Hofstadter-Conway $10,000 sequence
You are encouraged to solve this task according to the task description, using any language you may know.
The definition of the sequence is colloquially described as:
- Starting with the list [1,1],
- Take the last number in the list so far: 1, I'll call it x.
- Count forward x places from the beginning of the list to find the first number to add (1)
- Count backward x places from the end of the list to find the second number to add (1)
- Add the two indexed numbers from the list and the result becomes the next number in the list (1+1)
- This would then produce [1,1,2] where 2 is the third element of the sequence.
Note that indexing for the description above starts from alternately the left and right ends of the list and starts from an index of one.
A less wordy description of the sequence is:
a(1)=a(2)=1 a(n)=a(a(n-1))+a(n-a(n-1))
The sequence begins:
1, 1, 2, 2, 3, 4, 4, 4, 5, ...
Interesting features of the sequence are that:
- a(n)/n tends to 0.5 as n grows towards infinity.
- a(n)/n where n is a power of 2 is 0.5
- For n>4 the maximal value of a(n)/n between successive powers of 2 decreases.
The sequence is so named because John Conway offered a prize of $10,000 to the first person who could find the first position, p in the sequence where
|a(n)/n| < 0.55 for all n > p.
It was later found that Hofstadter had also done prior work on the sequence.
The 'prize' was won quite quickly by Dr. Colin L. Mallows who proved the properties of the sequence and allowed him to find the value of n. (Which is much smaller than the 3,173,375,556. quoted in the NYT article)
The task is to:
- Create a routine to generate members of the Hofstadter-Conway $10,000 sequence.
- Use it to show the maxima of a(n)/n between successive powers of two up to 2**20
- As a stretch goal: Compute the value of n that would have won the prize and confirm it is true for n up to 2**20
References:
- Conways Challenge Sequence, Mallows' own account.
- Mathworld Article.
Ada
<lang Ada> -- Ada95 version with Ada.Text_IO; use Ada.Text_IO;
procedure Conway is
package Real_io is new Float_IO (Float); Maxrange : constant := 2 ** 20; type Sequence is array (Positive range <>) of Positive;
function HCseq (N : Positive) return Sequence is
A : Sequence (1 .. N);
begin
A (1) := 1;
if N > 1 then
A (2) := 1;
for K in 3 .. N loop
A (K) := A (A (K - 1)) + A (K - A (K - 1));
end loop;
end if;
return A;
end HCseq;
S : Sequence := HCseq (Maxrange); Ratio : array (1 .. Maxrange) of Float; Mallows : Positive; M : Natural := 0;
begin
for k in 1 .. Maxrange loop
Ratio (k) := Float (S (k)) / Float (k);
end loop;
for N in 1 .. 19 loop
declare
Max : Float := 0.0;
Where : Positive;
begin
for K in 2 ** N .. 2 ** (N + 1) loop
if Max < Ratio (K) then
Max := Ratio (K);
Where := K;
end if;
end loop;
if (M = 0 and Max < 0.55) then
M := N - 1;
end if;
Put
("Maximun of a(n)/n between 2^" &
Integer'Image (N) &
" and 2^" &
Integer'Image (N + 1) &
" was ");
Real_io.Put (Max, Fore => 1, Aft => 8, Exp => 0);
Put_Line (" at" & Integer'Image (Where));
end;
end loop;
-- Calculate Mallows number
for I in reverse 2 ** M .. 2 ** (M + 1) loop
if (Ratio (I) > 0.55) then
Mallows := I;
exit;
end if;
end loop;
Put_Line ("Mallows number" & Integer'Image (Mallows));
end Conway;
</lang> Sample output:
Maximun of a(n)/n between 2^ 1 and 2^ 2 was 0.66666669 at 3 Maximun of a(n)/n between 2^ 2 and 2^ 3 was 0.66666669 at 6 Maximun of a(n)/n between 2^ 3 and 2^ 4 was 0.63636363 at 11 Maximun of a(n)/n between 2^ 4 and 2^ 5 was 0.60869563 at 23 Maximun of a(n)/n between 2^ 5 and 2^ 6 was 0.59090906 at 44 Maximun of a(n)/n between 2^ 6 and 2^ 7 was 0.57608694 at 92 Maximun of a(n)/n between 2^ 7 and 2^ 8 was 0.56741571 at 178 Maximun of a(n)/n between 2^ 8 and 2^ 9 was 0.55945945 at 370 Maximun of a(n)/n between 2^ 9 and 2^ 10 was 0.55493742 at 719 Maximun of a(n)/n between 2^ 10 and 2^ 11 was 0.55010086 at 1487 Maximun of a(n)/n between 2^ 11 and 2^ 12 was 0.54746288 at 2897 Maximun of a(n)/n between 2^ 12 and 2^ 13 was 0.54414475 at 5969 Maximun of a(n)/n between 2^ 13 and 2^ 14 was 0.54244268 at 11651 Maximun of a(n)/n between 2^ 14 and 2^ 15 was 0.54007107 at 22223 Maximun of a(n)/n between 2^ 15 and 2^ 16 was 0.53878403 at 45083 Maximun of a(n)/n between 2^ 16 and 2^ 17 was 0.53704363 at 89516 Maximun of a(n)/n between 2^ 17 and 2^ 18 was 0.53602004 at 181385 Maximun of a(n)/n between 2^ 18 and 2^ 19 was 0.53464544 at 353683 Maximun of a(n)/n between 2^ 19 and 2^ 20 was 0.53377920 at 722589 Mallows number 1489
ALGOL 68
- note: This specimen retains the original C coding style.
