deltarivers.pcx - anyone know when it's used other than for deltas?

[Quintillus]

Sounds like a silly question, but I've realized it is used in cases other than the obvious case. To quote a fairly informative guide, which I'd thought was accurate in this regard until recently, when I noticed what appeared to be deltarivers.pcx in use away from deltas when looking closely.

Next, you get the rivers and resources.pcx applied on. The rivers are applied on the edges of tiles, and if there are mountains involved, have waterfall graphics. If it passes next to a desert tile, it applies the floodplain.pcx graphic over the desert tile, and if it exits to water, it applies the deltaRivers.pcx file.

I was looking into deltaRivers.pcx tonight, and in my experiments, found that some of this description doesn't appear to be fully accurate. I'm wondering if any of the resident experts can fill in some details.

What I did was add a scenario search folder to a randomly-generated map in the Firaxis editor, and add only the deltaRivers.pcx from Balthasar's Winter Terrain set to the appropriate art folder, leaving mtnRivers.pcx as the default. This made it easy to tell when it was using deltaRivers.pcx, and when it wasn't.

After looking around the random map, I saw deltaRivers.pcx being used in most areas, so I created a non-random section, which looks like this:

It appears that deltaRivers is always used next to desert (thus becoming flood plains; the flood plains graphics are also visible), but in other areas there's a pattern to when it's used, and when it isn't. But the pattern occasionally changes.

I spent enough time trying to figure out the rhythm for forests and mountains back in the day to not try to solve it for these. Still, my searches didn't turn up much definitive, a bit surprising since I would have expected some posts over the years about trying to combine deltaRivers from one pack for the deltas with mtnRivers from another pack, and noticing some differences like this.

Anyone else ever notice this? I thought I knew which parts of terrain generation were random, but it would seem otherwise...

 

[Ozymandias]

I thought I knew which parts of terrain generation were random, but it would seem otherwise...

... And where might this info be found?

 

[Quintillus]

... And where might this info be found?

Only in my head, so let's see if I can convert it into a post!

Forest/jungles/marsh and mountains/hills/volcanoes are the ones that I recall. I believe Flintlock has said the "randomness" is actually based on the game seed.

I'm less sure whether coastline graphics are random. I've noticed the surrounding tile affects coastline graphics, but am not sure if that's the only factor.

I believe the ocean/sea graphics (wSSS, wOOO) are also chosen randomly. Which reminds me of one of my ideas for an editor improvement... allowing that to be chosen by the map creator.

 

[tjs282]

I'm less sure whether coastline graphics are random. I've noticed the surrounding tile affects coastline graphics, but am not sure if that's the only factor.

IIRC, whether the tile co-ordinates are odd or even makes a difference to which terrain-tile option is chosen.

Especially noticeable with e.g. islands which consist of only 1 land-tile (or a strip of land-tiles 1 tile in width).

 

[Quintillus]

IIRC, whether the tile co-ordinates are odd or even makes a difference to which terrain-tile option is chosen.

Especially noticeable with e.g. islands which consist of only 1 land-tile (or a strip of land-tiles 1 tile in width).

That would make sense, but when I tested it tonight (in the Firaxis editor, starting with the blank map), this is what I got:

The upper-left grassland is 7, 11; the other rows are rows 14 and 18. There are those random larger blocks of land, for no apparent reason.

I can imagine that depending on the landmass the modder is trying to depict, these differences could be annoying.

 

[Ozymandias]

Only in my head, so let's see if I can convert it into a post!

 

[AnthonyBoscia]

The first one here is standard, new blank water map. The second was a generated map, 70% water, pangaea. The second one has the same pattern as your example above. This pattern also determines the fat or skinny plains and grassland tiles by the coast as well. It looks like the mtnriver tiles make skinny land, and deltariver tiles make fat land. I don't know how many patterns there are beyond these two.

 

[Blue Monkey]

Having worked on maps covering the expanse from Indian Ocean to South China Sea, I can tell you that the coastal tile issue is more complex than simply odd/even. Translate the coastal tiles to the next over set of odd or even coordinates and you don't always get the same visual results. I could never figure it out well enough to predict. Not so bad with one tile islands because it's just aesthetics. With two tile islands it gets more complex. Sometimes you end up with something more akin to conjoined single tile islands. Three-tiles and it enters the realm of undesired straits amongst other complications.

The delta effects under discussion made for some interesting results when modeling the mouth of the Ganges. Unfortunately experimental biqs are part of what I lost recently, so I can't post images.

@Quintillus - the image you posted is of horizontal lines of islands. Try putting single tile islands on a diagonal with coast (water) in between.

During mapmaking I've seen tiles change skinny/wide depending on how many water tiles are between them and other nearby land - or when editing that nearby landmass in any way that affects which coastal tiles are used for it, even if the change is not directly adjacent to the island. Placing a lake one land tile away from the coast, for example. Intermittent and unpredictable - in my experience. I've also seen the skinny/fat change simply by changing the single tile island terrain, like between grassland & plains. Again, intermittent and somewhat depending on map location, so very difficult to figure out.

 

[timerover51]

Hmmm, I may have to poke at this a bit.

 

[timerover51]

I ran a check in all three editiors that came on the game disc: Vanilla, Play the Eorld, and Conquests. All three gave the same results when placing individual grassland tiles in the ocean. You get a string of 3 to 5 skinny grassland tiles, then a standard grassland tile, then another string of 3 to 5 skinny tiles. I suspect that it has something to do with the algorhythm that generates terrain.

I have not duplicated Quintillus' delta results, but I did discover that in about 50% of the cases of a river running into the coast tile, I got a delta. The other cases were simply the river running into the coast tile, no delta. All in all an interesting study.