Erdős' conjecture on arithmetic progressions, often incorrectly referred to as the Erdős–Turán conjecture (see also Erdős–Turán conjecture on additive bases), is a conjecture in additive combinatorics due to Paul Erdős. It states that if the sum of the reciprocals of the members of a set A of positive integers diverges, then A contains arbitrarily long arithmetic progressions.
Formally, if
(i.e. A is a large set) then A contains arithmetic progressions of any given length.
If true, the theorem would generalize Szemerédi's theorem.
Erdős offered a prize of US$3000 for a proof of this conjecture at the time.[1] The problem is currently worth US$5000.[2]
The Green–Tao theorem on arithmetic progressions in the primes is a special case of this conjecture.