<lang algol68>PROC do sqnc = (INT max)INT: (
[max]INT a list; INT v; INT k1 := 2; INT lg2 := 1;
INT nmax; LONG REAL amax := 0;
INT out mallows number;
v := a list[1] := a list[2] := 1;
FOR n FROM 3 TO max DO
v := a list[n] := a list[v] + a list[n-v];
IF amax < v/n THEN
amax := v/n;
nmax := n ;
IF v/n >= 0.55 THEN out mallows number := n FI
FI;
IF 2r0 = (BIN k1 AND BIN n) THEN
printf(($"Maximum between 2^"g(0)" and 2^"g(0)" is about "g(-8,6)" at "g(0)l$, lg2,lg2+1, amax, nmax));
amax := 0;
lg2 +:= 1
FI;
k1 := n
OD;
out mallows number
);
INT mallows number = do sqnc(2**20);
printf(($"You too might have won $1000 with an answer of n = "g(0)$, mallows number))</lang> Output:
Maximum between 2^1 and 2^2 is about 0.666667 at 3 Maximum between 2^2 and 2^3 is about 0.666667 at 6 Maximum between 2^3 and 2^4 is about 0.636364 at 11 Maximum between 2^4 and 2^5 is about 0.608696 at 23 Maximum between 2^5 and 2^6 is about 0.590909 at 44 Maximum between 2^6 and 2^7 is about 0.576087 at 92 Maximum between 2^7 and 2^8 is about 0.567416 at 178 Maximum between 2^8 and 2^9 is about 0.559459 at 370 Maximum between 2^9 and 2^10 is about 0.554937 at 719 Maximum between 2^10 and 2^11 is about 0.550101 at 1487 Maximum between 2^11 and 2^12 is about 0.547463 at 2897 Maximum between 2^12 and 2^13 is about 0.544145 at 5969 Maximum between 2^13 and 2^14 is about 0.542443 at 11651 Maximum between 2^14 and 2^15 is about 0.540071 at 22223 Maximum between 2^15 and 2^16 is about 0.538784 at 45083 Maximum between 2^16 and 2^17 is about 0.537044 at 89516 Maximum between 2^17 and 2^18 is about 0.536020 at 181385 Maximum between 2^18 and 2^19 is about 0.534645 at 353683 Maximum between 2^19 and 2^20 is about 0.533779 at 722589 You too might have won $1000 with an answer of n = 1487
AutoHotkey
<lang autohotkey>Progress, b2 w150 zh0 fs9, CreateLists ... CreateLists(2 ** (Max:=20))
Progress,, Find Maxima ... Loop, % Max - 1
msg .= "Maximum between 2^" A_Index " and 2^" A_Index + 1
. " is " GetMax(2 ** A_Index, 2 ** (A_Index + 1), n)
. " for n = " n "`n"
Progress,, Find Mallows Number ... Loop, % 2 ** Max
If (n_%A_Index% > 0.55)
MallowsNumber := A_Index
msg .= "Mallows Number = " MallowsNumber
Progress, Off MsgBox, %msg%
- ---------------------------------------------------------------------------
GetMax(a, b, ByRef Item) { ; return max value of a(n)/n between a and b
- ---------------------------------------------------------------------------
Loop {
IfGreater, a, %b%, Break
If (Maximum < n_%a%)
Maximum := n_%a%, Item := a
a++
}
Return, Maximum
}
- ---------------------------------------------------------------------------
CreateLists(Lenght) { ; Hofstadter-Conway sequences (using lookups)
- ---------------------------------------------------------------------------
; create the sequence a_%A_Index% [ a(n) ]
; and the sequence n_%A_Index% [ a(n)/n ]
;-----------------------------------------------------------------------
global
a_1 := a_2 := n_1 := 1, n_2 := 1 / 2
Loop, %Lenght% {
IfLess, A_Index, 3, Continue
n1 := A_Index - 1
an1 := a_%n1%
nan1 := A_Index - an1
a_%A_Index% := a_%an1% + a_%nan1%
n_%A_Index% := a_%A_Index% / A_Index
}
}</lang> Message box shows:
Maximum between 2^1 and 2^2 is 0.666667 for n = 3 Maximum between 2^2 and 2^3 is 0.666667 for n = 6 Maximum between 2^3 and 2^4 is 0.636364 for n = 11 Maximum between 2^4 and 2^5 is 0.608696 for n = 23 Maximum between 2^5 and 2^6 is 0.590909 for n = 44 Maximum between 2^6 and 2^7 is 0.576087 for n = 92 Maximum between 2^7 and 2^8 is 0.567416 for n = 178 Maximum between 2^8 and 2^9 is 0.559459 for n = 370 Maximum between 2^9 and 2^10 is 0.554937 for n = 719 Maximum between 2^10 and 2^11 is 0.550101 for n = 1487 Maximum between 2^11 and 2^12 is 0.547463 for n = 2897 Maximum between 2^12 and 2^13 is 0.544145 for n = 5969 Maximum between 2^13 and 2^14 is 0.542443 for n = 11651 Maximum between 2^14 and 2^15 is 0.540071 for n = 22223 Maximum between 2^15 and 2^16 is 0.538784 for n = 45083 Maximum between 2^16 and 2^17 is 0.537044 for n = 89516 Maximum between 2^17 and 2^18 is 0.536020 for n = 181385 Maximum between 2^18 and 2^19 is 0.534645 for n = 353683 Maximum between 2^19 and 2^20 is 0.533779 for n = 722589 Mallows Number = 1489
C
<lang C>#include <stdio.h>
- include <stdlib.h>
int a_list[1<<20 + 1];
int doSqnc( int m) {
int max_df = 0; int p2_max = 2; int v, n; int k1 = 2; int lg2 = 1; double amax = 0; a_list[0] = -50000; a_list[1] = a_list[2] = 1; v = a_list[2];
