Placing cities on ressources?

[Tomice]

In vanilla, people have created a list of ressources you should or shouldn't settle on. Since yields are very different in GEM, what do you think about the topic?

I'm not talking about purely strategic decisions like settling directly on iron to protect it from pillaging. It's all about the economic optimum.

 

[SlothACID]

Could you link me to the list mentioned above?

 

[Tomice]

Sorry, I couldn't find it any more.

What I found is that the higher the potential yield later on is, the less should you settle on the ressource - logical. In vanilla, there was a patch where yield increases through later techs were added (e.g. Dynamite---> Mines +1), which was seen as reason NOT to settle on ressources any more. Of course, there are exceptions, like if you want to sell a ressource ASAP to finance another settler (The famous tradition 4 city opening).

But vanilla is a different story, we play a significantly improved game here! Note that boosting yields from improvents more than once was one of Thals earliest ideas and among the many that made it into vanilla

I don't know if Thal changed the yield you get from the city tile itself if placed on a ressource (in vanilla basically only tiles with 3 base food increased the yield of the city tile). If he has already balanced this, there might already be a clear strategy to follow.

If a consistent design isn't yet implemented, we might start discussing it. But first we'd need to find out what the status quo is, which is no small task given the many combinations of base terrains, ressources and improvents. Not to forget freshwater access.

I guess I'll just do a bit of basic research myself, then If I find something useful, I'll post it here.



EDIT: FIRST RESULTS

I used the Ingame Editor Mod to place cities in various spots. From barren desert to riverside grassland, a city seems to usually give 3 and 3. Exceptions to this are luxuries, which give 1 extra gold, which seems logical. Cattle, sheep and wheat don't change the city tile yield, settling on them is a waste it seems. A discussion if that's appropriate might be useful. Also, AFAIK, granting my civ all techs did not change the yield in any city, even after waiting for 1 turn and seeing the increased yields on other tiles.

I did find something very strange, however:
If you place a city on a hill with either iron or horses (giving 3 even without improvement), the city will have no less than 7 3!!! Placing a city on grassland horses only gives standard yields, though.

Is this a bug or a feature?

 

[Naeven]

I don't think you should get any yield bonus from settling on a resource. Or maybe +1 for strategic resources since you can't see them from the beginning.

I also think yield bonuses like this leads to simpler settling decisions = less fun.
Example: I can't optimize all resources I want under 1 city, but if I put it on the resource in the middle, it works out!
That decision should come with the cost of the yield IMO. I also think this happens all the time and forces me slightly off the perfect spot.

If you want to settle on it to protect it, that's fine. But that stratigic decision should cost you the bonus yield.

A minor "real life" note; I don't see it as practicall to get "bonus yield" from wheat/cows/goats/marble/other if there is a whole city on it!
farms/plantations/quarries/other take space, space you can't build a city on

 

[Thalassicus]

It's neither a bug nor a feature, because the mod does not directly change this. I have not seen what might control city-on-resource yields in the game files, though I haven't done a thorough search. It would be an interesting topic to explore further.

 

[Tomice]

Well, then it seems like I've discovered a little secret here

If we talk about balancing it, I think the middle ground is best. Neither should we get the same yields as if there was the right improvement on the tile (making it a no-brainer to place the city there for instant yield without need for a worker and superior protection), nor should it always be bad.

As of now, here's the status quo:

- Hills + Strategic: Always settle
- Flat + Strategic: Only settle for protection/instant effect
- Luxury: Often worth settling if the alternative city spots are all useful, fertile spots that can be improved well. If deserts are available, settle there!
- Food ressource: Avoid settling if possible

Note that this is no in-depth analysis, only something done in 20 minutes of toying around with a cheating mod. Please correct me if you have other opinions/thoughts!

Also, I'd say that at least the food ressources should be changed, since they are not very strong anyway, and settling on them is a total loss of their effect. They should at least get 1 extra food for the city tile similar to the extra gold from luxuries (which keep their main benefit anyway) . Also the 7 production from iron hill cities appear like a bug IMO.

 

[mitsho]

I really didn't think of that before. I just naturally avoided settling on them...

The logical payoff would be that they receive a bonus food but not the boosted things from later techs. So that it could pay off early on but it would make really no sense later on. The extra 3 production on strategic ressource hills does seem quite like a bug imho I agree.

 

[Tomice]

It is an old civ tradition that settling on special tiles is usually bad. But it's worth a thought if it has to be like that. Ideally, both options should have their uses.

We also shouldn't forget that while we lose extra yields because the city tile cannot be improved, we'd lose an improvable tile anyway, which might be something as valuable as a riverside hill.
In other words: If a city placed over a ressource gets only the extra yield from the ressource, but not the yields from the improvement, we'd already neither lose nor gain anything!

I'll correct my table above: If we settle on spices as example, we "only" get 1 gold extra and "lose" the extra yield from the plantation. But we gain the option to put a village or farm on another tile, so we're not as bad off as it first seemed.

