[SPOILERS - NO PLAYERS] The Final Clusterfuck: The PB38 Map and Lurker Thread

[RefSteel]

Ah - thanks, Krill. That should be easy to test at least....

Also, in the absence of further comment from Bacchus, here is my rough Paint interpretation of what he posted:



Black dots represent possible civ starting positions, solid blue is sea, light green is a mix of terrains for civs' cores and primary contested zones, off-white is desert, brown is mountainous, and dark green is a jungle belt. Note this is adapted somewhat from what Bacchus described: I added an inland sea and a large gulf just so that nobody is completely cut off from water before conquest, but this still represents a weakness of the map (bearing in mind that every map concept will have weaknesses).

The "spraypaint" regions are just to show that the dominant terrain is broken up by something else: Freshwater lakes in the jungle area, hills/"mountain passes" in the "mountain range" area, especially near the northern and southern ends of the continent, and islands in the seas (inland and otherwise) plus the borderlands between the different map regions. To improve the natural look, a lot of rivers flowing down from the central mountain range through a narrowish hilly and then flat zone to the inner sea would represent a continental divide, with numerous rivers likewise running from the central mountains down through jungle-choked grassland hills throughout the jungle region to the gulf.

(I also recommend mostly a smaller number of larger islands rather than the ridiculous number of tiny islands the spraypaint implies, but a few little islands mixed in are also a good idea.)

Finally, there are a lot of balance issues with this concept as portrayed, though some may be improved via land area and/or resource distribution. This is just my quick(ish) and dirty attempt to visualize (and slightly tweak) the "Olympic Rings" suggestion as I understand it.

[Cornflakes]

Made progress! It seems there is some issue with placing strategics. Using the following script:

Code:

new_group 136 82 => $PB38
import_mask_from_ascii "working/template.txt" => @baseMap
new_mask_from_shape *random 136 82 => @rand_field

new_weight_table >= 0.97 => %luxury,
                >= 0.94 => %plantation,
                 #>= 0.91 => %strategic,
                 >= 0.86 => %landfood,
                 >= 0.01 => %any_land,
                 => %test1

new_weight_table >= 0.5 => %landfood
                >= 0.01 => %any_land
                => %asdf

mask_intersect @rand_field @baseMap => @random_base
generate_layer_from_mask @random_base %test1 => $PB38.bg


flatten_group $PB38

Worked when the %strategic line was commented out.
Screenshot:

   

WB saves attached in the zip. The one titles "working" works and the one titles "broken" does not work. The only difference between the two when I generated them was the one line mentioned above being commented out. But this is some good progress. Might have a little time today to play with customizing and overlaying regions on the map (e.g. desert band, jungle band)

[Cornflakes]

I've figured out the layering in GJ's program, and managed to split the template into regions, extract the regions in the program, and randomize the tiles in a regions based on a region-specific mix of terrain. Screenshots below. I've defined 7 regions

A #tundra
B #grassland
C #jungle Lite (to smooth transition to jungle)
D #jungle Heavy
E #desert Lite (to smooth transition to desert)
F #desert Heavy
G #plains

On the test map (screenshots below) the regions were assigned basically from top to the middle of the map, and then mirrored to the bottom.

To do:
1) finalize big picture geography ... i.e. stick with current "N" shape? Olympic rings? Something else?
2) determine regions within the big-picture
3) determine terrain distribution within each region (e.g. TUNDRA REGION: 50% tundra, 25% grassland, 25% plains). Also need to determine forest & hill distribution % in each region.
4) Once the above have been determined, I can make the necessary adjustments to generate the base map (including improving the coastline)
5) Then pass off to someone to polish ... add rivers, starts, strategic resources (which seem bugged in GJ's program), and any other balance considerations

Screenshots of current progress:

   

   

   

[Mardoc]

This is looking really nice!

To do:
1) finalize big picture geography ... i.e. stick with current "N" shape? Olympic rings? Something else?

The N seems sufficient.  Although - any chance you could add in the interstitial islands too?

2) determine regions within the big-picture

I would suggest you shrink the 'pure' regions and expand the 'transition' regions.  Would probably make the pure regions slightly less pure, as well - we don't want the land to be useless, just to delay settlement somewhat and reduce interaction until the players have whittled down their numbers a bit.