for (n=3; n <= m; n++) {
v = a_list[n] = a_list[v] + a_list[n-v];
if ( amax < v*1.0/n) amax = v*1.0/n;
if ( 0 == (k1&n)) {
printf("Maximum between 2^%d and 2^%d was %f\n", lg2,lg2+1, amax);
amax = 0;
lg2++;
}
k1 = n;
}
return 1;
}</lang> Results
Maximum between 2^1 and 2^2 was 0.666667 Maximum between 2^2 and 2^3 was 0.666667 Maximum between 2^3 and 2^4 was 0.636364 Maximum between 2^4 and 2^5 was 0.608696 .... Maximum between 2^18 and 2^19 was 0.534645 Maximum between 2^19 and 2^20 was 0.533779
C++
<lang cpp>
- include <deque>
- include <iostream>
int hcseq(int n) {
static std::deque<int> seq(2, 1);
while (seq.size() < n)
{
int x = seq.back();
seq.push_back(seq[x-1] + seq[seq.size()-x]);
}
return seq[n-1];
}
int main() {
int pow2 = 1;
for (int i = 0; i < 20; ++i)
{
int pow2next = 2*pow2;
double max = 0;
for (int n = pow2; n < pow2next; ++n)
{
double anon = hcseq(n)/double(n);
if (anon > max)
max = anon;
}
std::cout << "maximum of a(n)/n between 2^" << i
<< " (" << pow2 << ") and 2^" << i+1
<< " (" << pow2next << ") is " << max << "\n";
pow2 = pow2next;
}
} </lang> Output:
maximum of a(n)/n between 2^0 (1) and 2^1 (2) is 1 maximum of a(n)/n between 2^1 (2) and 2^2 (4) is 0.666667 maximum of a(n)/n between 2^2 (4) and 2^3 (8) is 0.666667 maximum of a(n)/n between 2^3 (8) and 2^4 (16) is 0.636364 maximum of a(n)/n between 2^4 (16) and 2^5 (32) is 0.608696 maximum of a(n)/n between 2^5 (32) and 2^6 (64) is 0.590909 maximum of a(n)/n between 2^6 (64) and 2^7 (128) is 0.576087 maximum of a(n)/n between 2^7 (128) and 2^8 (256) is 0.567416 maximum of a(n)/n between 2^8 (256) and 2^9 (512) is 0.559459 maximum of a(n)/n between 2^9 (512) and 2^10 (1024) is 0.554937 maximum of a(n)/n between 2^10 (1024) and 2^11 (2048) is 0.550101 maximum of a(n)/n between 2^11 (2048) and 2^12 (4096) is 0.547463 maximum of a(n)/n between 2^12 (4096) and 2^13 (8192) is 0.544145 maximum of a(n)/n between 2^13 (8192) and 2^14 (16384) is 0.542443 maximum of a(n)/n between 2^14 (16384) and 2^15 (32768) is 0.540071 maximum of a(n)/n between 2^15 (32768) and 2^16 (65536) is 0.538784 maximum of a(n)/n between 2^16 (65536) and 2^17 (131072) is 0.537044 maximum of a(n)/n between 2^17 (131072) and 2^18 (262144) is 0.53602 maximum of a(n)/n between 2^18 (262144) and 2^19 (524288) is 0.534645 maximum of a(n)/n between 2^19 (524288) and 2^20 (1048576) is 0.533779
C#
<lang csharp> using System; using System.Linq;
namespace HofstadterConway {
class Program
{
static int[] GenHofstadterConway(int max)
{
int[] result = new int[max];
result[0]=result[1]=1;
for (int ix = 2; ix < max; ix++)
result[ix] = result[result[ix - 1] - 1] + result[ix - result[ix - 1]];
return result;
}
static void Main(string[] args)
{
double[] adiv = new double[1 << 20];
{
int[] a = GenHofstadterConway(1 << 20);
for (int i = 0; i < 1 << 20; i++)
adiv[i] = a[i] / (double)(i + 1);
}
for (int p = 2; p <= 20; p++)
{
var max = Enumerable.Range(
(1 << (p - 1)) - 1,
(1 << p) - (1 << (p - 1))
)
.Select(ix => new { I = ix + 1, A = adiv[ix] })
.OrderByDescending(x => x.A)
.First();
Console.WriteLine("Maximum from 2^{0} to 2^{1} is {2} at {3}",
p - 1, p, max.A, max.I);
}
Console.WriteLine("The winning number is {0}.",
Enumerable.Range(0, 1 << 20)
.Last(i => (adiv[i] > 0.55)) + 1
);
}
}
} </lang> Output:-
Maximum from 2^1 to 2^2 is 0.666666666666667 at 3 Maximum from 2^2 to 2^3 is 0.666666666666667 at 6 Maximum from 2^3 to 2^4 is 0.636363636363636 at 11 Maximum from 2^4 to 2^5 is 0.608695652173913 at 23 Maximum from 2^5 to 2^6 is 0.590909090909091 at 44 Maximum from 2^6 to 2^7 is 0.576086956521739 at 92 Maximum from 2^7 to 2^8 is 0.567415730337079 at 178 Maximum from 2^8 to 2^9 is 0.559459459459459 at 370 Maximum from 2^9 to 2^10 is 0.554937413073713 at 719 Maximum from 2^10 to 2^11 is 0.550100874243443 at 1487 Maximum from 2^11 to 2^12 is 0.547462892647566 at 2897 Maximum from 2^12 to 2^13 is 0.544144747863964 at 5969 Maximum from 2^13 to 2^14 is 0.542442708780362 at 11651 Maximum from 2^14 to 2^15 is 0.540071097511587 at 22223 Maximum from 2^15 to 2^16 is 0.538784020584256 at 45083 Maximum from 2^16 to 2^17 is 0.537043656999866 at 89516 Maximum from 2^17 to 2^18 is 0.536020067811561 at 181385 Maximum from 2^18 to 2^19 is 0.534645431078112 at 353683 Maximum from 2^19 to 2^20 is 0.533779229963368 at 722589 The winning number is 1489.
