arrakis terraforming victory?

[Deliverator]

Here's a great passage from page 368 of Dune, that says a lot about terraforming (and I believe confirms the presence of ice at the poles):

'It has been calculated with precision,' Stilgar whispered. 'We know to within a million decalitres how much we need. When we have it, we shall change the face of Arrakis.'

A hushed whisper of response lifted from the troop: 'Bi-lal kaifa.'

'We will trap the dunes beneath grass plantings,' Stilgar said, his voice growing stronger. 'We will tie the water into the soil with trees and undergrowth.'

'Bi-lal kaifa,' intoned the troop.

'Each year the polar ice retreats,' Stilgar said.

'Bi-lal kaifa,' they chanted.

'We will make a homeword of Arrakis - with melting lenses at the poles, with lakes in the temparate zones, and only the deep desert for the maker and his spice.'

'Bi-lal kaifa.'

'And no man ever again shall want for water. It shall be his for dipping from well or pond or lake or canal. It shall run down through the qanats to feed our plants. It shall be there for any man to take. It shall be his for holding out his hand.'

'Bi-lal kaifa.'

I'm wondering whether we could have another terraforming improvement built on polar terrain called Melting Lense - (the FFH2 graphic, Mirror of Heaven, could be used). This would produce a fair amount of water, but destroy the polar ice around it. Once the ice is gone, it can be terraformed as normal. I like the idea of having a diminising polar ice cap as the game goes on. This might only make sense on the Arrakis mapscript.

 

[davidlallen]

I have locally changed terraforming so that it only happens on an improved plot, and if there is an unimproved plot, the terraforming is removed. Any kind of improvement counts, including solar farms. That means you can pillage, and slow down terraforming. Strictly, this does not slow down the terraforming victory, which happens when you have seven buildings. But if you want the benefit from terraforming, you have to build improvements. I have also added an improvement, anchor grass, which gives one water and can be built on any land terrain (including desert and rugged). I have not autoplayed it much yet, but this "should" allow the AI to build if it cannot find any better improvement.

I am thinking of another change, but it involves the AI having "perfect vision". In order to slow down the terraforming victory indirectly, I would like the AI's to get less friendly for each Reservoir of Liet built, even if the AI's cannot really detect the building. This may not be too bad a violation; for example, when World Wonders are built, there is a global message every player can see.

Some of the AI's are against terraforming, either because it is a Hated Civic, or because the AI has independently chosen the Spice Industry civic value. Each such AI would get a -2 diplomacy for each Reservoir you own. This uses python pPlayer.AI_setAttitudeExtra and TXT_KEY_MISC_ATTITUDE_EXTRA_BAD. The message is:

-6: "Your terraforming is threatening the spice!"

In addition, when a civ builds five reservoirs, all these AI's would declare war unless they are friendly (not bloody likely!) or a vassal.

What do you think of these changes? Too much, bad idea, not enough?

 

[Ahriman]

I have no problem with this; IIRC, you could see how many spaceship parts had been built by the enemy AIs. You can also easily detect the locations of the reservoirs because they have changed the tiles all around them.

I'd also make sure that they tech requirements for the second of the terraforming buildings was much higher; the first one (1 tile ring) can be quite low, but the second should be much higher - or there shoudl be a third one that is much higher that actually is needed for the tforming victory.

A couple of issues with the mechanic as it stands though:
a) I've noticed that sometimes building a reservoir turns the surrounding tiles to plains (+1 water), and then immediately the next turn to grasslands (+2 water). Shoudln't it take longer? Or have some tiles stay plains?
b) It doesn't work on many terrain types; desert (land tiles) and badlands and such; how about these still terraform to a new but weaker terrain type? For the visuals alone.
c) Should the number of reservoirs needed to win be scaled to map size, like domination victory % are? So on larger maps you need more cities/reservoirs to win?

 

[davidlallen]

You can also easily detect the locations of the reservoirs because they have changed the tiles all around them.

