I did consider putting a town on the coast 1N from the northern oasis, but that's only distance 3(.5) from Oasenstadt. I know RCP isn't nearly as important in Conquests as in Vanilla, but if I understood Alexmann's corruption article, a perfectly equidistant ring does still minimize the 'distance' component of the Conquests corruption algorithm output -- is that right?
I'm not an expert on this, but the way I understood the corruption articles on the Vanilla/PtW corruption model and the C3C corruption model, the difference is like this: in the old model, all towns with same distance share the same "rank" value (allowing some extreme exploits like jumping your palace to a remote isolated area and leaving almost all your cities world-wide with zero corruption...), while in the new model, this possible "exploit" was fixed by giving every city a different "rank" value.
So in our example, if we have 7 cities at distance 4 and then add an eighth city at distance 3:
a) in Vanilla/PtW, the former 7 cities would first have rank 1 (with pretty much neglible corruption), and then, when the eighth city is founded, they would all drop to rank 2 (increasing corruption in all of them significantly), while the new city now gets rank 1. If the new city would be founded at distance 4 as well, then all 8 cities would keep rank 1 and only have minimal corruption.
b) in C3C, the equi-distant towns have rank 1 - 7 (according to their order of founding). If we add an 8th city at distance 3, they would now get rank 2 - 8, while the new town gets rank 1. If we add the 8th city at distance 4, it would get rank 8, while the first 7 keep their ranks. In both cases, we would end up with a total of 8 cities having rank 1 - 8, so once these 8 cities are fully grown, there is no difference at all in terms of corruption. (However, in the period following directly after the foundation of the 8th city, there is some impact: the older 7 cities can be assumed to have grown a bit and be well-developed, while the new-found town is size 1 and doesn't produce much yet. So the well-developed cities all get a slight corruption penalty, while the still unproductive size-1 town gets the best corruption value... This is of course not so desirable. But as I said, this effect goes away, once the new town is also fully developed and can therefore take full benefit of its low corruption. And this effect can be avoided/minimized by founding the distance-3 town quite early in the expansion phase.)
Well, to make a long story short: in C3C other factors (access to a river or to early resources) are much more important than the above slight effect on corruption, so Ring-City-Placement is basically "dead" in C3C. We don't need to pay attention to it.
If so, and we're going to build a distance-3 ring with maximum coastal cities, then placing cities at the cardinal points relative to Oasenstadt (hills 2W, plains at 3NW, 3SW and 2N) would give us 4 coastal cities. But wouldn't this be too crowded?
It might get a bit too crowded, but we'll see once we have a better map of our lands. We should make sure that every first-ring city has 12 tiles to work on, because that makes the most out of our multiplier buildings. For example: if we can work 6x12 = 72 tiles with 6 cities (case a), we need 6 markets, libraries and universities. If we cramp 8 cities into these same 72 avaiable tiles (case b), we need 8 markets, libraries and universities. The total output would be approx. the same (not quite, because in case b, cities #7 and #8 would have higher corruption so the output of the tiles worked by these cities would be reduced compared to case a), but the investment would be higher (480 shields more needed for the additional markets, libs and unis) and the maintenance cost for the additional 6 buildings would cost -8gpt.
But if we can work additional tiles (especially high-commerce coastal tiles) that can't otherwise be accessed, we should build additional towns.
If we're going to use the distance-4 ring after all, then my vote would be to send our first couple of settlers along the riverbanks: one downstream (WNW) to build 'Kolossusstadt'(?) on the coastal plains as suggested
followed by one upstream (SE), aiming to build on the river (where it crosses the ring).
Yes, I like these two spots too. Especially as there is more floodplains upstream, which might result in a good spot for a worker-pump or (still hoping for a wheat...
) a second settler pump.
I know a granary is essential for a settler farm, but do we need to build it so early? Oasenstadt will grow so fast, might it not be better to put those 60 shields into 2 settlers and/or a settler and some workers? Also, we've got so much floodplain available that I'd almost be inclined to break the 'mine green, irrigate yellow' rule of thumb and mine some plains or desert (not the oases, obviously!), which would be much quicker than mining the hill.
I had the same feeling. One possible reason for a settler before granary I have already given above: in order to secure Colossus. If we discover let's say a wheat upstream, that might be another reason that justifies another early settler. Also: with 3 irrigated floodplains, our growth rate is so high and the shield output so low, that during the time it takes to finish the granary, our single worker would not be able to keep up with tile improvements to match Oasenstadt's growth. We would end up working unimproved tiles (with only 1 food), slowing down our expansion rate... During the time it takes to build 2 settlers or a settler and a worker, the worker has time to prepare more tiles and the capital can be kept down to size 4-6.
However, I'm not fond of poprushing, even in floodplain starts like this. The results that cracker got in his early experiments (see cracker's Opening Play Site) did not appear too promising to me, and the happiness penalty is too high a price to pay, especially on Emperor or higher, where happiness is always a problem and considering that we want to get out of Despotism asap, so we need every single gold coin we can get for research.