Finally found some time for a simple test with the standard settings. I wanted to test pure home cities first for clarity before adding a foreign city in.
3 cities as specified on the diagram one size 7 and 2 size 3 with the middle one having created all the archers... all 11 archers will be fighting nearest liverpool but having been created in Birmingham and London being more than twice the size.
The idea being I create some enemy, fire all my archers then delete the enemy... and repeat each turn as my WM increases, I can watch my cities
So here for example a few turns in
And here is the whole test with some limiting results, its like its decided to just provide 1 set of WW based on pop first and then proximity (no english troops were hurt in the making of this)
Here is the final T42 city screen
I was a bit confused so I increased liverpool to pop 7 and this was the result on T43 (ww accurate at the start of a turn) - no fighting this turn
So as its 2:10 am I up all 3 cities to pop 10 to see what it looks like and on T44 it looks like this - with 4746 WW (11.865*400) - no fighting this turn
T45 just for luck, upped to 13 pop in each ... wow - if you are going to war a lot, have small cities. WW is no problem at all.
A bit late to think more but pleased as the test does tell a lot of interesting info... feel free to jump in with your theories. To me it looks like cities can only get as many WW as amenities and it will allocate as many WW as a city has amenity requirements to the highest priority targets probably in the order of foreign, capital, closest. One suspects additional WW from deaths is then added? to the specific city the units came from.
A reminder of the parameters, the first set are the relevant ones