My only concern is that this kind of implies perfect satellite surveillance, and rules out the idea of Fremen hiding their work by bribing the spacer guild.

I'd also make sure that they tech requirements for the second of the terraforming buildings was much higher

That seems fine.

a) I've noticed that sometimes building a reservoir turns the surrounding tiles to plains (+1 water), and then immediately the next turn to grasslands (+2 water). Shoudln't it take longer? Or have some tiles stay plains?

There is a bug in the numeric calculation in 1.3.6 and previous. Instead of a 2% chance to upgrade after your first reservoir, it is 98%. I will release the fix in my next patch; maybe it will be too slow then.

b) It doesn't work on many terrain types; desert (land tiles) and badlands and such; how about these still terraform to a new but weaker terrain type? For the visuals alone.

It occurs on badlands, but not rugged or desert. Badlands upgrades to arid first. I can apply the same to the other two.

c) Should the number of reservoirs needed to win be scaled to map size, like domination victory % are? So on larger maps you need more cities/reservoirs to win?

I will see if I can find a way to implement this. I don't know how the scalable victory parameters work, it may be in the sdk.

 

[Ahriman]

Changing these percentages will fix a lot of these.

Another thought; there needs to be some way to prevent exploitation of civic switching. So you can't just go pro-tform, build the reservoirs, then switch to pro-spice and still run a good spice economy.

So the civic check probably needs to get tied into the terraform % chance. If its 2% chance per turn, that needs to be 2% chance per turn IFF civic = pro-tform.

For further differentiation, and desert rugged could go to badlands, then to arid, then plains, then grassland.

 

[davidlallen]

Another thought; there needs to be some way to prevent exploitation of civic switching.

As mentioned in the release note, if you have a reservoir and you are not running paradise civic, the reservoir is destroyed at the end of the turn. It is terrible, what can happen when your civ goes into anarchy

For further differentiation, and desert rugged could go to badlands, then to arid, then plains, then grassland.

I like this, but I need a little more infrastructure to remember what the original terrain type was. If it goes up via terraforming, and then the reservoir is destroyed, it needs to go back down to the right level.

 

[Ahriman]

if you have a reservoir and you are not running paradise civic, the reservoir is destroyed at the end of the turn

Cool, that works.

I like this, but I need a little more infrastructure to remember what the original terrain type was.

Hmm, good point. Well, one way would be for the terraforming to only degrade by one step if pillaged. Or have it never degrade past arid (the best non-terraformed terrain type).
Another would be to actually have different upgrade versions depending on what the core was; you could have multiple grasslands that looked the same and had the same benefits but were actually different terrain types.
I guess the last would be if you could actually create an array somewhere that stored the base terrain type of every tile on the planet and store that somewhere.
But no idea how to do that.

 

[SwordOfJustice]

Those diplomacy changes sound great, David. Very appropriate that other civs react to you building Reservoirs and approaching that victory condition.

Cheers,
Sword

 

[davidlallen]

This thread has been dead for a while, but the relevant history is here and I am sure most people will find it.

The present implementation of fresh water and terraforming is slightly complicated and parts of the complexity are not really useful. This post shows the current design, a proposed new design, and some other good suggestions I don't plan to implement just yet.

Current design

Fresh water is spread by wells, windtraps and catchbasins to all adjacent plots. The sdk does not spread fresh water to coast/ocean; so if one of these objects is adjacent to desert waste, the desert waste does not count as fresh water.

The effects of fresh water on your economy are pretty limited, which is by mistake; it should have more effect. In 1.5 the qanat can be built and gives +1 water. In 1.5.1 (not released yet) I think the qanat will be removed, and cottages on fresh water will have +1 hammer. Anyway, there is some effect.

Spice and sandworms will not enter certain areas. They will not enter fresh water. But since the game does not count fresh water in coast, they must check separately to avoid fresh water generators adjacent to coast. They will not come within two plots of a Reservoir of Liet.

