Population

[Juvenal]

I have just finished my first EB0.8 campaign (Casse). I very much enjoyed the experience - it was the biggest step-change for me in RTW since my first mod (RTR4).

I notice that population levels seem to max out within 40-50 years (I was obliged to exterminate most of my later settlement captures).

I wouldn't pretend to know what population rise is historically correct, but the current rate feels rather high. The effect on gameplay (for me) is that by the time I have money for more than one army, I no longer need to worry about casualties because of the inexhaustable supply of men from my settlements.

Please don't take the above as a criticism, but I would really like to know what is the current thinking on population.

PS. If this has already been discussed, could someone please give me a link.

[Gazius]

The current thinking is that it maxes out way too fast in the community, and I've heard talk of this being fixed in a later version, not .81. This is one reason I'd suggest playing with huge units, even if you have to lower the quality of anything else.

[PseRamesses]

Well, I´ve been whining about growth rates since RTW came out. It´s unrealistic that a settlement can outgrow itself imploding in squalor and unrest and almost a full rooster of troops is needed for garrisoning.

A series of speciall buildings could be constructed to hamper immigrants, right?! Or why not "buildable" decrees? Takes one year to build (implement) and reduces growth by x% with higher growth reduction the larger the settlement gets etc. Buildable decrees, laws etc would be a huge improvement to the game IMHO. what do you all think?

[blacksnail]

This is addressed somewhat in 0.81, but we will need feedback with population examples of longer-term play to help us determine the appropriate balance between population and squalor on a faction-by-faction basis. This is something that takes a lot of in-depth playtesting to balance right, and as this is an open beta a lot of the playtesting comes from the community.

[Juvenal]

I wasn't satisfied with just a vague feeling that growth was too high, so I have tried to model an upper limit on population growth. I am not familiar with demography maths, so I am doing this empirically.

Assumptions
Settlement population is number of males aged 15-50 (men of fighting age)
Couples are of same age (for simplicity).
Couples have children between ages 15-40 (i.e. for 25 years).

Mortality Rates (males)
Infant (0-14) 50%
Adult (15-44) 0%
Mature (45-64) 0%
Elderly (65+) 100%

Birth Rate
Assume 30% of adult females (15-40) stop producing children before age 40, due to death in childbirth, disease, malnutrition, accidents and loss of fertility.
1 couple, married at age 15 would have 16 children (out of a potential 25) in 25 years. This gives 4 boys and 4 girls surviving to age 16.
4 boys per couple in 25 years means 0.16 male children coming of age per year per men between 15 and 45.

Settlement Demographic
Assume men are evenly distributed between 15 and 50 years of age.
Proportion of new parents per year is 25 out of 35 = 71%
Coming-of-age rate per year is therefore 16% of 71% = 11.4%
Death rate is number of men becoming 65 = 1/35 = 2.9%

Conclusions
Result is 8.5% growth per year. That is no more than 2.2% per turn.

Even if infant mortality is zero, we only get 5% growth per turn.

Growth should be additionally limited by adult male mortality (disease, accidents, crime) and adult male sickness (making them unfit to fight). These could be offset by health and law buildings.

In my recent Casse campaign I was sometimes getting 8% per turn from recently exterminated settlements, so I think growth rates need to be cut back drastically.

Any comments or criticisms on my method would be welcome.

[JeffSteel]

In the case of exterminated cities having high growth rates for a while, wouldn't people start migrating to a recently "cleared out" city in hopes of obtaining cheap land or something? Kinda like a "land rush" of sorts.

[Juvenal]

@JeffSteel

I think the settlement population represents all available fighting men in the entire province (less unrecruitable 400 men in what must be 'reserved' occupations).

Of course you could take the view that there is a pool of currently-unavailable men somewhere in the hinterland who might become available under the right circumstances. However this is an unhelpful concept because such an extra population is not recorded by the game.

Anyway - I personally would keep well away from a settlement with two thirds of the former population now forming a gently smoking heap outside the front gate.

[blacksnail]

In my recent Casse campaign I was sometimes getting 8% per turn from recently exterminated settlements, so I think growth rates need to be cut back drastically.

Extermination/enslavement is a bad example. There is always a boost to the growth rate of a city affected in that manner.

Any comments or criticisms on my method would be welcome.

