cottage spam -- a test

[xifeng]

When I first play Civ4, I always wonder when should I build cottage. Should I build some farms to boost population first, or work on cottage right away?

I just ran a small test. A city has 10 tiles of grassland, 5 tiles of plain, and 5 tiles of hill. The first time, I build 5 farms on the grassland and focus on food first, then work on 10 cottages.
The second time, on the same map I build 10 cottages on grassland first, and then build 5 farms on the plain.
The city has a palace, a global theater, a granary(more realistic), 15 health (support the city grow to 16 pop). A river run through 9 tiles of the city, but I think it is almost irrelevant coz the same map is used in both scenarios.

Farm first approach:

year turn population gold per turn
4000 0 1 10
3680 8 2 11
3560 11 3 12
3440 14 4 13
3320 17 5 14
3240 19 6 16
3160 21 7 18
3040 24 8 19
2920 27 9 20
2800 30 10 22
2680 33 11 25
2560 36 12 28
2360 41 13 31
2160 46 14 34
1800 55 15 39
1320 67 16 46

At the 67th turn, the city will produce (base income)
5 farms * 0 = 0
7 villages * 3 = 21
3 hamlets * 2= 6
1 cottage * 1 = 1
total = 28/turn
The city has assets of 1692 golds (including income from palace and river bonus).


Cottage first approach:

year turn population gold per turn
4000 0 1 11
3600 10 1 12
3560 11 2 14
3280 18 3 16
3000 25 4 19
2720 32 5 23
2400 40 6 26
2040 49 7 31
1680 58 8 34
1320 67 9 37

At the 67th turn, the city will produce (base income)
0 farms * 0 = 0
5 villages * 3 = 15
2 hamlets * 2= 4
2 cottage * 1 = 2
total = 21/turn
The city has assets of 1533 golds (including income from palace and river bonus).

Conclusion: the farm first approach is superior. Not only it generats more income, but also gives the player flexibility to switch to hill tiles for production.

 

[civn]

Nice work dude . I think your research show that it's better to build farm on grass than on plain tiles.

It's probably good idea to mention who's the leader . As for Washington, you'll also getting some extra hammers when switch to Universal Suffrage and your cottages matured; and since you head for cottage first, your population grow more steadily so you'll get less "it's too crowd".

Even though cottages are easy quick and dirty build, but who in the right mind will build 10 consecutive cottages early on, you can't use all 10 cottages that early. You better off alternate between farms and cottages.

When I heard ppl say cottage spam, it means build cottage if you are unsure what to build.

When I first play Civ4

 

[Hans Lemurson]

Commerce I view as "soft" production, whereas hammers are "hard production". In the beginning of the game, hard-production is much more valuable. Also, food allows quicker growth, so your cities reach their full potentials sooner. As your empire gets bigger, an expansion begins to wind down, you can start to focus more on soft-production, but having at least one city dedicated to it from the beginning is a good thing. Food is much more important than commerce, since larger cities are more productive. Cottage spam is best combined with intensive farming to make the cities grow so that they can actually use the cottages.

Farms-first is the way to go.

 

[xifeng]

when you look at the data of the 67th turn, you will find out that over time the farm first approach will have more towns than the cottage first approach.

It is 7 villages + 3 hamlets Vs. 5 villages + 2 hamlets.

I use a non financial leader to test. But I believe the trend will no change even if I use financial leaders.

 

[cabert]

interesting test, but not quite right, since you have tweaked the rules very much!
indeed, the growth of population is steadier when you farm first, but you obviously need to build this granary, and you hit the health and happiness cap quite soon without those special gifts...

 

[cabert]

when you look at the data of the 67th turn, you will find out that over time the farm first approach will have more towns than the cottage first approach.

It is 7 villages + 3 hamlets Vs. 5 villages + 2 hamlets.

I use a non financial leader to test. But I believe the trend will no change even if I use financial leaders.

i believe the financial trait makes the early gold income very different!

 

[xifeng]

Yes, the health and happiness factors do play a role in a real game. In a real game the city I use in the test will be my 5th or 6th city, but for a simplified test I just set it this way.

Without granary, the cottage first approach will suck more.

I will run a test for financial leaders.

 

[omni_paul]

Also floodplains yeild as much food as many farmed squares already, so cottaging them from the outset is a clear winner.

 

[malekithe]

It is, of course, quite true that farming first will eventually lead to higher total commerce production (or hammers for that matter). The downside, however, is that this increase in total output comes at a price. That price is short-term output. How much commerce are you willing to sacrifice right now in order to be gaining more in 50 turns? The answer to that question is very dependent on what point in the game you're at.