Clojure
<lang Clojure>(use 'clojure.contrib.math) ; for expt
- A literal transcription of the definition, with memoize doing the heavy lifting
(def conway
(memoize
(fn [x]
(if (< x 3)
1
(+ (conway (conway (dec x)))
(conway (- x (conway (dec x)))))))))
(def N (drop 1 (range))) ; natural numbers
- This is enough to compute them all. The rest is grouping and max-finding
(def all-conways (map conway N))
- All the powers of two
(def pow2 (map #(expt 2 %) N))
- Find the lowest power of two > n
(defn lowest-pow-higher [n]
(some #(and (> % n) %) pow2))
- Split the natural numbers into groups at each power of two
(def groups (partition-by lowest-pow-higher N))
- The conway numbers of each number in the group
(def C (map #(map conway %) groups))
- Each conway number divided by its index
(def ratios (map #(map / %1 %2) C groups))
- The largest value in each group of ratios
(def maxima (map #(apply max %) ratios))
(take 4 maxima) ; no rounding errors: still using ratios
- yields (1 2/3 2/3 7/11)
(def first-20 (take 20 maxima)) (apply >= first-20) ; yields true - the sequence is decreasing (map double first-20) ; to get them in decimal form</lang>
Common Lisp
<lang lisp>(defparameter *hof-con*
(make-array '(2) :initial-contents '(1 1) :adjustable t
:element-type 'integer :fill-pointer 2))
(defparameter *hof-con-ratios*
(make-array '(2) :initial-contents '(1.0 0.5) :adjustable t
:element-type 'single-float :fill-pointer 2))
(defun hof-con (n)
(let ((l (length *hof-con*))) (if (<= n l) (aref *hof-con* (1- n))
(extend-hof-con-sequence l n))))
(defun extend-hof-con-sequence (l n)
(loop for i from l below n do
(let* ((x (aref *hof-con* (1- i)))
(hc (+ (aref *hof-con* (1- x)) (aref *hof-con* (- i x))))) (vector-push-extend hc *hof-con*) (vector-push-extend (/ hc (+ i 1.0)) *hof-con-ratios*)))
(aref *hof-con* (1- n)))
(defun max-in-array-range (arr id1 id2)
(let ((m 0) (id 0)) (loop for i from (1- id1) to (1- id2) do
(let ((n (aref arr i))) (if (> n m) (setq m n id i))))
(values m (1+ id))))
(defun maxima (po2)
(hof-con (expt 2 po2))
(loop for i from 1 below po2 do
(let ((id1 (expt 2 i)) (id2 (expt 2 (1+ i))))
(multiple-value-bind (m id) (max-in-array-range *hof-con-ratios* id1 id2) (format t "Local maximum in [~A .. ~A]: ~A at n = ~A~%" id1 id2 m id)))))
(defun mallows (po2)
(let ((n (expt 2 po2))) (hof-con n) (do ((i (1- n) (1- i)))
((> (aref *hof-con-ratios* i) 0.55) (+ i 1)))))</lang> Sample session:
ROSETTA> (maxima 20) Local maximum in [2 .. 4]: 0.6666667 at n = 3 Local maximum in [4 .. 8]: 0.6666667 at n = 6 Local maximum in [8 .. 16]: 0.6363636 at n = 11 Local maximum in [16 .. 32]: 0.6086956 at n = 23 Local maximum in [32 .. 64]: 0.59090906 at n = 44 Local maximum in [64 .. 128]: 0.57608694 at n = 92 Local maximum in [128 .. 256]: 0.5674157 at n = 178 Local maximum in [256 .. 512]: 0.55945945 at n = 370 Local maximum in [512 .. 1024]: 0.5549374 at n = 719 Local maximum in [1024 .. 2048]: 0.55010086 at n = 1487 Local maximum in [2048 .. 4096]: 0.5474629 at n = 2897 Local maximum in [4096 .. 8192]: 0.54414475 at n = 5969 Local maximum in [8192 .. 16384]: 0.5424427 at n = 11651 Local maximum in [16384 .. 32768]: 0.54007107 at n = 22223 Local maximum in [32768 .. 65536]: 0.538784 at n = 45083 Local maximum in [65536 .. 131072]: 0.53704363 at n = 89516 Local maximum in [131072 .. 262144]: 0.53602004 at n = 181385 Local maximum in [262144 .. 524288]: 0.53464544 at n = 353683 Local maximum in [524288 .. 1048576]: 0.5337792 at n = 722589 NIL ROSETTA> (mallows 20) 1489
D
D v.2 <lang D>import std.stdio: writefln;
void hofstadterConwaySequence(int m) {
auto alist = new int[m + 1]; alist[0 .. 2] = 1;
double amax = 0.0; auto v = alist[2]; int k1 = 2; int lg2 = 1;
foreach (n; 2 .. m + 1) {
v = alist[n] = alist[v] + alist[n - v];
if (amax < v * 1.0 / n)
amax = v * 1.0 / n;
if ((k1 & n) == 0) {
writefln("Maximum between 2^%d and 2^%d was %f", lg2, lg2 + 1, amax);
amax = 0;
lg2++;
}
k1 = n;
}
}
void main() {
hofstadterConwaySequence(1 << 20);
}</lang>
Output:
Maximum between 2^1 and 2^2 was 0.666667 Maximum between 2^2 and 2^3 was 0.666667 Maximum between 2^3 and 2^4 was 0.636364 Maximum between 2^4 and 2^5 was 0.608696 Maximum between 2^5 and 2^6 was 0.590909 Maximum between 2^6 and 2^7 was 0.576087 Maximum between 2^7 and 2^8 was 0.567416 Maximum between 2^8 and 2^9 was 0.559459 Maximum between 2^9 and 2^10 was 0.554937 Maximum between 2^10 and 2^11 was 0.550101 Maximum between 2^11 and 2^12 was 0.547463 Maximum between 2^12 and 2^13 was 0.544145 Maximum between 2^13 and 2^14 was 0.542443 Maximum between 2^14 and 2^15 was 0.540071 Maximum between 2^15 and 2^16 was 0.538784 Maximum between 2^16 and 2^17 was 0.537044 Maximum between 2^17 and 2^18 was 0.536020 Maximum between 2^18 and 2^19 was 0.534645 Maximum between 2^19 and 2^20 was 0.533779
Fortran
<lang Fortran> program conway
implicit none integer :: a(2**20) ! The sequence a(n) real :: b(2**20) ! The sequence a(n)/n real :: v ! Max value in the range [2*i, 2**(i+1)] integer :: nl(1) ! The location of v in the array b(n) integer :: i, N, first, second, last, m
! Populate a(n) and b(n)
