Evaluating city spots

This is in part a response to Elear, but also something I'd like to simply bring up in general.

Elear said:

That's true. Could you point me towards a resource that explains what is "good placement" in the milking phase?

Click to expand...

I have been thinking a bit about scoring during a milk run and come up with a method that would allow to assign real values to city sites or clusters and so allow comparison between city sites and clusters. It isn't anything earthshattering, radically new or something.

By and large you just count food in order to figure out how many laborers and specialists and certain city spot could support, add the territory score and put that in relation to the number of dom limit tiles the site eats up.

For an ideal all grassland city without any foodbonuses that would mean:

There are 21 tiles overall.
20 of those yield 4 food => 80 Food
The City center yields another 2 (I'll just assume non-agri)
So overall such a city yields 82 Food.

40 of those go towards 20 happy laborers (I'll just assume they are all happy)
=> 20 x 2 (the score of a happy laborere) = 40
The remaining 42 food go towards 21 specialists
=> 21 x 1 = 21

In addition to the score from the population of the city there is also the territory score to consider. And 21 tiles territory simply mean a score of 21.

Adding all these up the happy laborers, specialists and territory score means that the absolute score potential of such a city is 40 + 21 + 21 = 82



But more meaningful is to put that score potential into relation to the amount of tiles such a city takes away from the dom limit. And of course this is the same here as the number of tiles that count towards territory, i.e. 21. So, the score potential needs to be divided by the number of dom limit tiles that are eaten up:
82 / 21 ≈ 3.9 (average max. score per dom limit tile)


Doing the same calculation for a similar all plains city (62 food) assigns such a city a value of
(40 + 11 + 21) / 21 ≈ 3.43


OK, that is not entirely unexpected. It get more interesting when Sea tiles are factored in. As it turned out in some thread around here (thanks esp. to Chamnix and EMan) those sea tiles do not count towards territory score, but they can still yield the score for housing a happy laborer. They also don't count towards the dom limit.

A city on a one tile island, where there are 8 coast and 12 sea tiles within the fat cross of the city would yield 42 food which can go towards 20 laborers and one specialist. There are also 9 tiles which count towards the territory score. Which all means that the absolute score potential of such a city isn't all that high and adding everything up nets merely 40 + 1 + 9 = 50 points such a city could score (on Chieftain, from turn 1 to 540).

But when putting that into relation to the the very few tiles such a city takes away from the dom limit it turn out that the relative value of such a city is much higher that that of even a maxed out, all grassland city:
(40 + 1 + 9) / 9 ≈ 5.56

Some "real world" examples:


Light Grey (77 food max):
(40 + 18.5 + 21) ≈ 3.79
Dark Grey (50 food max):
(40 + 5 + 14) / 14 ≈ 4.2


Light Grey (41 food):
(40 + 0.5 + 14) / 14 ≈ 3.9 (!)
Dark Grey (58 food):
(40 + 9 + 21) / 21 ≈ 3.33

(Leaving aside that the two cities' cultural borders would connect)

OCP vs tighter placements




==> (156 + 73.5 + 82) / 82 ≈ 3.8





==> (152 + 75 + 82) / 82 ≈ 3.77



OK, I hope I didn't miscount anything. Thoughts?

This is great, Emsworth.

But the one thing I am confused about is OCP...according to your calculations the highest scores yield from huge spaces between cities, but we almost never see this in any milk run.

The part about the sea tiles really helps me, I honestly thought that sea tiles might count towards the domination limit. So, you need to build some temples in such cities to get the full border expansions, then sell them off? I assume this will yield fewer cities in the end and is why CrpViewer tends to show steadily decreasing city counts in the "set-up" phase of milking.

Your calculations also seem sound, however, I think it still warrants more "real-world" analysis since the sort of scenarios you demonstrated are rarely the case with real city alignment. The method you proposed is very good in my opinion, though, and makes sense to me based on what we know of the scoring system.

-Elear

Elear said:

This is great, Emsworth.

But the one thing I am confused about is OCP...according to your calculations the highest scores yield from huge spaces between cities, but we almost never see this in any milk run.

Click to expand...

It is just where reality comes in to some extent. OCP for most parts of the map is merely an ideal, that might only work for some patches of the map. Any deviation from this ideal just means either you space closer, or you'll find entirely unworkable tiles in your territory. And there is plenty of reason to deviate. Mountains, Deserts, Coastlines, Food bonuses, rivers etc all will either force you to space a little tigher, or at least make it reasonalbe to do so.

