I agree about the lack of commerce in the beginning, but as you said, we needed production in Aachen.
I really do not see how we could have limited our expansion. Almost any city was settled to grab a resource and they're all good cities. Probably an earlier GLH+Colossus could have helped, but who even thinks to GLH+Colossus in a rainforest map?
Of course every city had its uses, no doubt about that. What I'm saying is that after a certain number of cities the extra maintenance from the new city slows the whole empire. Eg. we get the next economic tech 2 turns later. We should have calculated/tested the maintenance cost hits (by maintaining the test save for a little longer) and should have stayed more conservative.
Next game we should try setting PD-style research objectives and key starting points. In this game we had Oracle, CS, Engineering, Edu, Steel. Running some tests that get these key techs 1-2 turns sooner gets us the same number of turns in the end (or more due to snowball effects).
IMO, is the research we chosen the main mistake. We weren't able to use Mansa to our advantage.
And about the galley/WB you're right. I never opposed it, but never showed much love for it.
Paras are not the reasl problem. The real problem was that Sury and Qin has a tech lead and we were close to strike. I always recommended to play as if we need paras, long before we actually know it.
Playing like we will need paras is not the best course of action. Why would we eg. do a double bulb into Edu and invest hundreds (thousands?) of hammers into universities? Especially if we estimate that there is <50% chance that we need to tech that deep? If we knew these in advance, we could have planned some bulbs to have a decent Edu date and better late game research. It's time to balme the mapmaker I guess. Having the icy quarter as neighbiour gives us a chance to have a long term plan and execute it.
On the other hand, it really gave a unique feeling to this game, and was educational still