Triangle count bound via book number in dense graphs
Conjecture 1.1 · arXiv:1905.05312
Status partial high confidence
Conjecture 1.1 remains open in full generality. The authors themselves proved it for the boundary case b = n/6 and for the near-upper-boundary range 0.2495n ≤ b < n/4, leaving the bulk of the interval n/6 < b < 0.2495n unresolved. No subsequent paper resolving the remaining cases was found in the indexed literature as of May 2026.
Reviewer notes. The conjecture was partially proved in the source paper itself: it holds at b = n/6 and for 0.2495n ≤ b < n/4. Web searches and a Semantic Scholar citation query returned no follow-up paper resolving the remaining range n/6 < b < 0.2495n. The conjecture is classified as partial (proved in special cases by the authors) with high confidence that no full resolution has appeared.
Context
The authors introduce the 3-prism blow-up $S_{b,n}$ — formed by blowing up the 3-prism graph with four parts of size $b$ and two parts of size $\lfloor(n-4b)/2\rfloor$ and $\lceil(n-4b)/2\rceil$ — as the conjectured extremal construction. The conjecture implies an asymptotically tight bound on $t(n,b)$ for $n/6 \leq b \leq n/4 - \omega(1)$, fully addressing Mubayi's question about the $\gamma(\beta)$ dependency. A footnote notes the construction is essentially identical to one in an earlier arXiv version of Mubayi's paper.
Source paper
Books versus triangles at the extremal density
David Conlon, Jacob Fox, Benny Sudakov · 2019-10-20
https://arxiv.org/abs/1905.05312
PDF source