%I #48 Feb 17 2020 16:56:41
%S 1,2,4,6,12,18,28,44,88,100,152,240,370,556,882,750,1500,2250,2784,
%T 4284,6438,6062,9526,14856,22944,26164,39528,35122,54800,80940,81326,
%U 122422,244844,234934,356154,309068,388042,589796,900000,813466,1212450,1837030
%N Number of sequences of length n over {1, -1} with Erdős discrepancy <= 2.
%C The Erdős discrepancy of sequence s is defined to be the maximum of the absolute value of s(d) + s(2d) + ... + s(kd) over all k, d such that kd <= n.
%C a(n) = 0 for all n > A237695(2). - _Rainer Rosenthal_, Feb 17 2020
%H Ehit Dinesh Agarwal, <a href="/A181740/b181740.txt">Table of n, a(n) for n = 0..60</a>
%H Boris Konev and Alexei Lisitsa, <a href="http://arxiv.org/abs/1402.2184">A SAT attack on the Erdős Discrepancy Conjecture</a>, arXiv:1402.2184 [cs.DM], 2014.
%H Michael Nielsen, <a href="http://michaelnielsen.org/polymath1/index.php?title=The_Erd%C5%91s_discrepancy_problem">Erdős discrepancy problem</a>
%e For n = 3 the only sequences omitted are 1 1 1 and -1 -1 -1, so a(3) = 6.
%Y Cf. A237695.
%K nonn
%O 0,2
%A _Jeffrey Shallit_, Nov 08 2010
%E a(30)-a(41) from _Allan C. Wechsler_, Sep 19 2012
%E Title changed by _Rainer Rosenthal_, Feb 17 2020