Terrain Optimization (Phase I)

[DaveMcW]

The first assumption, that we have access to any sort of tile that we might want, is such that I find it very hard to see how this could be applicable. Conclusions 3, 4 & 5 saying that "X are never useful" are only true if you have an infinite choice of tiles, which is certainly *not* the case, so the conclusions are misleading at best, and probably more like worthless.

That's like saying nobody can play a flawless game of Civ4 without reloading. It may be true, but knowing all the tricks and theories behind a perfect game is very useful.

In this case, we want to know how close our city site comes to "max production", and DaviddesJ proposed a tool to measure that.

 

[BruceLeeee]

I agree with zienth. This is exactly like programmer oversimplifying to solve a problem, but it becomes so limited it doesn't really solve any.
I propose building a mod that tells you what improvement is best on a specific tile you are given. Can't be too hard since Civ4 already has an algorithm built-in. Just trick the game into thinking that you have 6 tiles with each improvement and see which one the game circles. Hopefully, we can modify the game algorithm with a better optimization formula.

 

[Zombie69]

Let's say that I think that anyone who picks one particular leader trait 90% of the time is missing the whole point of the game, and I don't really care whether my analysis is helpful to them, or not.

Let's say that Fireaxis should release a patch to balance all the traits, so that we don't have to restrict ourselves to one in particular.

 

[Smirk]

Let's say that Fireaxis should release a patch to balance all the traits, so that we don't have to restrict ourselves to one in particular.

I'm sure they are all just waiting for your every post to know what to do next.

Anyway no one if forcing you to only play finacial civs and whether its fixed or balanced in the future doesn't really matter. The debate of the effectiveness of this trait belongs in another thread. But you seem to have made up your mind while the rest of us are actually playing other civs, for other reasons and other game goals. For example, what good is financial for an early conquest?


Also its Fir Axis, the axis of an evergreen not fire. Ignore the symbolism of their company logo.

 

[DaviddesJ]

Let's say that Fireaxis should release a patch to balance all the traits, so that we don't have to restrict ourselves to one in particular.

As I already said, you don't have to restrict yourself.

I'm actually quite happy if some traits are better than others, as it gives another way to modify the difficulty of the game than just adjusting the handicap for the computer opponents.

This still seems pretty irrelevant to this thread. As I said, if you have a financial leader, just multiply all of the commerce values by 4/3 (or whatever factor you think is appropriate).

 

[nalbano]

wow.. what an unintelligable load of crap. was the intention of the op to be of help to anyone here?

 

[Gato Loco]

Nice effort here. I'm sure this will be useful in deciding how to develop cities. In my own consideration of hte subject, I've come to similar conclusions. Rather than saying X is or isn't useful, I think it would be more productive to rank different combinaitons in order of utility so they can be applied to real-world situations.

1. The city has access to an unlimited number of tiles of every "basic" terrain type (i.e., plains or grassland, hill or flat). The tiles may or may not have rivers; this doesn't really affect the analysis, as the city just gets +1 commerce for each river, regardless of other choices.

Obviously a very off-putting assumption. In its defense, if this assumption is used to rank different combination in order of usefulness, the ranking can then be used to decide among various possible combinations for a city site.

Assumptions 2 and 3 seem pretty reasonable, though certain special resources shouldn't necessarily be worked in every situation (e.g. ice/furs or desert/iron).

4. We assume that each cottage/hamlet/village/town is a village, worth 3 commerce; there may be some towns at this point, but there will also be some cottages and hamlets that haven't grown to villages yet. This value could be varied, with some extra complexity in the analysis.[\quote]

This is a very tricky point. I suspect it's an ok approximation for the early game, but cottages are tricky because of their small immediate value and immense long-term value. The value of a cottage in the middle ages isn't the 2-3 commerce its making now but the 7 commerce it'll be making later on with full growth and printing press/free speech, so you can't really treat cottages like other improvements. They're an investment for hte future.

5. We will also simplify the analysis by pretending that a city can allocate fractional citizens, e.g., 1/2 citizen on a grassland/flat/farm and 1/2 on a grassland/hill/mine. Again, this could be made more precise, with some extra complexity. Note that you can usually achieve an equivalent result to a fractional citizen by moving one citizen back and forth between different tiles on successive turns.

This is the secret to useful terrain analysis. Because food yields in one spot affect your population assignment elsewhere you have to consider not individual tiles but tile combinations (or yield-per-food).

3. Workshops are never useful, under these theoretical conditions, because hill/mine (-1f +3p) is strictly better than flat/workshop (-1f +1p).[\QUOTE]

Pretty much aggree. Even in a real game, you're in a pretty bad way if you have so few hills that you're reduced to using pre-guilds workshops to get enough hammers.