One thing you're missing is Peaks.  Don't need many, maybe just 5-10%.  Actually, I'd probably go for 5% everywhere except High Desert and High Jungle, where I'd go for 10%.

Also - although the tundra looks really nice, it's really hard to balance; I would either reduce it to a couple tiles, or eliminate it entirely (to be added to offshore islands or the like).  Fits better in a SP game where people can just ignore the land until they've settled everything else.  I guess you could keep it if you changed the tundra region to just a nominal amount of tundra, 10% or less, just enough to be noticable.

3) determine terrain distribution within each region (e.g. TUNDRA REGION: 50% tundra, 25% grassland, 25% plains). Also need to determine forest & hill distribution % in each region.

Jungle heavy and Desert heavy should both have maybe 10% more plains, in addition to being shrunk somewhat.  This is an MP consideration, really - they look great from a SP perspective but we don't want to be too stingy for the players.

On forests and hills - be generous, especially with the forests.  I would sprinkle at least 25% forest into all biomes, except High Desert, and possibly aim more for 40-50% in the non-Jungle areas.

I would probably aim for 25% hills, uniformly in all biomes.  Maybe 30-35% in jungle/desert.  Players like hills too, although unlike forest you can have too many hills.

Also, can you sprinkle in some Oasis tiles into the desert and desert transition areas - maybe 5% in transition and 10% in High Desert?  That will cover all the terrain features except floodplains, which can't really go in until rivers are placed anyway.

Finally, from a 'natural' perspective - can you add some nominal amount of desert to the grassland and plains biomes?  Maybe 2% in grass and 3% in plains?  Just enough to add a little interest, not enough to shift the value of the terrain really.

4) Once the above have been determined, I can make the necessary adjustments to generate the base map (including improving the coastline)

So, is it just strategics that aren't getting placed?  I don't see any resources at all in your latest screenshots, but maybe that's just that the screenies are from an earlier pass?

Also, RB maps tend to be lusher, which tends to mean additional resources.  Can I get you to shift your table by adding a percent or two to each band?

I can't remember - does GJ's tool have any provision for ocean food?  I know he also uses ocean for blank tiles, so it might not - but it would be handy if it would add the clams/crabs/fish

5) Then pass off to someone to polish ... add rivers, starts, strategic resources (which seem bugged in GJ's program), and any other balance considerations

Still a lot of work, but it's starting to get in the range where I might be able to find the time.  Or we could play hot potato for a while, people doing what they can then posting the partial map for someone else to work on.

[RefSteel]

I agree with Mardoc on pretty much everything.

To make the map look a bit more natural, I'd love to combine the jungle and desert bands latitudinally but make the transition go from mostly-flat Jungle along the outer coast, to hilly Jungle, to hilly Desert with more peaks than elsewhere, to mostly-flat Desert - but that might end up being more effort than it's worth. On the other hand, just widening the transition zones might be enough to reduce the striped look and make it feel more natural overall.

With 25 non-mirrored starts required, I was thinking I might try to build a few independent of the map, so they could be copy-pasted in wherever with GJ's tool if desired. (I don't know if this would actually make things easier or not....)

If that would be helpful, I might try diversifying from Krill's plans some: Some starting tiles would have incense as suggested, others might have wine, and others spices. Also, while most starts would have double farm resources or double pasture resources, I would want to give one or more a pair of second-ring forest deer, just to increase the variety. I'd also be sorely tempted to give some starts a first-ring lake forest (or lake stone or lake marble) just to give Fishing a little bit of value for some of the players. (Short of a seafood start, which is bad in itself, or very low tech costs, I don't think there's anything that makes Fishing an actually-good starting tech....) Just spitballing ideas though, mainly. (But I'm definitely willing to help with the map either by building starts, playing hot potato with Mardoc, or whatever else would be helpful in my limited time.)

[Cornflakes]

Thanks for the comments! This will be enough to get me to the next iteration (probably next weekend). 

I rolled a couple of random huge pangeas to spot-count hills & forests. I found fairly consistently about 40-50% forest coverage and 30% hill coverage (outside of heavy desert/jungle areas) which is right in line with your guidline The terrain and resources will be placed in multiple passes within my Python script, and then translate into the actual map terrain within GJ's program.