If this test were run again, this time with 6 farms instead of 5, I think you'd find the output in the 6 farms case to be higher still. So, why choose 5 farms over 6? The answer must come down to your time-preference for commerce. How long are you willing to delay the pay-off? The reason cottage spamming in the beginning is so popular is because at the start of the game, people have a very high time-preference. They need bronze-working, writing, and alphabet and they need them now. It doesn't matter how much commerce you're producing in 50 turns if you have any army of horse archers breathing down your neck and no counter available. You can't throw loaves of bread at the invading army.

The other thing that can alter this experiment is your unrealistic (in my opinion) population cap. If I'm looking at a city that needs to grow from 1 to 16 population, I figure it's probably a city founded or conquered late in the game, and will be relatively inconsequential in the outcome. If I throw down 5 farms in a city founded before 1AD, though, I'm going to be able to grow to maybe 10 population (7-8 may be more realistic). So, what do I do with those farms when I hit 10 population? Let them sit around doing nothing, or replace them (and their very nice grassland/river tiles) with cottages? Most people would do the latter. So really, you're only farming for growth up to some intermediate cap. Afterwards, your farms will likely be torn down in favor of some more productive improvement.

Farms are nice, but they are rarely an end. They are merely a means. While they may be a very efficient means in many situations, that is not always the case. Only a combination of intuition and in-game experience can truly determine what the "optimal" number of farms is for a given city.

Attempting to take cottage spamming down a peg or two is a very noble endeavor, which I can get behind whole-heartedly. However, all you've shown is that most cities would benefit from a farm or two along with their mass of cottages. You're still going to need to get commerce eventually, the more the better. Cottages remain the most effective way of doing that. If there were an equally effective way of generating beakers and gold that remained potent throughout the game (or even situationally superior), then there would be a real alternative to large numbers of cottages. You can argue about specialists, but I'm a firm believer that a specialist-based economy can never hold a candle to a cottage-spammed empire highlighted by the occasional specialist. Not in the early game, not in the mid game, and certainly not in the late game.

The only real alternative to cottages eschews commerce altogether after a certain point. It is the early warmongering strategy. In this strategy, hammers are king as you attempt to push out as large a military as quickly as possible, using your plunder to fuel any further neccessary research. For the mostly peaceful player, however, there is no real alternative to cottages. Knowing to farm first, just makes you a more efficient cottage spammer.

 

[xifeng]

same test using financial leaders:

Farm first approach:

year turn population gold per turn
4000 0 1 10
3680 8 2 11
3560 11 3 12
3440 14 4 13
3320 17 5 14
3240 19 6 17
3160 21 7 20
3040 24 8 21
2920 27 9 22
2800 30 10 24
2680 33 11 28
2560 36 12 33
2360 41 13 38
2160 46 14 41
1800 55 15 37
1320 67 16 56

At the 67th turn, the city will produce (base income)
5 farms * 0 = 0
7 villages * 4 = 28
3 hamlets * 3= 9
1 cottage * 1 = 1
total = 38/turn
The city has assets of 1971 golds (including income from palace and river bonus).


Cottage first approach:

year turn population gold per turn
4000 0 1 12
3560 11 2 16
3280 18 3 19
3000 25 4 23
2720 32 5 28
2400 40 6 32
2040 49 7 38
1680 58 8 41
1320 67 9 44

At the 67th turn, the city will produce (base income)
0 farms * 0 = 0
5 villages * 4 = 20
2 hamlets * 3= 6
2 cottage * 1 = 2
total = 28/turn
The city has assets of 1817 golds (including income from palace and river bonus).

The gap between the total commerce produced are slightly closer, but the Farms approach is leading 10 base commerce/turn at the 67th turn.

So if a financial leader uses farm first approach, her advantage will be even bigger.

 

[cabert]

Yes, the health and happiness factors do play a role in a real game. In a real game the city I use in the test will be my 5th or 6th city, but for a simplified test I just set it this way.

then it won't have the palace commerce, and you'll have only the center square and river commerce! big difference here!

Without granary, the cottage first approach will suck more.

it's stupid to run a food rich strat without granary isn't it?
my point wasn't to test without granary but to point out that you need to build it! some good chopping to do

I will run a test for financial leaders.

i'm anxious to see it. I'm pretty sure early commerce is very different.

edit : you're faster than i am

 

[xifeng]

malekithe, I agree with what you said.
In fact, I cottage spamming every game coz I use financial leaders. As you predict, I just want to find out a way to spam more efficiently.

The reason that I use 5 farms is purely random. It is 1/3 of the workable tiles and it provide the food for me to work at the other 2/3 tiles. If the growth of the city will be capped at size 10, of course a player will choose build 3farms instead of 5. With food resources the optimum farm number will be even smaller.

