Nares: you're wrong in suggesting that I had an initial bias. I genuinely wanted to find out which approach was better - and I still do. I tried to make my testing as unbiased as possible. The answer about NC first or later may well prove to be situational, but it would be really interesting to establish the circumstances that favour NC first or NC later (and how much later). But I want that answer to be established by evidence rather than opinion - that was kinda the point of my original post! I'd really love to see someone (preferably with more experience than me) make a similar test from a different start point even if they come up with a different conclusion.
Monthar, that's great - but for a valid test you need to run two scenarios to an equivalent position. For example NC + 3 satellite cities as in my original post. One can't tell very much from looking at an optimised NC build by itself.At turn 50 I completed the NC with an NC before expansion strategy and I'm at 21 BPT.