I agree with every word of your post, mrparkersdogbite - including, particularly, what a brilliant job Neilando has done - especially considering the limited information at his disposal. I was going to make exactly the same points myself about the buttresses alternating between different thicknesses, with the thinner ones seeming to extend outwards slightly less than the thicker ones.
Neilando, thanks again for your brilliant work. The only additional point that I would make is that if you wish to refine your model even further, as well as varying the thicknesses of the columns you could use the white lines in the first image below, as well as the ground plan shape in the second image, to try and guess the exterior shape of the stands. Obviously, there would have to be some guesswork involved in tallying everything together, as you would be attempting to interpret a high-elevation 2-D perspective image in 3-D.
Some of the following may be redundant because I wrote this post earlier but couldn't send it because SSC was playing up on me: if your model suggests that it might be possible that the roof "ribs" do indeed converge radially on the centre of the pitch - in the way that I described in my last post - would it be convenient to model such a roof (including with the "ribs" giving way to radial cable "spokes" that extend a short distance to a circular ring of cable)? I completely understand if this would be too onerous a task (I'm afraid that I'm ignorant of the amount of work involved in creating such a 3-D model), particularly since I think that the oddly-faceted planes of the roof would make modelling convergent features extremely difficult. (I imagine that the only perfect way to do it would be to model vertical 2-D plane surfaces centred on the centre of the pitch which then fan out radially and bisect the contoured roof. Obviously, I wouldn't ask you to consider constructing something that complex, but I imagine that it might be part of how HdM's structural engineers modelled the roof shape and dynamics.)
So, to be clear (because much of this post was written before your latest post) I understand what you are saying about the length of the brick-clad roof "ribs" if they were extended over the pitch - and how much light they'd block, but have you considered only extending the brick ribs as far as the stands and then have them replaced by very thin lines representing cables? Do you think that the brick "ribs" would still be unfeasibly long? I envisage that these tensioned cables emerging from the ends of the brick ribs might be as short a length as possible before connecting to a circular cable tension ring that would be as wide a diameter as possible, i.e. almost as wide as the pitch. Sorry if none of this makes any sense because either it doesn't seem feasible or I haven't been very clear in what I'm saying (or both

). I know that this post is a bit of a mess.
the white lines in this drawing seem to illustrate the shape of the stadium exterior
this ground plan diagram gives the overall shape of the stadium's footprint from above
Love Neilando's renders of the shell of the stadium though obviously we all understand that there is some guesswork involved. Thanks to him for his efforts anyway. Just two observations that I'd add - in the sketches, the buttresses clearly seem to alternate between being thicker then thinner (you can see this from the sketch of the West Stand from the main gate). You can also see this in the sketch of the new Stamford Bridge walkway but in this picture the thinner buttresses actually appear to be set in closer to the stadium structure than their thicker equivalents. I've no idea how tricky it would be to incorporate these features into an updated 3D model but if you could then it'd give us an even closer approximation of the current design.