Well one explanation given by someone in this forum is that the plan was you'd run the ring round the core of the house, and any sockets on the outside would be spurs. That way, the ring itself is relatively short for the area covered, and the spurs are no longer than branches off a radial would be.
Given this arrangement, I could well imagine the extra length of smaller cable needed to bring the other end of the ring back to the fuse would be more than offset by reducing the CSA of the "trunk" of the radial. There's also the factor, especially given some of the discussions regarding competence of some people, that the whole circuit can (for most installations) be done with one size of cable without any thought or maths needed on the part of the installer (who could just do the whole lot in 7/029 as long as he didn't exceed the guidelines for area covered).
I feel I should point out, that the early ring installs offered in most cases up to a medium sized home, just one single ring for the whole home. The ring run under the first floor, going up and down to sockets. It progressed from that to an upstairs and downstairs ring, then as we have now - a kitchen + ground + first floor ring. As my home is built on a concrete slab, all wiring is still in the ground floor ceiling.
My home progressed with rewires, from single ring with very few sockets, to a more extensive single ring, to what I have now - an upstairs ring, downstairs ring, loft/workshop radial and a garage radial. To the basic scheme I have added numerous outlets, as they have been needed and opportunity presented itself. I have never counted the number of sockets, but I would guess at something in excess of 50. I have no substantial loads at all, biggest are washer and dishwasher, but I do have lots of tiny loads of a few watts.
I don't have a separate ring for the kitchen, with no plans to install one - loading is so light on my present ring.