Theory is we need to keep within the limits to ensure the automatic disconnection of the supply and volt drop are within the limits published in BS 7671. In the days of the fuse, exceeding the loop impedance would mean it may take a little longer to disconnect, and a jump from 5 seconds to 6 seconds was not really too much of a problem.
However as we moved to the MCB this changed, the MCB is two devices in one, a thermal and a magnetic trip, the latter is basic to protect against the short circuit, and with a B rated MCB works at between 3 and 5 times the thermal device setting, and works very fast, around 0.01 of a second, where the thermal part can take many minuets to trip, so now going over the allowed loop impedance can mean a huge change in the disconnection time, so it is very important to keep within the limits.
It is allowed to use inquiry to work out the loop impedance, so in theory you can get out the original installation certificate and look at what the loop impedance was, or the R1+R2 and calculate how much cable you can add. And it does not need to be measured with a loop impedance meter, one can use a low ohm meter using at least 200 mA to measure the resistance with.
However the point is you can't simply add more cable without working out what you already have and how close to the limits you are. At least when you have some one leaning over your shoulder (LABC inspector) to see your doing it right.
The IET
publish model forms step one is to down load the forms and see if you think you can fill them in, if you can then maybe the LABC inspector will accept your signature. For me it is relativity easy, I have the test gear, or can borrow it, some shared with my son, so I can complete the forms, and even if I didn't have the test gear, I have filled in enough to like with the Emma Shaw case make a good guess at what the results should be, I would not guess, as don't want there to be a court case for me.
I am trying to make this a real world answer, and in the real world the last thing we want to hear is "Do it again and do it properly" if LABC was not involved I may have said use near enough engineering and just use a socket tester with loop, or get an EICR done after the work is completed.
But we have seen the problems on other threads getting the LABC to accept work done.