To make the map look a bit more natural, I'd love to combine the jungle and desert bands latitudinally but make the transition go from mostly-flat Jungle along the outer coast, to hilly Jungle, to hilly Desert with more peaks than elsewhere, to mostly-flat Desert - but that might end up being more effort than it's worth. On the other hand, just widening the transition zones might be enough to reduce the striped look and make it feel more natural overall.

I think I can accomplish this by widening the transition zones, and turning the horizontal bands into elliptical bands. I've already got the code for elliptical bands written since I was using that in the Civ 6 maps so it's not a big deal at all to translate that into the Python. It would look something like the following which I've sketched into the minimap with Paint. Also once the forests get distributed throughout the rest of the terrain that will reduce the striped look somewhat as well. This has the added benefit of providing more useable "plains" region to the middle zone, while providing more of a buffer around the inside corners of the "N".

   

[Krill]

Do not under any circumstances give double deer as starting food. It needs AH, it's a straight penalty.

[Cornflakes]

To address the islands comments ...

Yes, the island regions should be fairly easy to add. I can do 3 larger island like so? [right and left edge crossing x-wrap?]

   

Alternatively I could omit the one on the x-wrap, but that favors the civs along the inner-coast. Also are these astro-only islands? It will be hard if not impossible to grant equal access to coast-reachable islands. Are we going for concentrated blobs or snaky continents? 

Regarding size, I'll need probably 3900-4200 tiles on the main continent based how the coastline-roughening ends up. Roughly 5 % of that would be in the desert belt, so if we exclude that and the peaks from the "land tiles" count we get roughly 3600-3900 land tiles, leaving roughly 600-900 for the islands ... in the grand scheme that doesn't sound like a lot

With 25 non-mirrored starts required, I was thinking I might try to build a few independent of the map, so they could be copy-pasted in wherever with GJ's tool if desired. (I don't know if this would actually make things easier or not....)

Regarding starting location, yes getting started in advance would help speed things up down the line. It would also allow us to provide starting screenshots to the players at an earlier date. The only difficulty I see with that is the rivers from starting locations blending into the overall map. I don't know how GJ's tool handles rivers in the overlay. Might be easier to give the players screenshots without rivers and the caveat "imagine a river flowing past your capital" so that we have the freedom to draw the rivers in the finished map only.

[Cornflakes]

Another alternative for the island regions ...

   

This gives 2 smaller islands for each group of 5 players, 10 islands total. Each island about 60-90 tiles. I like the look of this better than 3 larger blobs.

[Mardoc]

With 25 non-mirrored starts required, I was thinking I might try to build a few independent of the map, so they could be copy-pasted in wherever with GJ's tool if desired.  (I don't know if this would actually make things easier or not....)

Honestly, most of the headache in making starts is in the judgment, not the work to either copy/paste or manually recreate them.  Especially if Cornflakes is wrapping his head around GJ's tool - I don't use it often enough to remember the commands.

So it would definitely be helpful to start working on these, and to post them so we can all weigh in.

Plus, if we get those settled we could even post them to the threads to let the players start civ-picking.

A couple general thoughts:
With a map like this, you want to make settling in place the obvious choice.  Giving them a plains hill helps there, but the only reliable way I have to do that is to spend the time to imagine playing out the start.

Yes, the island regions should be fairly easy to add. I can do 3 larger island like so? [right and left edge crossing x-wrap?]

I would definitely put islands in each area you marked, but rather than one massive island go for 3-4 smaller ones (although still big enough to be worth contesting).  Keep in mind the scale - a 300 tile island is about 20-25 cities of land, bigger than a mainland empire in itself.

Personally I would go for coastal access but opinions differ.  Fortunately that's an easy thing to retrofit by adding a few tiny island chains.  In your initial setup, I'd go for Astro-only while we discuss, and we can change it later if we want. I don't think it would be bad to use a few more tiles on islands, especially in the area furthest from the mainland (arctic/Antarctic far from land); the further a tile is from the start the less it matters to the game.

I'm having trouble coming up with an opinion on the elliptical approach, I think mostly because I don't know which bands your colors correspond to.

Rivers are a pain to draw, but not that ridiculous that we should leave them out; they affect yields and tile improvement plans too much.  For flexibility, just put the caveat 'this (or a rotation or mirror of this) is your start'.  That gives enough flexibility to draw them however the larger continent needs them drawn.