Terraforming will only occur within one plot of a catchbasin or two plots of a Reservoir of Liet. Only "rock" terrain upgrades, so that I can easily downgrade the terraforming if the plot is no longer near a Reservoir of Liet. Grass downgrades to plains to rock, and then stops; I do not need to store what was the original terrain.

Furthermore terraforming will only occur if there is an improvement on the plot. That is why the otherwise useless "anchor grass" improvement exists. If there is a plot near a Reservoir of Liet, and you cannot think of anything else to build usefully there, you can build this, and then the terrain will improve. Part of the motivation for this was to enable pro-spice people, who fear terraforming, to pillage your plots and slow down the terraforming. But most people don't realize this and don't care, since to have any real effect on terraforming you need to raze the city with the Reservoir.

Proposed new design

I am thinking of combining these three effects: fresh water, spice/sandworm "do not enter", and terraforming improvement into a simpler fresh water effect.

Wells, windtraps and catchbasins will spread fresh water within all adjacent plots, even if they are coast or ocean.

Reservoir of Liet will spread fresh water within two plots, even if they are coast or ocean. (There is going to be a trick about removing this fresh water, when the Reservoir is destroyed somehow, but I will have to figure out how.)

If you have the arrakis paradise civic, any rock plot with fresh water will terraform. So plots which are only near windtraps, and not near any city, would also terraform. If needed we can lower the chance to terraform so the total amount of terraformed terrain stays similar to today's.

Take the anchor grass improvement out of the game since the complexity it introduces today is not actually known / used / needed.

Good ideas for the future

Anchor grass should speed up transformation. That is possible, but it is unlikely the AI will understand it.

Worse terrain, such as rugged or badlands, should also upgrade, but it may take more time for them to upgrade compared to rock. This is possible, but then I need to store the original terrain type in order to undo back to the proper level.

 

[Ahriman]

If you have the arrakis paradise civic, any rock plot with fresh water will terraform. So plots which are only near windtraps, and not near any city, would also terraform. If needed we can lower the chance to terraform so the total amount of terraformed terrain stays similar to today's.

Hmm. Interesting ideas.

Potential issues:
i) a large number of plots will already be adjacent to either windtraps or groundwater; usually at least 50%, sometimes much more. This means that there would be very little incentive to build catchbasins and reservoirs, since you can also get most of their benefits without them. Potentially you could tie the global terraform rates into the number of catchbasins and reservoirs you controlled, so there was also at least a passive bonus? And just give the building a high AI rating so that they build it anyway.

ii) fresh water access doesn't have any graphical indication. It would be annoying if terraform chance and terrain yields were significantly affected by a mechanic that was invisible except for mouseovers. Maybe there could be some kind of visual indicator, like the aqueduct building water channels? So these channels could go from the windtrap and groundwater well into adjacent tiles?
iii) why is important that desert tiles gain fresh water? Are you going to have desert waste tiles change somehow, like having oasis features appear in them for a +1 water bonus?
iv) potentially you could also have other features, like low chances of a scrub forest appearing, with a +1 hammer bonus.
v) interactions with the mapscript/heightmap changes. If there are terrain types rock, rugged, saltpan, graben with mesa and sink terrain changes, which ones should be effected, and what should they upgrade to?
One way would be to have graben-sink, rock-sink, rugged-sink, rock and rugged all upgrade into plains and then grassland as current, but saltpan and mesa tiles are unaffected.

You could just allow them all to deplete back to rock after pillage if needed, since the variation is fairly small; it doesn't really matter if you have rugged -> plains -> grassland downgrade to plains, downgrade to rock. Its not that important to store the original terrain type; arguably once a rugged has transformed into a plains, you've also been clearing the land and such, so its basically rock.
You could also have rugged terrain have a lower chance of turning into a plains than the others.

The +1 hammer bonus from sinks should be maintained, otherwise shifting from a 2h1c graben/sink to a 2f grassland could even be a downgrade.

