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 there are two infinite sets and such that agrees with the set of primes up to finitely many exceptions.
Asks whether the largest integer that some -element set with cannot write as a sum of its elements, repetition allowed, is asymptotically .
Asks which -element set with leaves the most integers that are not sums of its elements, repetition allowed, and whether it is .
Asks for the largest integer that is not a nonnegative integer combination of , for not a prime power.
Asks how many of the partial products of integers can be squares, and whether more than can be for every .
Asks whether every finite coloring of the integers has two distinct integers of one color whose sum is a square, or a th power. Khalfalah and Szemerédi proved it for the values of every non-constant polynomial that takes an even value.
Asks whether , where counts the consecutive pairs of an infinite with , and how large can be; both answered by Erdős and Szemerédi in 1980.
Asks for the size of the largest with for all , and whether Erdős's construction attains it. Chen showed the size is , and Chen and Dai showed the construction is not optimal for infinitely many .
Asks whether, when grows faster than , the sum of over pairs in , divided by the square of that sum, must tend to infinity; disproved by Tao, whose construction also gives the optimal growth threshold.
Asks for the size of when , whether it can be arbitrarily large, and whether it is at most for large .
Asks whether, for every infinite and every , the ratio of to is unbounded, where counts the elements of dividing .
Asks for the growth rate of the density of integers with a divisor in , and whether those with exactly one such divisor have density .
Asks how large a family of subsets of can be when no member is the union of two other distinct members.
Asks whether, for every , almost all have fewer than dyadic intervals containing a divisor of , where is the number of divisors.
Asks whether, for every , almost all have fewer than pairs of divisors , where is the number of divisors.
Asks whether some gives infinitely many such that every prime divides ; yes, by a 2026 AI construction erdosproblems.com accepted. Erdős's 1979 form, which asks the opposite, has the answer no.
Asks to estimate , the largest such that no integer strictly between and is composed only of primes dividing .
Asks whether every lacunary sequence () has an irrational with not dense in ; proved by Pollington and de Mathan. Peres and Schlag give a separation of order , and the erdosproblems.com wording is trivially true.
Asks whether the largest number of points in a disc of radius with all pairwise distances at least from the integers is , and even below ; proved by Sárközy (the first bound) and Konyagin (the sharp exponent ).
Asks whether, for some , the largest number of points in a disc of radius with all pairwise distances at least from the integers tends to infinity with . erdosproblems.com misprints the limit variable; Sárközy's power lower bound proves the modified Statement.
Asks whether converges over the integers that are sums of distinct proper divisors of themselves while no proper divisor of has that property.
Asks whether some finite set of primes grows without bound when one repeatedly adds every prime that is a sum of three distinct members; yes, by Vinogradov's three-primes theorem, an observation erdosproblems.com accepted.
Asks whether the positive integers can be permuted so that every two consecutive terms sum to a prime; yes, by an unpublished construction of Odlyzko reported by Erdős and Graham in 1980 and accepted by erdosproblems.com.