Map and province improvements for 1.4

[Swarbs]

Do those cluster rules not then act to limit city placement flexibility, particularly in situations when there are preplaced cities? It would seem that if you have lots of preplaced cities in the 1200AD scenario, and limited amounts of overlap allowed, this will mean that there are fewer sites available to found cities and thus greater predictability.

That's why I think a higher separation rather than limits on overlap would give the AI more flexibility to consider more city sites when an area is already quite full with preplaced cities. Civs with smaller territory could still have a two tile overlap.

Personally, I would favour an approach where some important cities are always founded, but the rest are as random as possible. Caen, for example, is very important for the viability of the 1st English UHV, so should always be there (I argued in the past for it being autofounded by the AI in 600AD and added to the 1st Frankish UHV otherwise the human player has no incentive to go there at all), but then just let the rest of France be as random as possible.

 

[gilgames]

I agree with swarbs. Plus teach AI to skip awful places like brest with one tile on my ss or pula for venice even memel between konigsberg and windau, etc. Randomness alway good adds to versatility. 14 tile overlap seems too much.

 

[The Southerner]

Also the primary issue is not even the poor placement, rather that the AI settle the same cities in most games, so there is too little variance.

Or not enough variance! It's a tough one. In my ideal world the 500AD scenario would have the most randomness (as there is more space on the map). The 1250AD map is mostly already settled. I would hope to see particular spots settled by the AI in keeping with historical accuracy and good city placement - but it doesn't usually happen! Especially in Eastern Europe - the AI settles some terrible locations.

 

[gilgames]

well, if there would be an alternative indy city spawn?! For example in france: Bordeaux, Toulouse, Lyon, Marseilles, are now. We could add a few more like Metz, Montpelier, Rennes, etc. idk and rng would choose 4 of them. It would make more interesting the first french uhv, and gibe more flavor to the games!

 

[AbsintheRed]

Do those cluster rules not then act to limit city placement flexibility, particularly in situations when there are preplaced cities? It would seem that if you have lots of preplaced cities in the 1200AD scenario, and limited amounts of overlap allowed, this will mean that there are fewer sites available to found cities and thus greater predictability.

That's why I think a higher separation rather than limits on overlap would give the AI more flexibility to consider more city sites when an area is already quite full with preplaced cities. Civs with smaller territory could still have a two tile overlap.

Personally, I would favour an approach where some important cities are always founded, but the rest are as random as possible. Caen, for example, is very important for the viability of the 1st English UHV, so should always be there (I argued in the past for it being autofounded by the AI in 600AD and added to the 1st Frankish UHV otherwise the human player has no incentive to go there at all), but then just let the rest of France be as random as possible.

In vanilla BtS there are also some similar rules. IIRC it corresponds to the AI not settling cities if it overlaps in more than 7 tiles with other already established cities.
This is a rule which is applied in all situations, which makes it very rigid of course.
In RFCE the values are civ-specific (also with some other modified rules), and allow closer cities for most civs, which makes it more flexible in most situations

If I get it right, you are suggesting that we should totally get rid of this tile-overlap approach in RFCE, and only use actual distance in all cases?
So you (and more importantly, the AI) would be able to settle cities even with 20 tile overlap, if the city spot itself is at least 3 tiles away from all other cities?

...O O C O O
O O O O O O O
O O O O O O O
C O O X O O C
O O O O O O O
O O O O O O O
...O O C O O

Does it really help if you are able to settle X in this situation?
The AI would probably choose the same C citysite places in most games anyway, which would force X to more or less the same position anyway.
I still think that the main issue is not with distance/overlap, but with the lack of variance in placement.

Or not enough variance! It's a tough one. In my ideal world the 500AD scenario would have the most randomness (as there is more space on the map). The 1250AD map is mostly already settled. I would hope to see particular spots settled by the AI in keeping with historical accuracy and good city placement - but it doesn't usually happen! Especially in Eastern Europe - the AI settles some terrible locations.