 

[davidlallen]

there would be very little incentive to build catchbasins and reservoirs, since you can also get most of their benefits without them. Potentially you could tie the global terraform rates into the number of catchbasins and reservoirs you controlled, so there was also at least a passive bonus?

The incentive to build RoL is to win, and catchbasin is a prerequisite for that so it takes longer. I guess I did not mention the Terraforming Counter in the previous post, but the game counts the total number of catchbasins and RoL to decide the percent change of terrain upgrading. This is the global terraform rate you mention; I would keep that part of the design.

ii) fresh water access doesn't have any graphical indication. It would be annoying if terraform chance and terrain yields were significantly affected by a mechanic that was invisible except for mouseovers.

Fresh water in vanilla is also invisible unless you mouseover; but players quickly learn it is adjacent to all lakes and rivers. Similarly, with the exception of city buildings, you would quickly learn it is adjacent to all wells and windtraps.

iii) why is important that desert tiles gain fresh water? Are you going to have desert waste tiles change somehow, like having oasis features appear in them for a +1 water bonus?

To unify the sandworm/spice "do not enter" check. Presently there are three separate steps, which is complex. Fresh water? Adjacent to windtrap/well? Within two plots of RoL?

You could just allow them all to deplete back to rock after pillage if needed, since the variation is fairly small

I'm not sure about that. It allows an exploit: if you have a lot of rugged terrain, change to paradise civic until all your rugged changes to plains, then change civics back.

 

[Ahriman]

This is the global terraform rate you mention; I would keep that part of the design.

I would not make this global; I would make it per player. You need to be capturing the benefits yourself from building more catchbasins in your city, not be spreading them out to anyone who adopts the civic. Otherwise there is little incentive to build them.

Catchbasins and Reservoirs need to have their own benefits separate from the victory condition.
Otherwise there is no point in building a catchbasin (with its water penalty) until the late game when you pursue the victory condition.

Fresh water in vanilla is also invisible unless you mouseover; but players quickly learn it is adjacent to all lakes and rivers. Similarly, with the exception of city buildings, you would quickly learn it is adjacent to all wells and windtraps.

But in vanilla fresh water doesn't effect tile yields, with the exception of farms in the very late game, and even then farms are the resource you are using to spread fresh water.
And in your proposal you also get fresh water from catchbasin/reservoir, and you can't identify the location of those (and which tiles have fresh water) without going into the city.
Unless you gave them a very large obvious building graphic?

Besides, lakes and rivers obviously provide fresh water, and even graphically look like fresh water. Windtraps and catchbasins aren't nearly so obvious.

To unify the sandworm/spice "do not enter" check.

Ok.

It allows an exploit: if you have a lot of rugged terrain, change to paradise civic until all your rugged changes to plains, then change civics back.

Not much of an exploit; in order to terraform you must build an improvement on the tile, and then wait a long time.
And the only reason why rugged is worse than rock is because of the longer improvement construction time. And you already built the improvement.

So the only possible gain here would be from building a cheap improvement, waiting a few dozen turns, changing civics, and then changing back and building a more expensive improvement on the tile. Hardly worth it to save a few worker turns.

If you're worried about exploits, then there is a much larger one with Harkonnen slaves; you can have them work on an expensive improvement until 1 turn before they finish, then move them away so they don't finish it, and finish the thing with a normal worker.
There are always going to be minor exploits.

And isn't there a much larger exploit? If you don't actually need catchbasins to maintain terraforming, then you switch to paradise and build some catchbasins, get some terraforming done, and then change back (losing the catchbasins), but unless your tile is pillaged you maintain the terraformed grasslands.

 

[Ahriman]

Another thought:
If you're going back to having Paradise civic terraform plots anywhere (ie not just in BFCs), then why don't you return to the 3% (or x%) rule for terraforming victory; you must have 3% of the world's land territory be grasslands and inside your cultural borders.

Rather than needing Y reservoirs.

So you build reservoirs and catchbasins to speed up the terraforming within your own borders.

And make sure that they are documented in game as doing such.

