You were proposing a ring final on a 20A device with cable rated at less than 20A.
Was I - where have I said that then. I could make such a proposal, but only for money

.
Hi guys, could someone gave me, or direct me to, a general overview of how a house is wired. Am I right in thinking that wire/cable goes from the mains all around the house, then back to the mains?
You are describing a ring circuit. If you break the circuit, then electricity can still reach all the other sockets either side of it as both sides go back to the consumer unit.
.
.
It is becoming increasingly popular to run socket circuits as radials using a thicker cable, for safety reasons.
it is also becoming harder to install ring mains
Yes, because the increasing amount of thermal insulation in houses is making it harder to achieve an Iz of 20A with 2.5mm².
The same would be true of any circuit, surely.
Yes, but with radials you have the added flexibility of changing the OPD rating as well as changing the cable size - with ring finals you're stuck with 32A.
no you're not, there's nothing to say you can't install a 6mm ring or a 20A ring.....
I don't think a 20A RFC is recognised by the regs?
Yes it is, as is any other properly designed circuit.
Would it have to if I choose to define it as a ring final circuit.
It would then have to comply with 433.1.5.
What do you do about 523.8 and 433.4.2?
Ignore them.
Do you do that with any regulation that inconveniently makes your designs non-compliant?
433.4.2, seems to be suggesting a way of rectifying a badly designed/installed parallel circuit.
If parallel cables are not carrying substantially the same load its because there is something wrong. And no gibberish regulation will rectify the problem.
Ring final circuits serving socket outlets do/can/frequently have parallel cables which are not carrying substantially the same load.
Just because you don't understand it doesn't make it gibberish.
What is a "non-twisted single core cable"? I'm pretty sure there is no such thing, never has been and never will be.
So out of all the cables in all the world, which could be used in situations within the scope of BS 7671, there are no non-twisted single-core ones?
You being pretty sure doesn't mean that there are not.
And if you try to install "non sheathed cables in trefoil" your circuit will be inadequately mechanically protected and will not comply with the regs.
Will you please take a look at numbers 6/7, 10, 43, 45 & 50 in Table A2, and then explain the following:
1) Why those cables are inadequately mechanically protected.
2) In general terms why the arrangement of non-sheathed cables has any bearing on whether or not mechanical protection is required, and why any mechanical protection provided is less effective if the cables are in a trefoil arrangement.
The only necessary rules for parallel cables are same type, length and csa, run from point A to B with no other connection points or take offs, and connected between one protective device and one isolator - end of.
So your proposed 20A ring final circuit supplying socket outlets would not be a compliant circuit with parallel conductors.
I have no idea why ring final circuits have become associated with parallel circuits, they are chalk and cheese. And I doubt if anyone could give a sensible explanation.
Each point of use in a ring final circuit is served by two parallel conductors, but such circuits do not comply with any of the regulations for parallel conductors, hence the special exemption of 433.1.5 and the fact that in at least one parallel conductor regulation it is explicitly stated that the regulation does not preclude ring finals.