Yeah, my sentiments exactly

 

[AbsintheRed]

@ Swarbs:
Having said that, I agree with some of your points though. At least they are probably more valid in the 1200 AD scenario.
I mainly have the 500 AD in mind, but it's true that with many preplaced cities overlap can prevent further settling of many good city-sites.
On the other hand the 1200 AD scenario will never be good enough in the regard though, whatever ruleset we use.
We need to have many preplaced cities there, so AI city placement flexibility won't be significant anyway.

well, if there would be an alternative indy city spawn?! For example in france: Bordeaux, Toulouse, Lyon, Marseilles, are now. We could add a few more like Metz, Montpelier, Rennes, etc. idk and rng would choose 4 of them. It would make more interesting the first french uhv, and gibe more flavor to the games!

That's not a bad idea actually
Will think about it

 

[gilgames]

That's not a bad idea actually
Will think about it

Thinking further: it needs to be done in clusters in: Iberia, France, Germany+Poland, Hungary+Balkan, Italy, England, East Europe and North Africa. Some are obvious but need to tide up a bunch of province for that.

 

[DC123456789]

I do feel like many of the current starting cities are important enough that they should always exist, especially cities such as Bordeaux and Lyon that represent certain polities at the very beginning of the game.

 

[AbsintheRed]

I do feel like many of the current starting cities are important enough that they should always exist, especially cities such as Bordeaux and Lyon that represent certain polities at the very beginning of the game.

Yeah, obviously the most important cities should spawn in all situations.
But maybe it won't be bad if I added some alternatives for a couple preplaces cities.
For example in Hungary there is 50% chance that Gyulafehérvár spawns in 900AD, 30% that Kolozsvár, 20% that a 3rd city.
Didn't really look into it in details yet, but it definitely doable, and would probably make some situations much more unpredictable, which is great for gameplay
Do you guys have any cities in mind which would have realistic alternatives?

 

[DC123456789]

Possibly a choice between Caen and Rouen and Normany, or Calais and Ghent (though I'm unsure if this would screw up the French UHV).

Also, this could be useful for representing Brittany; there could be an equal chance of spawning Nantes (Naoned), Vannes (Gwened), or Rennes (Roazhon). Perhaps a city spawn in Wales as well? It's always struck me as extremely odd that the English can just settle there; I would recommend Caernarfon for Gwynedd, one of the most pre-eminent Welsh kingdoms, or possibly Dinefwr to represent Dyfed/Deheubarth as well (or even both).

 

[Swarbs]

I would definitely have Caen in a fixed place, directly south of the crab. It is too important to the 1st English UHV to have it in a less optimal position, particularly inland. It was also the centre of Christina development in Normandy.

Personally, I would have only Caen and Lyon as fixed cities in France. Other cities can spawn semi randomly at Vannes, Nantes, Rennes, Bordeaux, Toulouse, Aix-en-Provence, Marseille, Calais, Strasbourg, and Dunkerque (maximum one city per province or two in larger ones), to give four or five cities in France. That would have the advantage of making the 1st French UHV much less predictable and easy to achieve, and also give more randomness in France in general.

I agree with the Welsh spawn, but think there should be a balance between Caernarfon, Cardiff and Mathrafal (then capital of Powys, becomes Shrewsbury when captured by the English), on the hill S of the sheep. That way you make the UK a bit more dynamic, both for England and Scotland.

 

[AbsintheRed]

Great! I would like to have as many of these as possible
Made a new thread about it: http://forums.civfanatics.com/showthread.php?t=566726

 

[AbsintheRed]

@ Swarbs:
Having said that, I agree with some of your points though. At least they are probably more valid in the 1200 AD scenario.
I mainly have the 500 AD in mind, but it's true that with many preplaced cities overlap can prevent further settling of many good city-sites.
On the other hand the 1200 AD scenario will never be good enough in the regard though, whatever ruleset we use.
We need to have many preplaced cities there, so AI city placement flexibility won't be significant anyway.