4. Watermills are never useful, either, because hill/windmill (+1p +1c) is strictly better than flat/watermill (+1p).

Here's where you'll need to slightly modify your analysis to apply it to real conditions. For a reasonably large city the choice will never be between working either a hill/windmill or a flat/watermill. Rather it'll be about what to do with the hill tile and the flat tile that you already have. For example) 1/2 grass/hill/windmill + 1/2 grass/watermill for [+0.5p +0.5c], 1/2 grass/hill/mine + 1/2 grass/farm for [+1p], 1/2 grass/hill/windmill + 1/2 grass/flat/village for [+0.5p +1.5c], or 1/2 grass/hill/mine + 1/2 grass/flat/watermill for [-0.5f +1.5p]. You can see here that the real choice is between optimizing for hammers or for comerce, as well as whether to run a food debt to get a bit more hammers out of the tiles. It does show, however, that the watermill isn't all that good unless you're absolutely going for broke with the hammers, which would take you outside the scope of this thought experiment.

5. Windmills are never needed, either, because a hill/windmill (+1p +1c) is equivalent to 1/3 of flat/farm (+1f) plus 1/3 of hill/mine (-1f +3p) plus 1/3 of flat/town (+3c).

Aggreed, with the reservations mentioned above. Windmills are a second-best improvement and wouldn't ever be useful under ideal city conditions where you can build mines, farms, and cottages to your heart's content. The other conclusions are similar.

Your math does seem to be reasonable, particularly the part about balancing food. It takes a lot of thinking and calculating in-game to apply this but it's still useful in helping to form intuitions about city spaces. Also, I think that your assumptions are particularly well suited for phase I. At this point I usually cherry-pick the best spots with the terrain I want, so not having enough hills or fgrasslands isn't a problem. For phase II, I'd suggest making two sets of calculations, one for mature cities founded during phase I and one for newer backfill cities founded more recently in poor spots (e.g. fishing villages, all hills, limited grassland, no food bonuses, etc.) Such extreme cases become increasingly relevant later in the game and are the reason Firaxis gave us workshops, windmills, and watermills. I look forward to your future work.

 

[DaviddesJ]

Yes, if I do more with this, I'm going to focus more on how to improve the tiles you actually have, not what the ideal mix of tiles would be. As you say, that becomes more relevant later in the game, because when cities are only working 5-10 tiles, usually you can choose the best ones. But, even earlier, you might, for example, have the choice between plains/river tiles and grassland/non-river, and it would be nice to have some theoretical analysis of that tradeoff.

 

[fed1943]

IMO,the tricky point to analise is that:a city needs food and hammers,never needs commerce.So,it may be would better to study about the tiles ímprovements(farms,mines,windmills,watermills,lumbermills,but not cottages)decide the necessary or desirable,the others would be cottages,or make a steadyfast decision about the commerce basic output of that city and the remaining...Well,I never got a satisfactory conclusion.
Best regards,

 

[Smirk]

In my games the overall question has really boiled down to do I cottage that luxury or improve it. And this is increasingly being applied to more and more resources, all except early food based ones. Since resources are so far flung in most maps trading is hardly valuable so even when I do connect them, its generally only one.
The reasoning should be obvious, getting +3-4 gold instead of a couple hammers is better because that gold can support another city which in turn can easily create more hammers than any improved tile. And with a financial civ this is even better, that 2f7g cottage is in essence 2f7g + 2f1h1g+ 2f(minimum) and it continues to grow thru out the game.

I do still mine hills, but thats about it.

 

[DaviddesJ]

Since resources are so far flung in most maps trading is hardly valuable so even when I do connect them, its generally only one.

I don't understand what this means. If you have a resource, either it's not a duplicate for you (in which case, building the improvement is necessary to give you access to it), or it is a duplicate (in which case, you can trade it to another civ for usually a lot more than 3-4 gpt).

 

[Gato Loco]

In my games the overall question has really boiled down to do I cottage that luxury or improve it. And this is increasingly being applied to more and more resources, all except early food based ones. Since resources are so far flung in most maps trading is hardly valuable so even when I do connect them, its generally only one.
The reasoning should be obvious, getting +3-4 gold instead of a couple hammers is better because that gold can support another city which in turn can easily create more hammers than any improved tile. And with a financial civ this is even better, that 2f7g cottage is in essence 2f7g + 2f1h1g+ 2f(minimum) and it continues to grow thru out the game.

I do still mine hills, but thats about it.

Don't forget rice. Rice = sugar = floodplain. You cottage you're floodplains, don't you? And those pesky calander resources that give you a paltry +1 at the start. I'd be wary about other food resources because your cities do need to grow, but I could even see myself cottaging wheat in the right circumstances.

 

[Jamppa]

How about we set like 10 different city sites with different landscapes; then optimize how terrain should be improved... This way we could calculate some kind of average value, right?

It would also be nice to have an article of how to improve "specialized" cities to get the most out of your terrain. Anyone up for it here?

 

[Brighteye]

What would be good is if someone listed a few basic rules for beginners like me which others can then debate or alter depending on their opinion, so here are some. I start from terrain production values assuming complete health and happiness: this seems like a better basic assumption to me. I will ignore resources, although actually they are very important in choosing city sites. It's really very hard to separate choosing city sites from terrain optimisation, so there's a bit of both in here.