The only exceptions are tiles like deserts which can't be improved anyway. Placing a city on them is always beneficial, since the plains tile next to it could be farmed. Is this right/realistic? Might be another topic worth discussing!

 

[mystikx21]

Right, you're dealing in opportunity costs to place it there instead of over there.

In general: placing on food bonus tiles is a no-no. And easy to avoid.
Placing on iron hill sounds fine because of the glitch but most strategics only should get a +1 because you cannot see they are there from the beginning. Oil is an exception for me because even improved they're not great those wells and oil can be important enough to get faster.

Keep in mind that you don't sacrifice any increases from buildings to resource tiles. You only lose out on tech improvements or policy boosts on the improvements. Beliefs should still work too if they affect wine or silver say. So it might still be worth it on a lot of tiles by the later game for camps or (some) pastures or plantations because farms or villages might be better still from investment in tech and culture. Less so on mines or farms (or bananas).

 

[Tomice]

@Mystik: Good points!

It would also be important to understand how the yield on the city tile is calculated. I believe it follows an "at least" rule. If I'm right, the city tile gets "at least" 3 3. That would explain well why only the extra gold from luxuries shows up, because the city tile doesn't have any "guaranteed" gold.

In the case of e.g. settling on floodplain wheat, the tile would have 3 food without any improvement. Building a city on it gives "at least" 3 food, so we don't see a difference to building our city on desert. If there was a way to give a tile 4 food without improving it, we should see 4 food on the city tile. Hmm... if the city had a granary, we should see 4 food on the city tile, but only in the case of floodplain wheat. Plains wheat + granary would probably only yield 3 food.

But that's just a wild guess, and the 7 hills can't be explained this way. Probably really a bug.



IMHO, the city tile should not get guaranteed yields, it should just act as 2 2 improvement. On average, it would turn out the same for most tiles (since most of them give 2 yields except for tundra/desert). But all bonuses would apply unbiased and choosing where to settle would be a much less "gamey" decision. What sense does it make to place your city on the worst spot available? That's neither logical nor intuitive for new players.

Sure, we would make it tougher to settle in really bad territory, but that's a small issue compared to a clearer, more intuitive set of rules. And it could be rebalanced easily (e.g. with a building providing 2 food only buildable if the city is placed on bad terrain).

 

[Naeven]

I'd advice against this. I think personally it would result in more game restarts to get a "perfect" spot.
I agree it doesn't make much sense to settle on worst terrain. But neither does it makes sense to build cities on resources, as their yield represent the tiles extra affinity to a special improvement (not talking about the actual tradeable resource). A city improves cattle as much as a village or farm, but these don't give it extra yield. Also, I don't think placing city on worst terrain is a big problem, I'm much more concerned getting as many resources as possible within range.

Placing cities on resources are more "gamey" than placing cities on bad terrain. In the end you can place a city anywhere. Aslong as you gather people and transport resources there it can function as good as any city.

That said, I'm fine with giving cities +1 yield on resources for the tradeoff but we shouldn't make it a "thing" IMO.

 

[Tomice]

I doubt there'd be many restarts due to the suggested change (apart from restarts leading to a crash with GEM, I suppose it's the "load modded game from ingame" issue)

First, there already has to be a code in place that avoids starting ON a ressource (or have you ever seen it happen?). I doubt deserts under your first settler are that common either.

Second, there are 1000 better reasons to restart a game than 1 extra yield under your first city



Sure, it isn't the most important issue, and it is more important to fix the bugs with the current version. But it would still be nice to have less "illogical" rules to consider when placing your cities.
Also, since tile size is not exactly defined in Civ (ranging from equator divided by number of horizontal tiles to half the firing distance of the most primitive bows) it doesn't have to mean a city would block the whole tile for agriculture or mining.

 

[Naeven]

Actually I was more arguing against the concept itself. As i said, I'm fine on +1 yield but not more.

Also, since tile size is not exactly defined in Civ (ranging from equator divided by number of horizontal tiles to half the firing distance of the most primitive bows) it doesn't have to mean a city would block the whole tile for agriculture or mining.

I don't really get this point at all. Put this logic to extreme, any improvment should fit on the same tile as any other improvment. Thus all improvments should improve all resources equally.

 

[Thalassicus]

3 city + 1 hill + 1 iron should be 5. I'm not sure where the game might be getting the extra 2, since improving iron does not increase its yield (beyond what a mine normally gives).

Changing from minimum city tile yields to improvement-style yields is feasible. Look at function DoHillProduction in Civup/1_Core_Roots/CiVUP_Events.lua. It's where I give production to hills. I could modify this to add yields of terrain under the city. Alternatively, I could subtract one production if it's an iron tile, to compensate for whatever's going on there.

function DoHillProduction(hexPos, playerID, cityID, cultureType, eraType, continent, populationSize, size, fowState)
    local plot = Map.GetPlot(ToGridFromHex(hexPos.x, hexPos.y))
    if GameInfo.Yields.YIELD_PRODUCTION.MinCity >= 2 and plot:GetPlotType() == PlotTypes.PLOT_HILLS then
        Plot_ChangeYield(plot, YieldTypes.YIELD_PRODUCTION, 1)
    end
end
LuaEvents.NewCity.Add(DoHillProduction)

 

[mystikx21]

It seems to be just iron/horse hills. I don't know if it does it on other strategics or not.

