I have been doing some experiments with marble starts with Elizabeth lately.
1.- Parthenon.
Getting the Parthenon is quite easy. It is very expensive, so you have to connect marble. The benefit of the Parthenon is twofold: first you get additional GPP. My estimate is that, in a typical 15GP game, with 3 cities contributing to the pool, 80% of the last GA is free thanks to the Parthenon; second, the 10cpt itself, equivalent to 1 bombed GA if the Hermitage and the Parthenon are in different cities, 1.5 bombed GA if they are both in the same city. The disadvantages of Parthenon are two: it determines your early research and the hammers you put on it could have had other uses.
2.- CS slingshot.
Researching up to CoL before any AI builds the Oracle is very difficult. I have had 0% success out of 10 games, losing by 8-10 turns.
Marble is optional, Oracle is so cheap and CoL is so expensive that you can build it without a discount. The best way to have a high enough research capability is to beeline to Pottery and then fill your FPs with cottages while beelining to PH. But the turns lost on the wheel and Pottery are simply too many. The other way is to have gems or gold in the capital. One of them is not enough. Even beelining to PH-CoL from the very beginning is not enough.
Maybe Elizabeth is just not suited to this strategy.
3.- Oracle for CoL.
Simply getting the Oracle is easy enough. Marble is optional. The benefit is a free expensive tech and a religion. The disadvantage is it determines your early research. At least you don’t need to skip agriculture to have a chance.
I don’t know if it is worth it. Is it better than beelining Pottery and then Alphabet? Is it better than getting Alphabet form the Oracle?
Incidentally, if you end up getting a GProphet, the best use for it could be bulbing Theo, getting a religion and then Sistine Chapel with marble. I still prefer a GA, though.
4.- No cottage city.
I had a food rich map that was asking to be played with cottages in the capital and GPFarms in the rest of the cities. The result was disappointing (14xxAD). The problem was I was doing 100bpt by 1AD, instead of the usual 170/200bpt that are needed to win the liberalism race with the help of 1/0 GS. I think a cottage city different from the capital is needed.
I had an idea I am reluctant to try, because of my obsession with efficiency. Here it is: your second city is a 1 or 2 food resources with 4-5 cottages. Instead of farming it all, which doesn’t help a lot in the beginning because of the happiness cap, cottage all the FPs and work them. That will help with the early research output. Then, when the happiness limit disappears, farm over the cottages (oh, this sounds like a sacrilege to me!) to max the GPP generating capability.
5.- The theoretical limit.
In one of the games I took CoL from Oracle in the capital, then Parthenon in the capital, then used the GP for Theo and build the Sistine Chapel in the capital, where I put the Hermitage later on. The 3 WW were done in the BC. The capital had 10 cottages, 6 of them early ones. Only 2 cathedrals were built in the capital, since I got no more religions. With this setup, the capital went legendary without the help of any GA in less than 35 turns since I raised the culture bar.
Ending the research phase around 200AD and then 30-35 turns takes us to 1100AD.
Building only one of the three WW but building a third cathedral would add 4 turns to the final date. So with a faster research rate, with 4 religions and 1WW, the capital could reach legendary in 1000AD without the need to bomb a GA.
Now, I only need to do all that in the capital and at the same time get 13GA by 1000AD. The GPFarm will get 25000c easily if 8 GA are bombed there. The other city will have to get at least 3 cathedrals and at least 100 commerce+rawculture to be able to do it with the help of 5 GA. Am I dreaming?
EDIT: Is it possible to get 13GA by 1000AD?
Assuming NE is built in 195BC and no GPP is generated before that, there are 40 turns to amass 6499GPP. With Parthenon, 6 artists in the GPFarm and 4 artist in 2 other cities it can barely be done. Now, at 195 BC it will be difficult to have 6 artists instantly, and in addition there is always unused GPP in the cities at the end of the game, so aiming for 7-5-4 seems to be called for. It looks doable.