This observation is clearer when treating the 64-bit chunks simply as double-precision IEEE754 floating points. Then the set of pairs (i,j) for which xi/xj is ±2n for some n clearly draws lines with slopes close to powers of 2/e. But they don’t seem quite straight, so the slope is not so clear. In any case there is some pretty big long-distance correlation between xi and xj with rather different indices. (Note that if we explain the first line j≃(2/e)i then the other powers are clearly consequences.)
This observation is clearer when treating the 64-bit chunks simply as double-precision IEEE754 floating points. Then the set of pairs (i,j) for which xi/xj is ±2n for some n clearly draws lines with slopes close to powers of 2/e. But they don’t seem quite straight, so the slope is not so clear. In any case there is some pretty big long-distance correlation between xi and xj with rather different indices. (Note that if we explain the first line j≃(2/e)i then the other powers are clearly consequences.)