Alright if happiness is not the problem, how much can you grow your cities?
If cities will be built within 3 squares of each other to maximize number of cities on the map without crossing the domination limit, then there won't be 20 farms per city, but rather 10-12.
Say Sushi is giving +30
. Add food of 12 farms = 12 x 4 = 48. Total food generated is 78. You can sustain 39 citizens with that much food. 39 x 100 cities squeezed on a standard map = 3900 population points. Pretty damn good for a standard map.
On the other hand, growing to such city size will get gradually slower. At size 38, food surplus would be +2
, at size 37 +4
, at size 36 +6
. Growing from size 38 to 39 with a granary would take like (33 + (3 x 38))/2 = 76
(I have never had such a large city during my civ4 career)
Then with +2
surplus, it would take 38 turns to grow to size 39. Not worth the milking.
So while you can theoretically grow to size 39, 38 turns spent from growing from size 38 to 39 is such a long period in late game that the score benefit you get from having the extra population will be insignificant.
There must be a reasonable point to grow cities during the milking phase. I think it can be found through backwards growing the cities. Let me do the basic approximation quickly and you can tell me where I've missed anything.
Growing from size 37 to 38 with a granary, this time with a +4
surplus:
((33 + (3 x 37))/2)/4= 18 turns
Pretty good compared to the not practical 38 turns to grow from 38 to 39.
Growing from size 36 to 37 with a granay, this time with a +6
surplus,
((33 + (3 x 36))/2)/6= 12 turns
Growing from size 35 to 36 with a granay, this time with a +8
surplus,
((33 + (3 x 35))/2)/8= 9 turns
Growing from size 34 to 35 with a granay, this time with a +10
surplus,
((33 + (3 x 34))/2)/10= 7 turns
Growing from size 33 to 34 with a granay, this time with a +12
surplus,
((33 + (3 x 33))/2)/12= 6 turns
Growing from size 32 to 33 with a granay, this time with a +14
surplus,
((33 + (3 x 32))/2)/14= 5 turns
Growing from size 31 to 32 with a granay, this time with a +16
surplus,
((33 + (3 x 31))/2)/16= 4 turns
Growing from size 30 to 31 with a granay, this time with a +18
surplus,
((33 + (3 x 30))/2)/18= 4 turns (again 4 turns)
Growing from size 29 to 30 with a granay, this time with a +20
surplus,
((33 + (3 x 20))/2)/40= 3 turns
Before I stop, here is the calculation for a city size 20 and 25.
Growing from size 24 to 25 with a granary, this time with a +30
surplus,
((33 + (3 x 25))/2)/30= 1.8 turns ( roughly 2 turns)
Growing from size 20 to 21 with a granary, this time with a +40
surplus,
((33 + (3 x 20))/2)/40= 1.16 turns (roughly growth every turn, just like WastinTime suggested)
--------------------------------------------------------------------
Summary:
For an overcrowded map, where most cities don't have more than 12 farms to work and sushi giving +30
, it takes the following number of turns for cities to grow.
Cities until 20 population grow almost 1 citizen every turn.
Cities size 20-25 growth every 1-2 turns.
Cities size 25-29 growth every 2-3 turns.
Cities size 30-31 growth every 4 turns.
Cities size 32 growth every 5 turns.
Cities size 33 grow every 6 turns.
Cities size 34 grow every 7 turns.
Cities size 35 grow every 9 turns.
Cities size 36 grow every 12 turns.
Cities size 37 grow every 18 turns.
So total growth phase until size 20 lasts 20 turns. Until size 25 lasts 27 turns. Until size 30 lasts 40. Until size 35 lasts 71 turns. Until size 39 lasts (71+12+18+38)= 139 turns
Waiting your cities to grow 139 turns to max size is pretty long, so better is aiming for size 30 (40 turns) or 35 (71 turns)....
So ideal city size at the end of the milking phase is somewhere between 30-35....
---------------
Btw, have I forgotten to take the effects of unhealthiness into account? Does each unhealthy citizen also reduce the food surplus by -2
in addition to the regular 2
each citizen consumes?
I haven't played civIV long time so I am starting to forget the game mechanics, someone refresh my memory
-------------------
Anyway, I will assume the above calculation are correct. For a standard size map with 1000 max pop score, 100 cities * 30-35 citizens each = 3500 population score. Then a standard map indeed could have 3.0-3.5x max population score within 40 turns.
----------------
Until how long later could a larger map size achieve same 3.5-4x max population score and still score as high as the standard map size with same population to max pop score ratio?! Well settler and sushi executive production is the limiting factor. Once sushi is spread to all cities it is a matter of only 40 turns to reach practical milking limitation on any map size.
Producing 300 settlers or producing 100 settlers... Producing 300 executives or producing 100 executives? And would it take faster or slower to dominate a larger map size? I would assume slower, but again you might be getting so many great generals and experienced army that could make your advances even faster.
Those are the questions you need to answer to determine which map size would score higher, because growing your cities to size 30-35 would take similar amount of time on any map size.
Poprushing a 300
settler would take 3-4 citizens. Poprushing a 200
executive would take 2-3 citizens. So poprushing both would take. 500/90 = 5-6 citizens.
Say you had 50 cities with sushi before the milking phase all at size 15. In 3 turns, you will have 100 sushi cities.
Then you must wait some turns for the new cities to grow so you can poprush them. The new cities need granaries to be built as well for faster growth, but let's say you won't worry about unhappiness for a long time, so you will poprush the granary, the settler, and the executive within 30 turns.
Growth happens almost every turn for granary containing cities lower than size 20, so if you wait 15 turns you will have again a city size 15. Also you poprushed the granary as soon as the city size was 3-4, took you 1 turn only.
Now, 100 cities poprush, 100 settlers and 100 executives. You end up with 200 sushi cities. Then, you wait another 5-6 turns for cities to grow their population back. Poprush another set of 100 settlers and executives. You got yourself a 300 city empire.
300 city empire all with sushi took 3+1+15+5=24 turns of poprushing.
---------------------------------
Conclusion:
Total milking phase for a large map would last 24 turns to poprush the settlers and the executives plus 40 turns to grow all cities to size 30-35.
24+40 = 64 turns to stuff a large map size with 300 cities and 9k-10k citizens without crossing the domination limit.
On a standard map size producing 50 executives and 50 settlers would take only a few turns, but again those cities would have to wait 40 turns to grow to size 30-35. Thus total milking phase for a standard map would be roughly 45.
64 turns vs. 45 turns... the difference between milking a standard size map and a large map size with 3x more pop score is only 20 turns.
The advantage of playing a smaller map size is, well, you would have spent less time producing those settlers and executives. And you would have spent less time conquering the world, assuming similar domination speed for both map sizes. How many turns would that make in total? 20-30 turns? 50? or even 100? As I discussed in earlier posts, those number of turns are not good enough to compansate the higher score you would get with the same popscore/maxscore ratio for a larger map.
I know both of you prefer smaller map sizes, but showing the patience to play a large map would pay handsomely in the final score. That is the way the game is.
. I still enjoyed the victory and will try as Cyrus b4 I tackle deity. 
.