 

[davidlallen]

I would not make this global; I would make it per player. You need to be capturing the benefits yourself from building more catchbasins in your city, not be spreading them out to anyone who adopts the civic.

In FFH, hell terrain is a global counter which is shared among players. In DW, the global ecology does not care from within whose border the moisture is coming from.

And isn't there a much larger exploit? If you don't actually need catchbasins to maintain terraforming, then you switch to paradise and build some catchbasins, get some terraforming done, and then change back (losing the catchbasins), but unless your tile is pillaged you maintain the terraformed grasslands.

Another omission from the above post, but it is captured elsewhere in the thread; you lose terraformed terrain *either* by having no improvement, or by changing away from the paradise civic.

If you're going back to having Paradise civic terraform plots anywhere (ie not just in BFCs), then why don't you return to the 3% (or x%) rule for terraforming victory; you must have 3% of the world's land territory be grasslands and inside your cultural borders.

The earlier reason for not doing this was that the AI will not understand about building windtraps and other improvements outside its BFCs.

 

[Ahriman]

In FFH, hell terrain is a global counter which is shared among players. In DW, the global ecology does not care from within whose border the moisture is coming from.

In FFH, factions benefit/suffer differently from the rising Armageddon counter, and no-one except Sheaim really has any incentives to deliberately devote resources to raising it (through the Elegy of the Sheaim ritual). Sheaim have incentives to build Elegy rituals, because they benefit from a rising AC while their rivals suffer.

But if building more catchbasins helps many of your rivals as much as it helps you, there isn't much point in doing it; better to just free-ride on the other civs.

As for realism; terraforming Arrakis isn't putting moisture out into the atmosphere; its using the water you have stored to irrigate plant growth and try to start generating a self-sustaining ecology on a local level.

You make these plantations and palmeries and grass growth in your own territory.
So absolutely from a realism perspective it matters who is storing the water, because they control where the water gets used.

The terrain transformation represents the success of ecological projects on the local level, producing plants in the area.

So, from both gameplay and realism, it makes more sense that the catchbasins/reservoirs benefit the person who built them, rather than just anyone.

 

[davidlallen]

Reanimating this thread. In the 1.5 feedback thread, we have discussed (again) about changing the terraforming victory condition. This thread is the more appropriate one to discuss that, since some of the previous design discussions are in this thread.

In 1.5.4, terraforming spreads to any fresh water plot; catchbasins and wells spread fresh water within one plot; reservoirs of liet spread fresh water within 2 plots; and victory comes to any civ with 7 reservoirs of liet. Several players have commented that this victory is too easy. You can choose the Private Property civic, shut down your economy, and hurry the reservoirs without much trouble. Several players have suggested that the victory condition should be terraforming 3% of the planet's land area inside your own cultural borders.

As discussed earlier in this thread and also in the older spacer guild mechanic thread, this condition only makes sense if there is a way for other players to fight it. As currently defined, terraforming is a terrain type and there is no way for other players to pillage it. We could change the pillage action to remove terraforming, but the AI would not intentionally use this to prevent a terraforming victory. We could add an auto-pillage type action, so that any enemy walking on a terraformed plot would automatically un-terraform it; but again the AI would not use this intentionally.

So, the question for discussion is: suppose we changed the terraforming victory to happen as soon as 3% of the total land area is terraformed by one player. Would this be an improvement over the current situation, or only a different wrong answer? What is the right way to implement a terraforming victory?

 

[Ahriman]

condition should be terraforming 3% of the planet's land area inside your own cultural borders.

To clarify: 3% might not be enough, I would say X% for now. X will probably need to be > 3% (though keep in mind that only rock, rugged and graben/salt can be terraformed; polar and mesa can not, and rugged takes a long time).

this condition only makes sense if there is a way for other players to fight it.