a(1:2) = [1, 1]
b(1:2) = [1.0e0, 0.5e0]
N = 2
do i=1,2**20
last = a(N)
first = a(last)
second = a(N-last+1)
N = N+1
a(N:N) = first + second
b(N:N) = a(N:N)/real(N)
end do
! Calculate the max values in the logarithmic ranges
m = 0
do i=1,19
v = maxval(b(2**i:2**(i+1)))
nl = maxloc(b(2**i:2**(i+1)))
write(*,'(2(a,i0),a,f8.6,a,i0)') &
'Max. between 2**', i, &
' and 2**', (i+1), &
' is ', v, &
' at n = ', 2**i+nl(1)-1
if (m == 0 .and. v < 0.55e0) then
m = i-1
end if
end do
! Calculate Mallows number
do i=2**(m+1), 2**m,-1
if (b(i) > 0.55e0) then
exit
end if
end do
write(*,'(a,i0)') 'Mallows number = ',i
end program conway </lang>
Output:
Max. between 2**1 and 2**2 is 0.666667 at n = 3 Max. between 2**2 and 2**3 is 0.666667 at n = 6 Max. between 2**3 and 2**4 is 0.636364 at n = 11 Max. between 2**4 and 2**5 is 0.608696 at n = 23 Max. between 2**5 and 2**6 is 0.590909 at n = 44 Max. between 2**6 and 2**7 is 0.576087 at n = 92 Max. between 2**7 and 2**8 is 0.567416 at n = 178 Max. between 2**8 and 2**9 is 0.559459 at n = 370 Max. between 2**9 and 2**10 is 0.554937 at n = 719 Max. between 2**10 and 2**11 is 0.550101 at n = 1487 Max. between 2**11 and 2**12 is 0.547463 at n = 2897 Max. between 2**12 and 2**13 is 0.544145 at n = 5969 Max. between 2**13 and 2**14 is 0.542443 at n = 11651 Max. between 2**14 and 2**15 is 0.540071 at n = 22223 Max. between 2**15 and 2**16 is 0.538784 at n = 45083 Max. between 2**16 and 2**17 is 0.537044 at n = 89516 Max. between 2**17 and 2**18 is 0.536020 at n = 181385 Max. between 2**18 and 2**19 is 0.534645 at n = 353683 Max. between 2**19 and 2**20 is 0.533779 at n = 722589 Mallows number = 1489
F#
<lang fsharp>let a = ResizeArray[0; 1; 1] while a.Count <= (1 <<< 20) do
a.[a.[a.Count - 1]] + a.[a.Count - a.[a.Count - 1]] |> a.Add
for p = 1 to 19 do
Seq.max [|for i in 1 <<< p .. 1 <<< p+1 -> float a.[i] / float i|] |> printf "Maximum in %6d..%7d is %g\n" (1 <<< p) (1 <<< p+1)</lang>
Outputs:
Maximum in 2.. 4 is 0.666667 Maximum in 4.. 8 is 0.666667 Maximum in 8.. 16 is 0.636364 Maximum in 16.. 32 is 0.608696 Maximum in 32.. 64 is 0.590909 Maximum in 64.. 128 is 0.576087 Maximum in 128.. 256 is 0.567416 Maximum in 256.. 512 is 0.559459 Maximum in 512.. 1024 is 0.554937 Maximum in 1024.. 2048 is 0.550101 Maximum in 2048.. 4096 is 0.547463 Maximum in 4096.. 8192 is 0.544145 Maximum in 8192.. 16384 is 0.542443 Maximum in 16384.. 32768 is 0.540071 Maximum in 32768.. 65536 is 0.538784 Maximum in 65536.. 131072 is 0.537044 Maximum in 131072.. 262144 is 0.53602 Maximum in 262144.. 524288 is 0.534645 Maximum in 524288..1048576 is 0.533779
Go
<lang go>package main
import (
"fmt"
)
func main() {
a := []int{0, 1, 1} // ignore 0 element. work 1 based.
x := 1 // last number in list
n := 2 // index of last number in list = len(a)-1
mallow := 0
for p := 1; p < 20; p++ {
max := 0.
for nextPot := n*2; n < nextPot; {
n = len(a) // advance n
x = a[x]+a[n-x]
a = append(a, x)
f := float64(x)/float64(n)
if f > max {
max = f
}
if f >= .55 {
mallow = n
}
}
fmt.Printf("max between 2^%d and 2^%d was %f\n", p, p+1, max)
}
fmt.Println("winning number", mallow)
}</lang> Output:
max between 2^1 and 2^2 was 0.666667 max between 2^2 and 2^3 was 0.666667 max between 2^3 and 2^4 was 0.636364 max between 2^4 and 2^5 was 0.608696 max between 2^5 and 2^6 was 0.590909 max between 2^6 and 2^7 was 0.576087 max between 2^7 and 2^8 was 0.567416 max between 2^8 and 2^9 was 0.559459 max between 2^9 and 2^10 was 0.554937 max between 2^10 and 2^11 was 0.550101 max between 2^11 and 2^12 was 0.547463 max between 2^12 and 2^13 was 0.544145 max between 2^13 and 2^14 was 0.542443 max between 2^14 and 2^15 was 0.540071 max between 2^15 and 2^16 was 0.538784 max between 2^16 and 2^17 was 0.537044 max between 2^17 and 2^18 was 0.536020 max between 2^18 and 2^19 was 0.534645 max between 2^19 and 2^20 was 0.533779 winning number 1489
Haskell
<lang haskell>import Data.List import Data.Ord import Data.Array import Text.Printf
hc :: Int -> Array Int Int hc n = arr
where arr = array (1, n) $ (1, 1) : (2, 1) : map (f (arr!)) [3 .. n]
f a i = (i, a (a $ i - 1) + a (i - a (i - 1)))
printMaxima :: (Int, (Int, Double)) -> IO () printMaxima (n, (pos, m)) =
printf "Max between 2^%-2d and 2^%-2d is %1.5f at n = %6d\n"
n (n + 1) m pos
main = do
mapM_ printMaxima maxima printf "Mallows's number is %d\n" mallows where hca = hc $ 2^20 hc' n = fromIntegral (hca!n) / fromIntegral n maxima = zip [0..] $ map max powers max seq = maximumBy (comparing snd) $ zip seq (map hc' seq) powers = map (\n -> [2^n .. 2^(n + 1) - 1]) [0 .. 19] mallows = last.takeWhile ((< 0.55) . hc') $ [2^20, 2^20 - 1 .. 1]</lang>
Icon and Unicon
<lang icon>procedure main(args)
m := integer(!args) | 20
nextNum := create put(A := [], 1 | 1 | |A[A[*A]]+A[-A[*A]])[*A]
p2 := 2 ^ (p := 1)
maxv := 0
every n := 1 to (2^m) do {
if maxv <:= (x := @nextNum / real(n)) then maxm := n
if x >= 0.55 then mallows := n # Want *this* n, not next one!
if n = p2 then {
write("Max between 2^",p-1," and 2^",p," is ",maxv," at n = ",maxm)
p2 := 2 ^ (p +:= 1)
maxv := 0
}
}
write("Mallows's number is ",\mallows | "NOT found!")