And then there is of course this thing about culture that is needed to take in all tiles if you use real OCP, whereas CxxxC a little offset just offers a placement pattern without any needed culture.

And of course that for a long time, you don't have hospitals anyway. And, and, and.

Elear said:

Your calculations also seem sound, however, I think it still warrants more "real-world" analysis since the sort of scenarios you demonstrated are rarely the case with real city alignment.

Click to expand...

You are right. It is also geared merely towards score potential while leaving aside other important issues. In a real game for example, it is impossible to make all the laborers in a one tile island city happy with just luxes, entertainers and what your level gives you in terms of citizens born content. Even using temples etc would still not cut it. Your only hope would be war happyness in large doses for the entire milking phase. Cities with lots of water tiles also grow quite slow and may be needed to be joined up to full size.

Good analysis here... thanks Lord Emsworth! Though, I feel that I should mention that such an analysis ignores how fast cities will grow and what sort of shield production you can get from such cities for improvements and units (and commerce also).

Lord Emsworth said:

In a real game for example, it is impossible to make all the laborers in a one tile island city happy with just luxes, entertainers and what your level gives you in terms of citizens born content. Even using temples etc would still not cut it. Your only hope would be war happyness in large doses for the entire milking phase.

Click to expand...

Hmmm... sounds like a slew of commerce. Put in a centralized Palace, a courthouse and a police station, and raise the luxury slider to 90% or 100%. As long as you don't have a tundra location, you'll have at least one specialist in any city which you can use as an entertainer for another happy citizen. Maybe that's still not enough, but it seems close.

temple, cath (x2), colloseum, cure for cancer, 1 for highest level, 1 entertainer = 12 content faces.

8*3+12*2+1 = 49 * .2 = 9 content faces at 100% lux = 21 content faces in total.

So, if you are willing to do 100% lux, all you need is a temple, cath, collosum, courthouse, harbor and market on a 1 hex island to get completely happy

Market with 8 luxes = 20 happy faces

AutomatedTeller said:

8*3+12*2+1 = 49 * .2 = 9 content faces at 100% lux = 21 content faces in total.

Click to expand...

Hey AT, can you provide a detailed explanation of this formula. I get the other stuff you're saying regarding the wonders, improvements and luxuries, but I'm not sure what the 8*3.... formula is all about.

He's counting commerce from 8 coast and 12 sea tiles, plus 1 for the center tile. The city will be 80% corrupt, so the net is only 9 commerce. At 100% lux, this is enough to make all 20 citizens happy, provided the other conditions are met.

That is correct - each sea tile gives 2 commerce, each coast gives 3 and the center gives 2 (I put 1, but it's actually 2), which is 50 so 10 content faces - so I guess you can drop the colloseum (which is more evidence that colloseums are useless!!)

Ah, but the city is a metro, which gives more commerce in the center tile. Between three and five, I believe, for non-commercials depending on government and if there is a river. (Or is this different in C3C?)

oh - forgot about that. yes, a metro gives an extra commerce in the center square (2 for commercial) Still, it's not going to get to 11 content at 80%. you could add a police station, of course, which makes it 15 or 16, if you want, and can then drop the cath. Doesnt' matter all that much, I don't think.

So it's the marketplace that makes sea tiles provide two commerce instead of one? I guess I never paid that much attention. I thought the marketplace would calculate commerce as:

((8 coast x 2) + (12 sea x 1)) x 150% = 42 base commerce

You're saying it's a 50% increase on each and every tile, with a rounding up for sea tiles so:

(8 coast x 3) + (12 sea x 2) x 100% = 48 base commerce

Both calculations disregard the city centre.

I'm sure you're right about this, just double-checking.

You get extra commerce on sea and coast tiles from a commercial dock.

Sorry.

I was assuming republic (or democracy), which give the commerce bonus. market gives no commerce bonus. It gives you happy faces and multiplies tax income by .5

If you are in a government without the commerce bonus, you get 21 fewer commerce, so 30 * .2 = 6, plus the MP's (2 for despotism, 3 for monarchy, I think commie is 3 and fasicm is 4, but I don't remember)

with a commercial dock, in republic you get another 20 commerce (1 for each water, none for center), which is about 70 * .2 = 14. But the pollution adds to global warming (I think a dock adds pollution)