Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Campos 2023 exponential improvement diagonal ramsey
theorem_1_1: The first exponential improvement of the Erdős–Szekeres upper bound on the diagonal Ramsey number, with the two explicit values of ε the paper gives in prose.
M. Campos, S. Griffiths, R. Morris and J. Sahasrabudhe, An exponential improvement for diagonal Ramsey, Annals of Mathematics (2) 203 (2026), no. 3, 869--932; DOI 10.4007/annals.2026.203.3.4 (Crossref record, issued May 2026; the page range is the one printed in the bibliography of Gupta, Ndiaye, Norin and Wei). Preprint arXiv:2303.09521 (v1 16 March 2023; v2 4 August 2025).
The copy read for this card is arXiv:2303.09521v2 [math.CO] 4 Aug 2025, 59 pages, with a text layer; printed page PDF page. The page numbers below are the arXiv pages and the journal text has not been compared. Pages 1--2, 42, 45 and 48 were read on rendered page images. The arXiv record names arXiv's non-exclusive distribution license (arXiv:2303.09521), every other right reserved.
Read status: claims checked for Theorem 1.1 and the prose giving the two values of (p. 2), Theorem 1.2 (p. 2) and Theorems 13.1, 13.9 and 14.1 (pp. 42, 45 and 48), read clause by clause on the page images; no proof was read.
Theorem 1.1 proves that for some constant and all sufficiently large , the first exponential improvement on the Erdős--Szekeres bound of 1935. The paper gives two proofs and says in prose (p. 2) that the first gives and the second ; neither value appears in a numbered statement. Theorem 1.2 gives the off-diagonal analog for , an exponential gain, when , over the previous best bound for that range, , reached through the improvements of Rödl, Thomason, Conlon and Sah. "The precise bounds we prove in this paper are stated in Theorems 13.1, 13.9 and 14.1" (p. 2): for every sufficiently large with (Theorem 13.1, p. 42), for (Theorem 13.9, p. 45) and for all with (Theorem 14.1, p. 48); only the last range reaches the diagonal, and Section 14 says that its proof "also provides a second, somewhat different proof of Theorem 1.1". The method is a book-algorithm argument that builds large monochromatic books inside a coloring and controls the density of red edges between two vertex sets, without the quasirandomness of the Thomason--Conlon approach, so it escapes that approach's barrier. For Erdős problem 77, which asks for , Theorem 1.1 gives ; the paper states no interval for the limit (the lower end comes from Erdős's 1947 bound , which the introduction quotes on p. 1), and it notes the subsequent optimization by Gupta, Ndiaye, Norin and Wei to (p. 2).
Contents
- Introduction (pp. 1--2): the history from Ramsey and Erdős--Szekeres through Thomason, Conlon and Sah's ; Erdős's 1947 bound and Spencer's factor of ; the Erdős--Szekeres bound (1) and its improvements, to when and, for all and , by a polylogarithmic factor in unpublished work of Rödl (footnote 1).
- Theorem 1.1 (p. 2): for some constant once is large; the two proofs give and (prose).
- Theorem 1.2 (p. 2): for some constant , whenever .
- Theorems 13.1 (p. 42), 13.9 (p. 45) and 14.1 (p. 48): the explicit bounds (for , large), (for ) and (for all ) times .
Compiled scope
Pages 1--2, 42, 45 and 48 were read on the page images; the proofs (Sections 2--14) were not read. Nothing here is independently reviewed.
Source: https://arxiv.org/abs/2303.09521.
Bears on. #77: Theorem 1.1 is the first bound placing strictly below ; the existence and the value of the limit are untouched.
No file of this source is held: no license on record permits its redistribution, and the card cites the edition it names above.