We're trying for slower, more historically plausible growth rates in most cases. For example, Rome had a steady, roughly 1% growth rate per year during our time period. We're going to try to match that as well as we can, but it's a very careful balance between population growth and squalor.

[MarcusAureliusAntoninus]

I got tired of the squalor, so I added law bonuses to the academic building branch, and doubled the bonuses for garrisons and entertainment. It's a little cheap, but I also capped the city level to "city" for most places and don't want to have every town everywhere angry. Just from that little bit, I can get a bunch more people before everyone starts freaking out. (Also, the AI is building a lot more academies.)

[Pode]

It seems to me that one of the major difficulties is the granularity of the pop growth rate. 1/2% is the hardcoded smallest step change per turn, and 1.005 ^ 4 = a MINIMUM growth rate of 2% per 4 turn year. AFAIK, this exponential growth / shrinkage rate is hardcoded (non-integer pop growth bonus in the EDB causes a CTD). Again AFAIK, the add_pop script can only deal with a fixed number, not a formula (-487 instead of -1% of 48700). You guys know <lots> more than I do about scripting, but I don't see a way you can pull off any growth rate less than 2% / year short of regular bouts of worldwide single turn plague to bring the average down.

[QwertyMIDX]

At the risk of sounding redundent. I just posted this in the suggestions thread.

I think we should all wait and play some on the next build before worrying about this too much. We slashed BFL's (which provide the base rate of population growth among other things) by over 50% across the board. We from an average of about 5 to one a little under 2. That is an average drop in growth of 1.5% right there. My next goal is actually to make large and huge cities require lots of time and money invested in infrastructure to get. That's for the build after next, but rest assured, we're working away at it.

[McHrozni]

I wasn't satisfied with just a vague feeling that growth was too high, so I have tried to model an upper limit on population growth. I am not familiar with demography maths, so I am doing this empirically.

Assumptions
Settlement population is number of males aged 15-50 (men of fighting age)
Couples are of same age (for simplicity).
Couples have children between ages 15-40 (i.e. for 25 years).

Mortality Rates (males)
Infant (0-14) 50%
Adult (15-44) 0%
Mature (45-64) 0%
Elderly (65+) 100%

Birth Rate
Assume 30% of adult females (15-40) stop producing children before age 40, due to death in childbirth, disease, malnutrition, accidents and loss of fertility.
1 couple, married at age 15 would have 16 children (out of a potential 25) in 25 years. This gives 4 boys and 4 girls surviving to age 16.
4 boys per couple in 25 years means 0.16 male children coming of age per year per men between 15 and 45.

Settlement Demographic
Assume men are evenly distributed between 15 and 50 years of age.
Proportion of new parents per year is 25 out of 35 = 71%
Coming-of-age rate per year is therefore 16% of 71% = 11.4%
Death rate is number of men becoming 65 = 1/35 = 2.9%

Conclusions
Result is 8.5% growth per year. That is no more than 2.2% per turn.

Even if infant mortality is zero, we only get 5% growth per turn.

Growth should be additionally limited by adult male mortality (disease, accidents, crime) and adult male sickness (making them unfit to fight). These could be offset by health and law buildings.

In my recent Casse campaign I was sometimes getting 8% per turn from recently exterminated settlements, so I think growth rates need to be cut back drastically.

Any comments or criticisms on my method would be welcome.

Yeah, you forgot migrations. Settlements are only city population, the majority of folks lived in the countryside. At the height of the Roman Empire, perhaps 30% of population were city dwellers (admittedly this wasn't surpassed until the 1870s or so), but for the most of the history and most of the regions, the number was probably closer to 5-10%.

So what happens after you exterminate a settlement? People from the sorrounding areas move closer to the city - cities were usually near rich farmland (that's what determined most city sites to begin with) - and soon came to live in the city.
You would also probably encourage some migrations by yourself through colonists, to improve the control of the area. Romans were particularily fond of that, but hardly unique.

McHrozni

[Juvenal]

@McHrozni

Thanks for your thoughts...

It is not entirely clear (to me) in the game what the population figure actually represents.

The extermination mechanism seems to imply it is the population of settlements - because extermination is performed instantly when a settlement is captured.