end</lang>
Output:
->hc Max between 2^0 and 2^1 is 1.0 at n = 1 Max between 2^1 and 2^2 is 0.6666666667 at n = 3 Max between 2^2 and 2^3 is 0.6666666667 at n = 6 Max between 2^3 and 2^4 is 0.6363636364 at n = 11 Max between 2^4 and 2^5 is 0.6086956522 at n = 23 Max between 2^5 and 2^6 is 0.5909090909 at n = 44 Max between 2^6 and 2^7 is 0.5760869565 at n = 92 Max between 2^7 and 2^8 is 0.5674157303 at n = 178 Max between 2^8 and 2^9 is 0.5594594595 at n = 370 Max between 2^9 and 2^10 is 0.5549374131 at n = 719 Max between 2^10 and 2^11 is 0.5501008742 at n = 1487 Max between 2^11 and 2^12 is 0.5474628926 at n = 2897 Max between 2^12 and 2^13 is 0.5441447479 at n = 5969 Max between 2^13 and 2^14 is 0.5424427088 at n = 11651 Max between 2^14 and 2^15 is 0.5400710975 at n = 22223 Max between 2^15 and 2^16 is 0.5387840206 at n = 45083 Max between 2^16 and 2^17 is 0.537043657 at n = 89516 Max between 2^17 and 2^18 is 0.5360200678 at n = 181385 Max between 2^18 and 2^19 is 0.5346454311 at n = 353683 Max between 2^19 and 2^20 is 0.53377923 at n = 722589 Mallow's number is 1489 ->
J
Solution (tacit): <lang j> hc10k =: , ] +/@:{~ (,&<: -.)@{: NB. Actual sequence a(n)
AnN =: % 1+i.@:# NB. a(n)/n MxAnN =: >./;.1~ 2 (=<.)@:^. 1+i.@# NB. Maxima of a(n)/n between successive powers of 2</lang>
Alternative solution (exponential growth):
The first, naive, formulation of hc10k grows by a single term every iteration; in this one, the series grows exponentially in the iterations.
<lang j> hc10kE =: 1 1 , expand@tail
expand =: 2+I.@;
tail =: copies&.>^:(<@>:`(<@,@2:))
copies =: >: |.@(#!.1 |.)~ 1 j. #;.1 #^:_1 ::1:~ ]~:{.</lang>
Example: <lang j> ] A=:1 1 hc10k @]^:[~ 2^20x 1 1 2 2 3 4 4 4 5 6 7 7 8 8 8 8 9 ...
AnN A
1 0.5 0.666667 0.5 0.6 0.666667 ...
MxAnN@AnN A
1 0.666667 0.666667 0.636364 ...
MxAnN@AnN@hc10kE 20
1 0.666667 0.666667 0.636364 ...</lang>
Java
with corrections to 0 indexing.
<lang java>public class HofCon {
public static void main(final String[] args)
{
doSqnc(1<<20);
}
public static void doSqnc(int m)
{
int[] a_list = new int[m + 1];
int max_df = 0;
int p2_max = 2;
int k1 = 2;
int lg2 = 1;
double amax = 0;
a_list[0] = a_list[1] = 1;
int v = a_list[2];
for (int n = 2; n <= m; n++)
{
v = a_list[n] = a_list[v] + a_list[n - v];
if (amax < v * 1.0 / n)
amax = v * 1.0 / n;
if (0 == (k1 & n))
{
System.out.printf("Maximum between 2^%d and 2^%d was %f\n", lg2, lg2 + 1, amax);
amax = 0;
lg2++;
}
k1 = n;
}
}
}</lang> Output:
Maximum between 2^1 and 2^2 was 0.666667 Maximum between 2^2 and 2^3 was 0.666667 Maximum between 2^3 and 2^4 was 0.636364 Maximum between 2^4 and 2^5 was 0.608696 .... Maximum between 2^18 and 2^19 was 0.534645 Maximum between 2^19 and 2^20 was 0.533779
Oz
A direct implementation of the recursive definition with explicit memoization using a mutable map (dictionary): <lang oz>declare
local
Cache = {Dictionary.new}
Cache.1 := 1
Cache.2 := 1
in
fun {A N}
if {Not {Dictionary.member Cache N}} then
Cache.N := {A {A N-1}} + {A N-{A N-1}}
end
Cache.N
end
end
Float = Int.toFloat
for I in 0..19 do
Range = {List.number {Pow 2 I} {Pow 2 I+1} 1}
RelativeValues = {Map Range
fun {$ N}
{Float {A N}}
/ {Float N}
end}
Maximum = {FoldL RelativeValues Max 0.0}
in
{System.showInfo "Max. between 2^"#I#" and 2^"#I+1#": "#Maximum}
end</lang>
Output:
Max. between 2^0 and 2^1: 1.0 Max. between 2^1 and 2^2: 0.66667 Max. between 2^2 and 2^3: 0.66667 Max. between 2^3 and 2^4: 0.63636 Max. between 2^4 and 2^5: 0.6087 Max. between 2^5 and 2^6: 0.59091 Max. between 2^6 and 2^7: 0.57609 Max. between 2^7 and 2^8: 0.56742 Max. between 2^8 and 2^9: 0.55946 Max. between 2^9 and 2^10: 0.55494 Max. between 2^10 and 2^11: 0.5501 Max. between 2^11 and 2^12: 0.54746 Max. between 2^12 and 2^13: 0.54414 Max. between 2^13 and 2^14: 0.54244 Max. between 2^14 and 2^15: 0.54007 Max. between 2^15 and 2^16: 0.53878 Max. between 2^16 and 2^17: 0.53704 Max. between 2^17 and 2^18: 0.53602 Max. between 2^18 and 2^19: 0.53465 Max. between 2^19 and 2^20: 0.53378
Perl 6
This is written in relatively low-level primitives (for Perl 6) because rakudo still has some performance issues with long lists. One interesting feature, however, is the use of the max infix operator as an assignment operator, and using it on a Pair so that the maximum is calculated on the key of the Pair, while the value of the Pair carries the index of the maximum through to the end without having to keep a second variable. <lang perl6>my $POW = 20; my $top = 2**$POW;
my @a = (0,1,1); @a[$top] = 0; # pre-extend array
my $n = 3; my $p = 1;
loop ($n = 3; $n <= $top; $n++) {
@a[$n] = $p = @a[$p] + @a[$n - $p];
}
my $last55; for 1 ..^ $POW -> $power {
my $beg = 2 ** $power; my $end = $beg * 2 - 1; my $max; my $ratio;
loop (my $n = $beg; $n <= $end; $n++) {
my $ratio = @a[$n] / $n;
$last55 = $n if $ratio * 100 >= 55;
$max max= $ratio => $n;
}
say $power.fmt('%2d'), $beg.fmt("%10d"), '..', $end.fmt("%-10d"), $max.key, " at ", $max.value;
} say "Mallows' number would appear to be ", $last55;</lang>
Output:
<lang> 1 2..3 0.666666666666667 at 3
2 4..7 0.666666666666667 at 6 3 8..15 0.636363636363636 at 11 4 16..31 0.608695652173913 at 23 5 32..63 0.590909090909091 at 44 6 64..127 0.576086956521739 at 92 7 128..255 0.567415730337079 at 178 8 256..511 0.559459459459459 at 370 9 512..1023 0.554937413073713 at 719
10 1024..2047 0.550100874243443 at 1487 11 2048..4095 0.547462892647566 at 2897 12 4096..8191 0.544144747863964 at 5969 13 8192..16383 0.542442708780362 at 11651 14 16384..32767 0.540071097511587 at 22223 15 32768..65535 0.538784020584256 at 45083 16 65536..131071 0.537043656999866 at 89516 17 131072..262143 0.536020067811561 at 181385 18 262144..524287 0.534645431078112 at 353683 19 524288..1048575 0.533779229963368 at 722589 Mallows' number would appear to be 1489</lang> Incidentally, this takes three hours or so currently. We expect this to get much faster.
