This is working as intended, to avoid confusion if a subnet extends to the next higher octet we rotate the bits so all subnets with 0's in the shared octet are allocated first. If you had more nodes the system would fill in the gaps.
PR link https://github.com/openshift/openshift-docs/pull/8523
Doc looks good, bug verified
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2018:1816