%I #20 Aug 18 2022 21:11:29
%S 0,3,20,119,715,4523,30509,213343,1530983,11203550,83064263,620498643,
%T 4643259527,34592032908,254639722327,1832740718223,12680919388801,
%U 81678704122892,452951221016511,1574800035301944,8395299939524712,282240813012897282,4457697545906326118,58106920364272792945,693274802905577732102,7864635685729658131835
%N Similar to A057835 except using K * X / log(X), K=1.022.
%C This is an improvement over the classic X / log(X) approximation in the range many people work with.
%C pi(x), R(x), and li(x) are all asymptotically x/log x + x/log^2 x + O(x/log^3 x), so this approximation is good around exp(1/.022) ≈ 5 * 10^19. Asymptotically the best value for K would be 1. - _Charles R Greathouse IV_, Aug 18 2022
%F a(n) = abs(round(1.022*10^n/log(10^n)) - pi(10^n)). - _Charles R Greathouse IV_, Mar 22 2015
%F a(n) ~ 10^n/kn with k = 104.6629.... - _Charles R Greathouse IV_, Mar 22 2015
%e a(10)=11203550 via abs (455,052,511 - 443,848,961).
%t Table[ PrimePi[10^n] - Round[N[1.022*10^n/Log[10^n]]], {n, 23}] (* and absolute value thereof (orig entries 21-23 <0); courtesy of _Robert G. Wilson v_ *)
%o (PARI) a(n) = abs(round(1.022*10^n/log(10^n)) - primepi(10^n)) \\ _Charles R Greathouse IV_, Mar 22 2015
%Y Cf. A057835, A006880.
%K nonn
%O 1,2
%A _Bill McEachen_, Apr 23 2007