Problems
Every problem posed by Paul Erdős, with its status, references, discussion and proof claims.
Every problem posed by Paul Erdős, with its status, references, discussion and proof claims.
618 of 1,221 problems match
Every problem posed by Paul Erdős, with its status, references, discussion and proof claims.
Asks whether , where is the least with for all . Elliott (1969) proved it for the two-sided threshold, with the one-sided form covered only by Elliott's remark.
Asks whether can be -colored so that every monochromatic -term arithmetic progression starting at has for all .
For a sequence , asks whether must be unbounded in , and whether is possible.
Asks whether the least spherical cap discrepancy of an -point set on the unit sphere tends to infinity with .
Asks whether the discrepancy of an infinite plane sequence, measured against circles of radius and their area, is always unbounded, and how fast it grows.
Asks whether the arguments of a polynomial's roots are equidistributed with error , where is the number of nonzero coefficients and is a ratio of coefficient sizes.
Asks whether the -point sets on the unit sphere that maximize the product of all pairwise distances have spherical cap discrepancy .
Asks whether, for every increasing integer sequence and almost every , the discrepancy of satisfies .
Asks whether the sequence counting the independent sets of each size in a tree or forest is unimodal.
Asks whether, for almost every , the fractional parts visit every measurable set with frequency .
Asks whether, for every , the sequence over the primes fails to be well distributed.
Asks whether, for irrational , an interval with bounded discrepancy along the fractional parts must have length for some integer . The erdosproblems.com wording, which asks this of both endpoints, is false.
Asks whether, for every , almost every satisfies for infinitely many coprime if and only if .
Asks whether some infinite integer sequence has , where counts the fractions with whose denominator in lowest terms differs from every earlier .
Asks whether the measure of the set of with for some reduced with tends to a limit as , and what that limit is explicitly.
Asks whether has an asymptotic distribution function over as .
Asks to estimate , the largest such that Farey fractions of order at most places apart are similarly ordered, and whether . Marked solved on Cipollini's 2026 preprint, which gives , matching van Doorn's upper bound.
Asks whether every graph of girth at least five has an acyclic orientation that stays acyclic when any one edge is reversed. False, by Nešetřil and Rödl's graphs of large girth with a monotone cycle under every ordering of the vertices.
Asks for the fewest edges in a graph of dimension , where a graph's dimension is the least such that it embeds in with every edge a unit segment. The answer is nine, attained only by up to isolated vertices (House; Chaffee and Noble).
Asks whether every graph with edges has a -free subgraph with edges. True (Conlon, Fox and Sudakov); Bollobás and Erdős's first form, with , fails by Folkman's example.
Asks whether, for each , every graph on vertices with edges, , has at least edge-disjoint triangles. Proved by Győri (1988), a paper known only through attestation.
Asks whether every graph on vertices with edges, , has at least triangles. This is the Erdős-Rademacher conjecture, proved in full by Lovász and Simonovits.
Asks to determine or estimate how large must be, in terms of , for edges to force a cycle on of the vertices. Settled by Woodall's 1972 Corollary 11.1, which covers every and the smaller too.
Asks whether, for each fixed , the ratio of consecutive off-diagonal Ramsey numbers tends to . Answered yes by a 2026 manuscript hosted by OpenAI, attributed to an internal model and accepted by erdosproblems.com.