Here is a more list-oriented version. Note that @a is a lazy array, and the Z variants are "zipwith" operators. <lang perl6>my $n = 3; my @a := (0,1,1, -> $p { @a[$p] + @a[$n++ - $p] } ... *);
my $last55; for 1..19 -> $power {
my @range := 2**$power .. 2**($power+1)-1;
my @ratios = (@a[@range] Z/ @range) Z=> @range;
my $max = [max] @ratios;
($last55 = .value if .key >= .55 for @ratios) if $max.key >= .55;
say $power.fmt('%2d'), @range.min.fmt("%10d"), '..', @range.max.fmt("%-10d"),
$max.key, ' at ', $max.value;
} say "Mallows' number would appear to be ", $last55;</lang>
PicoLisp
<lang PicoLisp>(de hofcon (N)
(cache '*HofCon (format (seed N))
(if (>= 2 N)
1
(+
(hofcon (hofcon (dec N)))
(hofcon (- N (hofcon (dec N)))) ) ) ) )
(scl 20)
(de sequence (M)
(let (Lim 4 Max 0 4k$ 0)
(for (N 3 (>= M N) (inc N))
(let V (*/ (hofcon N) 1.0 N)
(setq Max (max Max V))
(when (>= V 0.55)
(setq 4k$ N) )
(when (= N Lim)
(prinl
"Maximum between " (/ Lim 2)
" and " Lim
" was " (format Max `*Scl) )
(inc 'Lim Lim)
(zero Max) ) ) )
(prinl
"Win with " 4k$
" (the task requests 'n > p' now)" ) ) )
(sequence (** 2 20))</lang> Output:
Maximum between 2 and 4 was 0.66666666666666666667 Maximum between 4 and 8 was 0.66666666666666666667 Maximum between 8 and 16 was 0.63636363636363636364 Maximum between 16 and 32 was 0.60869565217391304348 Maximum between 32 and 64 was 0.59090909090909090909 Maximum between 64 and 128 was 0.57608695652173913043 Maximum between 128 and 256 was 0.56741573033707865169 Maximum between 256 and 512 was 0.55945945945945945946 Maximum between 512 and 1024 was 0.55493741307371349096 Maximum between 1024 and 2048 was 0.55010087424344317418 Maximum between 2048 and 4096 was 0.54746289264756644805 Maximum between 4096 and 8192 was 0.54414474786396381303 Maximum between 8192 and 16384 was 0.54244270878036220067 Maximum between 16384 and 32768 was 0.54007109751158709445 Maximum between 32768 and 65536 was 0.53878402058425570614 Maximum between 65536 and 131072 was 0.53704365699986594575 Maximum between 131072 and 262144 was 0.53602006781156104419 Maximum between 262144 and 524288 was 0.53464543107811232092 Maximum between 524288 and 1048576 was 0.53377922996336783427 Win with 1489 (the task requests 'n > p' now)
PL/I
<lang PL/I> /* First part: */
declare L (10000) fixed static initial ((1000) 0); L(1), L(2) = 1; do i = 3 to 10000;
k = L(i); L(i) = L(i-k) + L(1+k);
end; </lang>
PureBasic
<lang PureBasic>If OpenConsole()
Define.i upperlim, i=1, k1=2, n=3, v=1
Define.d Maximum
Print("Enter limit (ENTER gives 2^20=1048576): "): upperlim=Val(Input())
If upperlim<=0: upperlim=1048576: EndIf
Dim tal(upperlim)
If ArraySize(tal())=-1
PrintN("Could not allocate needed memory!"): Input(): End
EndIf
tal(1)=1: tal(2)=1
While n<=upperlim
v=tal(v)+tal(n-v)
tal(n)=v
If Maximum<(v/n): Maximum=v/n: EndIf
If Not n&k1
PrintN("Maximum between 2^"+Str(i)+" and 2^"+Str(i+1)+" was "+StrD(Maximum,6))
Maximum=0.0
i+1
EndIf
k1=n
n+1
Wend
Print(#CRLF$+"Press ENTER to exit."): Input()
CloseConsole()
EndIf</lang>
Python
<lang python>from __future__ import division
def maxandmallows(nmaxpower2=20):
