I guess our judgement differs on this one. I'll try and write up a very detailed first spoiler so we can compare results.
Your calculations show me that ancient beakers are dearer than I assumed - getting some seafood online early (with Fishing first) is needed to kickstart research as well as growth/production.
However, I almost certainly want to REX before getting a third seafood improved (which takes many turns to pay off compared with a mine) since the map is crowded, and starting on a plains hill makes it so much easier to get the WBs and fogbusters built, not to mention the extra hammer while building the second and third settlers.
Great, I'll be careful with my first spoiler too.
I don't think we really disagree, but we have different goals.
We both think that the early game is crucial.
We both think that 2500BC-1000BC development will be faster if the visible resources are shared between 2 cities.
Now, I am playing for Culture. If I share resources in a non-undoable way (like moving and losing a fish for the capital forever), and if then I get a perfect development of my civ (9 cities, all of them good, peace, etc, etc), then I will have a good capital going for Legendary instead of having a superb capital going for Legendary. This will have a big impact on my finish date.
On the other hand, if I leave all these resources for the capital and then I am unable to get my civ correctly developped (stucked at 2 cities, etc...), then I will be in a hard position, much worse than if I had shared resources.
I am confident in my strengths and I am aiming for an Award, so I will take the risk.
I think that people aiming for fastest spaceship should do the same, as a superb capital is crucial for this victory condition too.
I consider that people going for fastest conquest should share resources climbing to a PH instead, as a superb capital won't even have the time to get completely developped.
Does this all make sense?