Easier to test two radials than a ring.
Also poorer fault coverage?
For example, many only do the R1+R2 test on a radial to verify (dead) that Zs will be met, and they should also check polarity at each socket as well. But you only know your worst-case R1+R2 if you also know the end socket, which for an EICR is unlikely. Though if you have just installed it then you would know.
If the radial is "pure" (i.e. only ever two cables at each accessory so no branching tree-like) then an end socket test gives you good coverage of the whole circuit, but in practice you might have a tree-like arrangement and so you have to test the "branches" to know that all are good.
Do you test RN+R2 as well? Probably not.
Where as for the RFC you are required to test all three: r1, r2, and rN and that gives you good knowledge of all circuit's connections are tolerably good from r1 and rN being practically identical, and r1/r2 ratio matching the CPC/L conductor size ratio (usually, assuming no metal bonding from more than one socket).
True, if there are spurs you do not have coverage of them in the end-end test and also for EICR probably no knowledge of where they might be, but when you do the socket polarity test using the figure-of-eight approach it gives you a
very good indication not only of any possibly bad switch/contact in a socket, but also the likely existence of a spur from measured resistance not being
identical to all other sockets.
Personally I don't like spurs for exactly that reason: they breaks the excellent fault coverage of the end-end test which is convenient to do at the CU without knowledge of the physical route and order of accessories.