My tests showed me what I always new, there's no free lunch! Everything has a trade off.
If we want early teching to BW we improve the sheep first. The sheep adds 2
which is significant early.
I believe the improved sheep will be 4/1/2, compared to the improved cows 4/2/0. So we can get some more commerce (which is admittedly about a 20% increase in our current total commerce) but will give up a hammer/turn. If we are building a settler at size 1, that is roughly a 15% difference (6 food-hammers/turn for the sheep compared to 7 for the cow). This should cause a 1 turn delay in building the settler at size 1. The sheep is also obviously not in the direction of the marble, as far as positioning the worker for road building or moving to the new city.
If we grow to size 2 before building the settler, our food-hammer base increases and the 1 hammer/turn difference becomes less significant (~12%).
If we want to settle city 2 the fastest, settler now, at size 1 will settle 4 turns faster than the next best option I believe. The only way this works is to have both the scout and the warrior fogbusting the path and the worker roading to the site immediately upon improving the cow.
As I noted previously, I am not sure that "city #2 the fastest" is the metric we should be optimizing towards. "City #3 the fastest" might be better, or even city #4. I generally dislike building a settler at size 1, and sending our worker off to road rather than improving a second strong tile at the capital just does not seem good to me. Especially with barb animals roaming around, which means the worker must be covered. And with bears out and about, nothing we have can effectively cover the worker when it is roading on flat ground.
The real question we need to ask is what will Peter build 1st and 2nd? If he builds settler first or second, which AI do at this level, and he wants the site, we can't be him anyway.
We know from demographics during my turnset that all of the AIs have grown to at least size 2 (some to size 3). So no AI has gone settler first. Some could have started a settler upon reaching size 2 -- no way to know for certain. With their starting workers improving tiles (also visibly from demographics) I doubt we can beat a size 2 AI settler.
I think I like Settler now, but we need to get the warrior and the scout into the area and hope to eliminate the bear somehow. If we are not able to do that settling T31 -vs- T35 won't be possible anyhow.
I do not really see how we can eliminate a bear right now. Best option would be the warrior on a forested hill, and even that will not give great odds. Warrior plus scout together could probably be pretty certain of doing so, but on flat ground even together they can't deal with a bear. Better to plan to use the settler's speed to outrun any bear problems, with warrior and scout providing info/visibility. Or build another warrior or two while growing and use them to escort.
I think growing to size 2 or even size 3 before starting the settler will "only" cost us a few turns, while also allowing time to build an extra warrior for escort duties. The larger capital will also allow for a more rapid third city, which I think is an important consideration.
What if horses lie in a different direction?
On the Astro question, we can afford to wait a bit before we decide. We know we want mining > BW next. That is at least T35 and probably later.
We will know about horses in 3 more turns. If they are available somewhere other than the marble site, what do we do? Is stronger military that important at this stage? How likely are we to have hammers available to build even one chariot before getting a second worker and another settler out? I think we could probably leave any horses for our third city, unless the resource is positioned to be very vulnerable to settlement by Peter or Suleiman.
I would like to also ask about what our new city will do once we have it? Without any way to pop borders (finish AH -> Mining -> BW seems set, so it is going to take a while before we could even think of teching Mysticism for monuments or Writing for libraries), the marble site is going to be one plains cow for quite a while. And that will be the far plains cow, meaning we have to send the worker along to improve it. Or I suppose we could keep the worker at the capital and have the marble city build another worker. Slow, but without any other tiles to improve for a while it might be an option?
The burned city site might actually be stronger, since it has wheat that can be improved and the grass hill to be mined without needing a border pop. But with a greater threat of losing the site to an AI, I think we have to go for the marble if we want to have it for the Oracle.
How do we intend to pop borders? We are going to need to do so for almost any site we choose early -- Pollux has set us up with some tough choices.