications: (1) for all $k\in \mathbb{N}$, given a $3$-colorable graph $G$, it is $\mathbf{NP}$-hard to find a proper $k$-coloring; (2) for all $δ>0$, given a $2$-colorable $3$-uniform hypergraph $G$, it is $\mathbf{NP}$-hard to find in it an independent set containing at least $δ$ fraction of the vertices. Our proof is a three-step construction that builds on the two-step framework of [Dinur, Khot, Kindler, Minzer, Safra]. In the outer-PCP step, we use quadratic equations to gain perfect completeness. We then construct a new middle PCP that performs low-rank tests while preserving a key covering property. Finally, we construct a new inner PCP based on a tensor of the standard Grassmann encoding with its low-rank variant due to [Golowich, FOCS 2023].