# Note: The first hc number is returned in hc[1];
# hc[0] is not part of the series.
nmax = 2**nmaxpower2
hc, mallows, mx, mxpow2 = [None, 1, 1], None, (0.5, 2), []
for n in range(2, nmax + 1):
ratio = hc[n] / n
if ratio > mx[0]: mx = (ratio, n)
if ratio >= 0.55: mallows = n
if ratio == 0.5:
print("In the region %7i < n <= %7i: max a(n)/n = %s" %
((n)//2, n, "%6.4f at n = %i" % mx))
mxpow2.append(mx[0])
mx = (ratio, n)
hc.append(hc[hc[n]] + hc[-hc[n]])
return hc, mallows if mxpow2 and mxpow2[-1] < 0.55 and n > 4 else None
if __name__ == '__main__':
hc, mallows = maxandmallows(20)
if mallows:
print("\nYou too might have won $1000 with the mallows number of %i" % mallows)
</lang>
Sample output
In the region 1 < n <= 2: max a(n)/n = 0.5000 at n = 2 In the region 2 < n <= 4: max a(n)/n = 0.6667 at n = 3 In the region 4 < n <= 8: max a(n)/n = 0.6667 at n = 6 In the region 8 < n <= 16: max a(n)/n = 0.6364 at n = 11 In the region 16 < n <= 32: max a(n)/n = 0.6087 at n = 23 In the region 32 < n <= 64: max a(n)/n = 0.5909 at n = 44 In the region 64 < n <= 128: max a(n)/n = 0.5761 at n = 92 In the region 128 < n <= 256: max a(n)/n = 0.5674 at n = 178 In the region 256 < n <= 512: max a(n)/n = 0.5595 at n = 370 In the region 512 < n <= 1024: max a(n)/n = 0.5549 at n = 719 In the region 1024 < n <= 2048: max a(n)/n = 0.5501 at n = 1487 In the region 2048 < n <= 4096: max a(n)/n = 0.5475 at n = 2897 In the region 4096 < n <= 8192: max a(n)/n = 0.5441 at n = 5969 In the region 8192 < n <= 16384: max a(n)/n = 0.5424 at n = 11651 In the region 16384 < n <= 32768: max a(n)/n = 0.5401 at n = 22223 In the region 32768 < n <= 65536: max a(n)/n = 0.5388 at n = 45083 In the region 65536 < n <= 131072: max a(n)/n = 0.5370 at n = 89516 In the region 131072 < n <= 262144: max a(n)/n = 0.5360 at n = 181385 In the region 262144 < n <= 524288: max a(n)/n = 0.5346 at n = 353683 In the region 524288 < n <= 1048576: max a(n)/n = 0.5338 at n = 722589 You too might have won $1000 with the mallows number of 1489
If you don't create enough terms in the sequence, no mallows number is returned.
Ruby
<lang ruby>class HofstadterConway10000
def initialize @sequence = [nil, 1, 1] end attr_reader :sequence
def [](n)
raise ArgumentError, "n must be >= 1" if n < 1
a = @sequence
a.length.upto(n) {|i| a[i] = a[a[i-1]] + a[i-a[i-1]] }
a[n]
end
end
hc = HofstadterConway10000.new</lang>
I run the sequence backwards to make it easier to find the Mallows number.
The first call to the hc.[] method will pass the value 2^20, pre-populating the entire sequence.
<lang ruby>p = -1 20.downto(1).each_cons(2) do |j, i|
max_n, max_v = -1, -1
(2**j).downto(2**i).each do |n|
v = hc[n].to_f / n
max_n, max_v = n, v if v > max_v
# Mallows number
p = n if p == -1 and v >= 0.55
end
puts "maximum between 2^#{i} and 2^#{j} occurs at #{max_n}: #{max_v}"
end
puts "the mallows number is #{p}"</lang>
output
maximum between 2^19 and 2^20 occurs at 722589: 0.533779229963368 maximum between 2^18 and 2^19 occurs at 353683: 0.534645431078112 maximum between 2^17 and 2^18 occurs at 181385: 0.536020067811561 maximum between 2^16 and 2^17 occurs at 89516: 0.537043656999866 maximum between 2^15 and 2^16 occurs at 45083: 0.538784020584256 maximum between 2^14 and 2^15 occurs at 22223: 0.540071097511587 maximum between 2^13 and 2^14 occurs at 11651: 0.542442708780362 maximum between 2^12 and 2^13 occurs at 5969: 0.544144747863964 maximum between 2^11 and 2^12 occurs at 2897: 0.547462892647566 maximum between 2^10 and 2^11 occurs at 1487: 0.550100874243443 maximum between 2^9 and 2^10 occurs at 719: 0.554937413073713 maximum between 2^8 and 2^9 occurs at 370: 0.559459459459459 maximum between 2^7 and 2^8 occurs at 178: 0.567415730337079 maximum between 2^6 and 2^7 occurs at 92: 0.576086956521739 maximum between 2^5 and 2^6 occurs at 44: 0.590909090909091 maximum between 2^4 and 2^5 occurs at 23: 0.608695652173913 maximum between 2^3 and 2^4 occurs at 11: 0.636363636363636 maximum between 2^2 and 2^3 occurs at 6: 0.666666666666667 maximum between 2^1 and 2^2 occurs at 3: 0.666666666666667 the mallows number is 1489
Tcl
The routine to return the nth member of the sequence. <lang tcl>package require Tcl 8.5
set hofcon10k {1 1} proc hofcon10k n {
global hofcon10k
if {$n < 1} {error "n must be at least 1"}
if {$n <= [llength $hofcon10k]} {
return [lindex $hofcon10k [expr {$n-1}]]
}
while {$n > [llength $hofcon10k]} {
set i [lindex $hofcon10k end] set a [lindex $hofcon10k [expr {$i-1}]] # Don't use end-based indexing here; faster to compute manually set b [lindex $hofcon10k [expr {[llength $hofcon10k]-$i}]] lappend hofcon10k [set c [expr {$a + $b}]]
} return $c
}</lang> The code to explore the sequence, looking for maxima in the ratio. <lang tcl>for {set p 1} {$p<20} {incr p} {
set end [expr {2**($p+1)}]
set maxI 0; set maxV 0
for {set i [expr {2**$p}]} {$i<=$end} {incr i} {
set v [expr {[hofcon10k $i] / double($i)}] if {$v > $maxV} {set maxV $v; set maxI $i}
}
puts "max in 2**$p..2**[expr {$p+1}] at $maxI : $maxV"
}</lang> Output:
max in 2**1..2**2 at 3 : 0.6666666666666666 max in 2**2..2**3 at 6 : 0.6666666666666666 max in 2**3..2**4 at 11 : 0.6363636363636364 max in 2**4..2**5 at 23 : 0.6086956521739131 max in 2**5..2**6 at 44 : 0.5909090909090909 max in 2**6..2**7 at 92 : 0.5760869565217391 max in 2**7..2**8 at 178 : 0.5674157303370787 max in 2**8..2**9 at 370 : 0.5594594594594594 max in 2**9..2**10 at 719 : 0.5549374130737135 max in 2**10..2**11 at 1487 : 0.5501008742434432 max in 2**11..2**12 at 2897 : 0.5474628926475664 max in 2**12..2**13 at 5969 : 0.5441447478639638 max in 2**13..2**14 at 11651 : 0.5424427087803622 max in 2**14..2**15 at 22223 : 0.5400710975115871 max in 2**15..2**16 at 45083 : 0.5387840205842557 max in 2**16..2**17 at 89516 : 0.5370436569998659 max in 2**17..2**18 at 181385 : 0.5360200678115611 max in 2**18..2**19 at 353683 : 0.5346454310781124 max in 2**19..2**20 at 722589 : 0.5337792299633678