Grassland is better than plains; if you like hammers you should be looking for 2:1 ratio of grassland to grassland hills, not for plains. You can farm plains for reasonable effect, but grassland is still better.
Unlike in previous Civ games farmed grassland gives you nothing but food, so it is only good for growing population.
This extra population is useful for mining hills but particularly for being specialists. I say this because if you use it to mine hills you're just doing what has been considered before and trying to optimise the average tile production, which can be done after founding a city, but if you want great people the specialist cannot support himself and so you absolutely must have farmed grassland to support him, and so a great person farm must be on grassland and floodplains.
I've seen a fair few comments about whether we should bother with farms. Two grassland farms still give a free population point, so you have a net bonus. Therefore farms are still worthwhile (before considering health and happiness). On plains they are worthwhile because the tile then supports itself and gives an extra one of those ever-elusive hammers.
So a very simple system to start your civilisation would be to farm river tiles, cottage other tiles and mine hills. Your priorities for cities should be river tiles, then grassland, always looking for a good number of hills within the city radius.
Once we have these basic rules we can then add extras. An important consideration is the financial trait, which adds an extra commerce to cottages on rivers. Thus you have a dilemma: you can reduce city growth by using cottaged tiles on a river but get extra commerce, or you can grab some commerce while growing the city by farming the river tile. Thus it's all dependent on how close you are to civil service for farming non-river tiles and whether you want short-term or long-term bonuses. I'd say that for greatest long-term bonuses you'd want to farm river tiles and cottage others, thus getting the commerce from the river while your city grows enough to use the cottages.
If your city is going to be a producer: your military unit city or other building city, you don't need to worry about cottages. Just mine those hills and farm other spaces to support the miners.
As previously mentioned, the base effects of workshops etc. are to optimise production of food/commerce/hammers in sub-optimal conditions. Thus you should be aiming to found cities in better places, at least to start with. Maybe once you have a bit more tech., and maybe state property, you can change the improvements around your core cities, but my scheme is a starting scheme.
I realise that on harder levels health and happiness are major concerns, but they change depending not only on difficulty level but on resources, which I haven't considered here. You can adjust the scheme to take account of your health/happiness by planning ahead. I have stated that extra population is never bad, but if you know the city size that you're going to start having problems you can reduce the number of farms accordingly. Farms help speed up growth, so even if you know you want two cottages in the end, it may be helpful to have farms to begin with so that you can be working those two cottages sooner. Thus you might have farms to support miners and growth, but when you give up on growth you can cottage over most of your farms and exchange your food production for cash. Either this, or support great people instead of production.
This leads me on to the point that I think that there's a choice between specialists and cottages. Your production cities will focus on hammers and not worry about this choice. A scientist gives you 3 science whereas a cottage gives you a variable amount of commerce, which in this thread has been said to average 3. However, specialists can add far more, such as 2 culture and 3 extra science, as well as great person points. On the other hand, 3 cottages support themselves whilst giving an average of 9 commerce, but specialists need 2 farms to support them, so the output is still only 6 science. How much do you value those great people?
I reckon that farms are very helpful to start with, and in producing cities (large hammer output), but I haven't seen great people help too much, and I'd say that three cottages (which will give far more output with a little time) are far more valuable.
Thus with the new health and happiness system you can't set a city and leave it, unless you cottage immediately and forget it, but there are some basic rules about improvements.

 

[Brighteye]

From what I've just said, I think there are a few things to look out for in city placement.
1. Floodplains and rivers
2. Grassland
3. Hills
4. Forest

There are 3 types of city: production, commerce and great people farm.
A production city needs hills, and might also have iron or copper to mine. To support these miners it must have tiles that produce excess food. Thus a production city wants a lot of grassland hills (about 2 for every 3 grassland in a resource-barren area) and grassland to support them. Plains are acceptable, but sub-optimal. Floodplains are also helpful, but could well be better used in a great-person farm. A production city only really needs rivers for fresh water and health. As long as there's a food resource to keep growth going before civil service a river isn't necessary.
A commerce city needs rivers and lots of grassland, with maybe a few plains or hills. Hills will be helpful to mine in order to give the city enough hammers to build the essential buildings, so try to have a couple in the radius.
A great person farm requires huge amounts of food, health and happiness, so look for floodplains, grassland, rivers to provide farms, and either a couple of hills or a lot of forest for building those helpful wonders. A farmed plains tile is a waste of one health; that one production would be better as half a specialist supported by a farmed grassland.
A city with a lot of plains will inevitably be a slow-growing city with a reasonable amount of production, but won't be the best of cities.
Coast is good for quick commerce, but cottages on grassland will soon overtake coast, and don't require a lighthouse, so coastal cities should be later foundings filling in the gaps, major naval ports or there for food resources.