Yes, half way round the ring is the worst case - and that's what matters for fault protection etc. Anywhere else on the ring will have a lower value - as demonstrated by the graph posted (whichever way up you look at it)
The different lines on the graph are irrelevant. The deviations from the simplistic formula only occur away from the centre point of the ring - where the resistance is lower than the worst case anyway.
But for a simple explanation ...
Consider a fault at the midpoint of the ring. From the MCB in the CU to the fault, you have two equal conductors in parallel - each of which is half the length of the ring and hence has half the resitance of the whole ring. So you have two resistors of value r1/2 in parallel, and two equal resistors in parallel give you half the value of each of them. So you have (r1 / 2) / 2, or r1 / 4.
The CPC has the same calculation, so from MET to fault you have r2 / 4.
So total is (r1 / 4) + (r2 / 4), which is better written as (r1 + r2) / 4
Again, this is the worst case, at any other pointbround thecring, the total resistance to the fault will be lower. So you don't need to considervthe effects of different line & CPC sizes etc. This even hold true if the cable CSA changes round the ring IF you consider "middle of the ring" to be in terms of cable resistance rather than distance.
It would be a diffetent case if the CSA variations were "odd" - eg if in one leg you had a larger line, and in the other leg you had a larger CPC. But that would be unusual !