 

[atreas]

@xifeng:
I have run this test already for many happines caps and starting populations and the result is (as you already saw yourself) a win for farming first. There is a slight delay in the start, but it is so small (in turns and commerce) compared to the continuous gain afterwords that it's clearly better to farm. Still, this is true ONLY FOR SPECIFIC SETUPS, for the following reasons.

The first drawback in your analysis is that you make a very convenient assumption for your case - that the city has no food resource available. The results are very different when there is a +3 Food resource (to say nothing for the case of a food resource in combination with some floods); in this case you shouldn't farm anything else, but start immediately with the food resource and then work on cottages.

The second problem is that it is very rare to have a city which can grow without halts from pop 1 to pop 15. One common case is to see the city growing from cize 1 to cize 5-7, wait for Monarchy/Calendar, and then grow to size 15. Since now the first city cap is very smaller, it's difficult to find a farming combination that is beneficial when you also have any food resource available. You should run two separate calculations (1 -> 6 and 6 -> 15, for example) to be more realistic. In the second test, one crucial parameter is not to enter very fast into the "plains zone", because it will slow you down.

For more information, you can also have a look at http://forums.civfanatics.com/showthread.php?t=165822

 

[xifeng]

For more information, you can also have a look at http://forums.civfanatics.com/showthread.php?t=165822

Oops! I didn't notice that this topic had been discussed recently.

I rethink it last night ,and I realize that building 4 farms upfront can be very useful even if the pop size cap is 8!
After the size reaches 8, you can always swithch the work force from farms to cottages. The city will reach size 8 at the 24th turn. At the same time, the cottage first city will be only at size 4. Both cities will be working 4 tiles of cottage. However, for the size 8 city you can now switch another 4 to the cottage tiles or change the farm tiles into cottage tiles if those tiles are more desirable.

It is true that weak beginning cash flow (commerce flow) is a problem; however, the flexibiliy option will be a very good compensation for it.

I think that is what Bezhukov means in the #52 reply of your post.

 

[Zombie69]

Conclusion: the farm first approach is superior. Not only it generats more income, but also gives the player flexibility to switch to hill tiles for production.

Only because your city didn't have any special resources. Since most cities have at least one food resource, or some flood plains, lakes or oasis, cottage first is normally superior.

Also, most cities don't have 15 health and happiness right from size 1, again making cottage first more efficient.

Edit : i see that atreas beat me to it.

 

[DaveMcW]

A more common case:

6 happiness limit
6 river grassland
1 worker
Pottery and Agriculture
NO buildings
NO improvements

I think you will find farms are much less useful in this case.

 

[xifeng]

A more common case:

6 happiness limit
6 river grassland
1 worker
Pottery and Agriculture
NO buildings
NO improvements

I think you will find farms are much less useful in this case.

Your assumption couldn't be the "more common" case... This city could be your 4th city or 8th city, not just the first one. At that time, happiness is not a huge problem and workers are available.

 

[jar2574]

These tests are interesting, and will be useful for the rare city that has no resources. I just hardly ever settle somewhere without any kind of resource.

If you've got food resources then cottage spamming is the way to go. Before settling in an area without food resources you should settle all available food resource areas, and cottage spam.

The exceptions are production cities, meant to create your military units, and GP cities. In production cities you should build only farms and production improvements. In GP cities farms and food improvements are what your workers need to build.

Before bothering to settle in bad locations it may be better to build military and take out opponent cities in good locations. In Civ4 location is everything.

 

[Tigger70]

You might want to take a look at a strategy article I wrote some while ago on city growth. The central thing I did there was to generate a chart for how fast cities will grow at different food surpluses. It turns out that the growth rate isn't linear in the way I had intuitively expected that it would be. In a certain sense, a 5F surplus is maximally effecient (again, in a certain sense). Back when I had time to play a while ago I built this into the strategy for growing commerce cities, aiming for a 5F surplus where feasible. At any rate the chart provides concrete data about changes in growth rates given different surpluses. The difference between a 4F and 5F surplus turns out to be a very different thing for example than the different between 5F and 6F.

 

[xifeng]

You might want to take a look at a strategy article I wrote some while ago on city growth. The central thing I did there was to generate a chart for how fast cities will grow at different food surpluses. It turns out that the growth rate isn't linear in the way I had intuitively expected that it would be. In a certain sense, a 5F surplus is maximally effecient (again, in a certain sense). Back when I had time to play a while ago I built this into the strategy for growing commerce cities, aiming for a 5F surplus where feasible. At any rate the chart provides concrete data about changes in growth rates given different surpluses. The difference between a 4F and 5F surplus turns out to be a very different thing for example than the different between 5F and 6F.

I feel like I am back to the Corporate Finance class in which we calculate the optimum capital structure for a firm.
If someone creates a simulator program to run these tests, we will have some better strategies for sure.