 

[Tomice]

@ Naeven:

It seems you understood me wrong. My idea would not increase yields from city tiles (on average)! Placing a city on a luxury would result in the same sum of yields it would have right now (1gold more than without the luxury). I'd only like to have a more consistent rule, which applies to strategics and food ressources in the same way.

In the current ruleset, Cities have a minimum, guaranteed yield of 6 (3 food + 3 prod). Depending on the underlying terrain, this rule "swallows" most additional yields. Usually only the extra gold from luxuries is applied, since the city tile gets no "guaranteed" gold.



If we move away from the "minimum" rule and change it to a "plus" rule, things could look like this:

Most tiles in Civ5/GEM give 2 yields before improving them (Grassland, Hills, Forest,...) and 1 more if it contains a ressource. If we ADD 2 2 to the city tile without ignoring the base terrain, we would receive the same yield we get now in most cases (a sum of 6 yields). Each ressource, not only luxuries, would grant 1 yield more (7 altogether). No city tile would ever get more than these 7 yields (before applying social policies, buildings, wonders and such stuff of course!).

So I'm not suggesting a significant increase of city tile yields, only a more consistent rule without "swallowed" yields, and a rule that treats luxuries and other ressources similarily.

I also hope this would make city placement more strategic by giving us multiple equally viable choices instead of having pseudo-choices where one is clearly better than the other.




A few more details:

• Cities would basically act as 2 2 improvements in my suggestion instead of following a seperate rule.
• When talking about city tile yields, we always have to take another tile into consideration - the one we'd build the city on otherwise and which can be improved normally by a worker because we didn't put a city on it (Important for deserts!).
• If we have 2 similar tiles (e.g. 2 flat plains) and one contains wheat, by the old rule placing the city on the wheat would be a waste of 1, because the extra yield from wheat would be "swallowed". In my suggestion, there'd be no difference in combined yields of the two tiles.
• Of course we don't always choose between equal tiles when placing our city, but for all "good" tiles (those who give 2 yields unimproved like hills, forest, plains...), the possible improvements should be roughly of equal value.
• An interesting topic are desert, tundra and other "bad" tiles. With the old rule, they are ideal for city placement, which is weird (most cities start as villages on the sweetest spot available. Only very large, mature cities take up enough space that the loss of farmland is a major concern).
• In my suggestion, it would still make some sense to build on them, because a city is the only "improvement" allowed on flat deserts. But the city tile would have only 4 yields, because the base terrain offers none.
  If we choose between desert and plains for city placement, we could either have a plains city tile with 6 yields, an useless desert tile, and our first citizen working a third tile. Or, as alternative, a desert city tile with 4 yields and the option to farm the plains tile for 1-2 extra yields. In this case, the first option - the realistic one - would give better yields, but if other factors demand it, placing the city on the desert tile could still be better overall.

Of course, with some severe bugs plaguing GEM right now, this is no high priority issue. But since the bug where iron/horse hills allow city tiles with whopping 10 yields needs to be solved anyway, we might as well discuss were we want to go afterwards. I wouldn't open another pandoras box before we understand where the "iron hills bug" exactly comes from, however.

 

[Naeven]

I understood what you ment in practice

sure it's only 1 yield from the resource, but with the other proposed city changes it's more yields moving around. we now gona have 4 2 cities, VS 2 4 cities. Sure that adds up to 6, same as before. But they are not equal yields. specialy not early on.
So now all of a sudden you gonna have some players starting on 5(1 food resource) 2 VS 2 4 cities. That's not equal starts at all, thus leading to more restarts giving me the best spots. And the AI probably won't have as good spot, therefor nerfing it.

Sure, the whole concept adds some flavor. But it's costing more then what it's worth.
I like equal starts, I advice against these changes.

 

[Tomice]

You have a point

But the map generation doesn't put our first settler on ressources, it's always a standard tile AFAIK. So there won't be 5 starts. I also don't think that the difference between 2 4 starts and 4 2 starts is that important compared to starting coastal, on a river, next to marble, close to deserts, enclosed by mountains, on a peninsula, .....

Let's see what others say, ok? In the end, Thal's the boss anyway

 

[Naeven]

Actually, that's even worse That means humans can move 1 tiles to get the 5vs2 cities, and the AI would just probably settle on start, making the gap even larger.

Would the AI understand that settling on a resource is a bonus?
Would the AI understand that if it's starting on a hill, it should probably move?

of course it's not up to me, that's why I'm very clear on the "I advice". Doesn't mean I'm gonna stop argue against it tho.

 

[Tomice]

Do you move away from your starting position? I always fear I'll move an yet unseen (strategic) ressource out of my city radius. Starting spots are heavily buffed compared to other terrain, I'm not sure if there are many situations where we can profit by doing it. Also, we lose the most valauble turn of all for snowballing