This doesn't make sense to me.
a) This condition is no harder to fight than against than the 7 reservoirs condition.
Effectively, the condition isn't going to be able to be met without a bunch of reservoirs, otherwise not enough tiles will have fresh water.
Invading a terraformer to take away some of their terraformed land will work, conquering cities with reservoirs will work, or sending in spies to destroy the reservoir catchbasin building (which should remove fresh water from many tiles and thus cancel their terraforming progress). In fact its easier to fight.
In the 7-reservoir method, blowing up a reservoir can just have it replaced with no real loss.
In the terraforming method, when the reservoir blows up, lots of terraformed tiles will no longer have fresh water and so will degrade, which will take a long time to terraform back up to grassland again.

b) The condition also isn't really any harder to fight against than a space-race victory, its just that terraforming can't easily be hurried by the player.

The AI won't intentionally stop a reservoir or space-race victory either, other than through normal war. AI won't deliberately work to stop *any* victory conditions; your friends won't fail to vote for you to stop you winning a diplomatic victory, members of your faith won't suddenly switch faiths and start inquisitioning their own cities to stop you winning a religious victory, and so forth.

So I don't see the problem.

It would also be easy to create an event trigger when you get close, that has all non-Arrakis paradise users declare war on you (ignoring vassals, allies, etc.).

The way I would put it is this; with the 7-reservoir method, you build the reservoirs and then win instantly. With the X-tiles terraformed method, you need to build a bunch of reservoirs *and* then wait for quite some time while the terraforming occurs. And there is no real way that this terraforming can be rushed, which eliminates the cheese of the rush-built reservoirs.

I think this would definitely be an improvement. It is more fluff-worthy, it means that victory takes *time* as well as hammers, and it can be counteracted effectively by the human player with spies (is there some way to give the AI a high preference for destroying reservoirs with spies?).
The trick will be in getting terraform speeds right, and getting X% right.

* * *

If we went this way, we would need to scale terraform time by game speed, and we would probably need the 3-range fresh water from reservoirs.

And we would probably need to make it X% of *non-Dunes* tiles. The outlying Dunes tiles on an Arrakis map (that are there just for aesthetics) shouldn't really count towards this. It would be nice if we could exclude them from domination/holy war type victories too.

 

[WarKirby]

Just a few thoughts from me.

A % of terraformed tiles in the world is the best idea, I think. I'd say take some example from FF's gone to hell victory - some method of tracking "responsibility" for terraforming, and awarding the victory to whoever contributed most, when most of the planet (read, >50% at least) is terraformed.

For it to be feasible, terraforming should be able to spread outside of your territory, and across the desert waste.

There need to be some consequences for doing this though. The sand worms aren't really menacing enough, since all they do is run around angrily outside your land accomplishing nothing. Maybe the sandworms should go crazy as more of the planet is terraformed, and start to become braver in their counterattacks, like somehow venturing onto "land" and eating terraformed plots back down to desert

 

[Ahriman]

A % of terraformed tiles in the world is the best idea, I think. I'd say take some example from FF's gone to hell victory - some method of tracking "responsibility" for terraforming, and awarding the victory to whoever contributed most, when most of the planet (read, >50% at least) is terraformed.

This seems infeasible to me. There is no real good way to get terraforming to spread outside your own territory, or to track "who caused it" (how would you even define that?).

And I see no reason for it to not be based off terraforming within your own cultural borders.

And David is very very against desert waste tiles getting any terraforming.

Sandworms going crazy makes no sense. Sandworms avoid water; it is poison to them, so there is no way that worms could start attacking a terraformer, whose lands are rich in water and plant life.
Like species facing habitat loss today on earth, the worms' response to terraforming is.... to die.

 

[AnotherPacifist]

If it's possible to account for all non-desert tiles from the start (i.e. the mod can know exactly at each moment how much terraformable tiles are terraformed), then the goal of controlling x% of terraformable tiles would be reasonable. Granted, you can't tell which civ caused the terraformed land, but all the more reason for civs to fight over good land. This goal should account for a non-domination win though (i.e. the x% should be less than what's needed for domination, preferably half or less to make it possible), otherwise it would just mean a domination victory.