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.
Estimates the least size of a random subset of an abelian group of order whose subset sums hit every element nearly equally often, and asks whether .
Asks whether, for each , every residue modulo every prime is a sum of boundedly many inverses of integers up to ; proved by Shparlinski, Croot and Glibichuk, whose bound has order .
Asks whether every dense subset of has a -term arithmetic progression whose common difference lies in , for every large enough set .
Asks for an estimate of , the supremum of over finite families of pairwise disjoint congruence classes with distinct moduli . The supremum is known to logarithmic scale.
Asks whether, for and a positive non-decreasing , the set of with exactly representations as a sum of two elements of always has lower density and upper density below some .
Asks how fast can tend to infinity for a set of infinite measure in which is never an integer for distinct .
Asks whether every primitive set , in which no element divides another, has as .
Asks whether, for every of positive measure and almost every , each large enough integer has for some integer .
Asks whether every two-coloring of admits an infinite set all of whose sums of products of distinct members share one color. The answer is no, by a 1995 theorem of Smith, as the erdosproblems.com thread deduced in 2026.
Asks whether some ensures that, for any primes , removing residue classes modulo each leaves at most integers .
Asks how large the larger upper logarithmic density of the two subset-sum sets must be when is split into two classes. Conlon, Fox and Pham showed that the minimum is .
Asks whether, for every starting value and gap bound , any long enough integer sequence starting at with gaps at most has two intervals with equal sums. The answer is yes, by Hegyvári's 1986 Theorem 3, with an explicit bound.
Asks whether integers must be equal when, for every , the primes dividing are the same as those dividing .
Asks whether some constant ensures that every nonconstant polynomial with and all roots on the unit circle has a path of length at most from to the unit circle lying in except at .
Asks whether the largest transitive subtournament that every tournament on vertices must contain has vertices. This first fails at and fails for infinitely many , and the exact value is known for and .
Asks whether a sequence of positive lower logarithmic density contains a divisibility chain whose count below , divided by , has at least that of divided by ; answered yes in 2026 by Alexeev and seven coauthors.
Asks whether when the cardinals strictly increase and the first exceeds .
Asks whether for every singular cardinal such that and are both -inaccessible.