I think I will combine the 2 rules, we will see how well it turns out:
The AI can always found cities if they are at least 3 tiles away from every other city
If there is a city within 2 tiles, the civ-specific city cluster rules are applied.
Also the basic rule that cities can never be on adjacent tiles is still valid.

This way small civs (Venice, Prussia, Austria, etc) still have the opportunity to found cities close to each other, while the minimum distance is 3 for bigger civs.
Some of the civs with lowest shared tile value (Russia for example) still got a big buff with this though.
Let's see what does it do for balance.

 

[gilgames]

I think that not the number of cluster or distance, but allow AI to build new city if the new would have 6 own land tile in bfc. 6 which is not common with other cities.

 

[AbsintheRed]

I think that not the number of cluster or distance, but allow AI to build new city if the new would have 6 own land tile in bfc. 6 which is not common with other cities.

As I said, we already have a civ-specific setting for that.
Adding another rule on top of that which allows additional options for the AI is not a bad thing. Even if it won't choose the new ones in most cases.

The only real question is whether we want to allow cities with less than 3 distance from each other or not.
If yes, than I always allow city placement in 3 distance, the number of overlapping tiles will only be counted if there is a city within 2 distance.
If no, city placement is never allowed with less than 3 distance, but the overlapping tile ruleset is always applied.

In both cases, smaller civs (civs with higher overlapping tile setting) will be able to have more cities in the same amount of territory.
The difference is in overall density - it will be much lower in the second case, and not only for small civs, but everyone.

 

[gilgames]

I wanted to point out, that to close and too few land tile cripple the civ. Having too much close city cripple the production for both city. 6-10 not overlapping tile might be sufficient for a city. That's all.

 

[Panopticon]

Sorry I don't see a better place for this. Any support for moving the English city from Exeter to Bristol in the 1200 scenario? Makes the naval game a bit more challenging.

 

[AtlantaMarty]

I am very excited for 2.0, and a larger map.

Until then, I think the game would be improved by expanding the map 5-10 tiles to the east, as this would make more room for the Arabs, and make Trebizond viable. Expanding to the east is easy, compared to expanding to the west, which would require all coordinates to shift.

What do you think?

 

[merijn_v1]

If we are going to expand it, we have to go all the way. Adding 5 columns is just as much work as adding 50 columns, as all maps various other things need to be changed to handle the new tiles. So we won't add 5 more tiles.

Besides, this is RFCEurope, not RFCArabia. And there is enough room for Arabia already. (North African coast)

 

[Swarbs]

If we are going to expand it, we have to go all the way. Adding 5 columns is just as much work as adding 50 columns, as all maps various other things need to be changed to handle the new tiles. So we won't add 5 more tiles.

Besides, this is RFCEurope, not RFCArabia. And there is enough room for Arabia already. (North African coast)

I see your point about Arabia, but personally I would like to see the map expand to include all of Eastern Europe, up to the Urals and the west edge of the Caspian. That would also have the advantage of including Baghdad, Mosul, Trebizond and other cities which were not part of Europe, but did play an important role in the region as a whole.

That would help make the AI Arabs more threatening and viable than they are at the moment, and would avoid the need for excessive revolt mechanics in Jerusalem, as the Arabs could respond to Crusades with their own armies built on a much stronger economic base. It would also make the Seljuks and Mongols more threatening, as rather than spawn a few units at a time in Anatolia and European Russia they could spawn en masse in Iran and the Ural passes and sweep west

Also in the long term we could add Georgia which was an important influence on Eastern Europe at the time and possibly the Golden Horde and Seljuks as playable civs.

Of course I understand this is going to be a lot of work (and I won't be able to do it!) so happy to wait until you guys feel you are in a position to make such a major addition.