The recruitment mechanism seems to imply it is the available male fighting age population in the whole province - otherwise why should my settlement having only 400 population stop me recruiting troops from elsewhere in the province?

Buildings seem to send mixed messages, some supporting settlement-only (cess-pits, temples, taverns) and others supporting whole province (farms, ports, roads, government).

I think the settlement also represents all other settlements in the province, therefore its population can likewise represent the whole population of the province.

The biggest argument against the hinterland view is that it is not modelled anywhere in the game - you only have one population number for a province, you can only control (or lose) the whole province, and you cannot benefit from your presence in a province you don't fully control.

Of course some things still fall outside of the model, such as mercenaries and rebels - they don't affect province population at all!

[McHrozni]

@McHrozni

Thanks for your thoughts...

It is not entirely clear (to me) in the game what the population figure actually represents.

The extermination mechanism seems to imply it is the population of settlements - because extermination is performed instantly when a settlement is captured.

The recruitment mechanism seems to imply it is the available male fighting age population in the whole province - otherwise why should my settlement having only 400 population stop me recruiting troops from elsewhere in the province?

Buildings seem to send mixed messages, some supporting settlement-only (cess-pits, temples, taverns) and others supporting whole province (farms, ports, roads, government).

I think the settlement also represents all other settlements in the province, therefore its population can likewise represent the whole population of the province.

The biggest argument against the hinterland view is that it is not modelled anywhere in the game - you only have one population number for a province, you can only control (or lose) the whole province, and you cannot benefit from your presence in a province you don't fully control.

Of course some things still fall outside of the model, such as mercenaries and rebels - they don't affect province population at all!

One thing is pretty clear, the system is quite imperfect

Take Carthage, for example. It usually grows to about 40,000 people, if you set your mind to it (low taxes, good leader), you can get it to about 50,000, but not much more. Even if this is just military age males fit for service, that translates to perhaps 150,000 people. Carthage had about 250,000 inhabitants at it's height.
Don't get me started with Rome and it's 1-1.4 million people at it's height, please Even with the above calculation, you only get slightly above 1/10th what it should be
On the other hand, many cities that were rather minor can get 25-30 thousand inhabitants eventually, with good care and investment. In some instances, I guess a "what-if" comes in, if the city was indeed developed to this extent, it could've grown this big, but in some cases, hardly

So, overall, the system is very imperfect, since it permits high growth while preventing to reach historic size of some cities - you can't import grain from Sicily and Egypt, which allowed Rome to grow that big.

As for which buildings do what, it is arguable. Okay, stone walls are a single wall around the largest settlement in the province, but wooden palisades would still ring larger villages in the countryside, at least in border areas. This is ignored by the system, obviously.
Watch towers, on the other hand, are province-wide installation, as are farming upgrades.
Many come as slightly dubius, a holy site can be something built in several places in one province, and a field of martial games likewise, but upgrades of both are rather one only affairs.

My thoughts? Stop thinking so much how the system could/should/might/would work, and enjoy what it offers A system that would allow you to micromanage every single detail would quite likely be too complex for enjoyment, or you'd have to leave it to the AI, killing the benefits of such a system (Master of Orion 3 anyone?).

McHrozni

[Juvenal]

@McHrozni

Don't forget that military unit sizes are scaled down, this means settlement population must be identically scaled down since recruiting a 160 man unit takes 160 from the population total.

What is the scale factor?

Well with game scale set to 'large', if an 80 man cohort represents a real cohort of 500 men (http://members.aol.com/FlJosephus2/romanArmy.htm) then the scale would be roughly 1:6.5

To get a simulated Rome population of 1,400,000, assume 350,000 men of fighting age (a quarter) meaning you need a game population of 54,000.

OK, this is bigger than Rome gets in my games, but it does actually look achievable.

This reminds me of another (hard-coded) problem - populations are the same whatever the unit scale (I think) - so population growth can only be properly balanced for one scale.


I agree that most cities should have a lower size cap, this would make Rome, Carthage etc much more important and interesting. I am sure I have read threads discussing this - but I can't remember whether it was EB or another mod.

-edit (remembered large scale unit sizes)

[Kull]

As QM pointed out above, slashing BFLs (basic farming levels) has changed the dynamic here quite a bit. In v.81 you will find that population increase buildings aren't something to be avoided like the plague.