Cradle 2.0????

[Maquiladora]

I got Cradle 2 working with the Apolyton Edition in a scenario folder, I just had to disable diplomod (withdraw part is buggy), and fix line 334 in CRASL_mapwonders.slc, in the part concerning pillaging of map wonders:

for(i=0; i <= WONDERTIMP.#; i = i + 1){

remove the second "=" so the loop doesn't go beyond the end of the wondertimp array. This generated an array out of bounds error so should be fixed.

Originally posted by hexagonian

The pillaging, however, stumps me

I think IW used the "CutImprovements" event rather than "PillageUnit" to allow tile imps under a wonder to be pillaged, but not the wonder itself. This of course still allows the AI to constantly try (and fail) to pillage a wonder tile imp, even when there are no other tile imps on the tile.

From looking at CRARR4_tileimp.txt it seems that the only tile imp you can build on the same tile as a wonder is a road/rr/maglev? But the wonder tile imps already have road/rr movement bonus anyway, so maybe you can disable the pillage order (for the AI) on a tile with a wonder altogether?

One thing I am considering as an addon option in Modswapper is to make all the infantry-type units Government-specific. You could lose a lot of units every time you switch Governments... and have to rebuild every time.

That would spice things up a bit. How does the AI deal with it?

===

Also I noticed that the upgrading at any time with the sleep command isn't implemented. The new sleep button implies it but it seems to be an old version of the unit updater, and without upgrading in forts and air bases either. If it's by design (it does make the decision to upgrade certain units more difficult) then it's fine, but the sleep button should be changed back. I waited on upgrading all my spearmen at once to save some gold and not cripple my support costs, and so wanted to upgrade them slowly, then obviously I couldn't.

[HuangShang]

Originally posted by Martin Gühmann


Well, when I can proof read it.

By the way, did you try the download I uploaded to Apolyton?

-Martin

Yes, but my disc is in school and I am not going back until this weekend. Does this look like it will run well on Apolyton Edition?

[Maquiladora]

Originally posted by HuangShang

Does this look like it will run well on Apolyton Edition?

I watched the 8 AI's on a huge map for 200 turns and it went fine. However this wasn't a real test as no human player was involved.

The only problem I noticed so far though is the AI tile improvement building. In the AE for some reason the AI often builds mines on grassland/plains etc rather than a lot of farms, and this impacts their growth quite heavily, so that might be a problem in the long term in Cradle. You can easily fix this by not allowing mines on grassland etc, that's if you don't want to build mines on grassland etc yourself (I don't).

Perhaps I'll test properly after I finish my original game.

[hexagonian]

In the game, where does the AI get it's priorities for terrain improvements???

ImprovementLists.txt doesn't seem to alter the priority too much, just gives what the AI needs to build if it is in a growth, production mindset

I had noticed in the past that in Cradle, the AI focused on Farms and Trading Posts almost exclusively. When playing, I had usually created a lot of Mines on everything, and slaved heavily to get my pop growth.

I decided to alter that for Cradle 2.0.

In Cradle 2.0, mines take away 5 food on all terrain except Mountains (Mountains have 0 to start with). I'm finding that I am building a better mix of TIs now, but this negative food benefit might affect the AI a lot more if it is now prioritizing mine construction through the Apolyton Pack settings.

That would spice things up a bit. How does the AI deal with it?

No idea yet...

I suspect that there will be a little more AI expansionism through conquest, because there will be times that certain AIs will be vunerable. On a small map, this would greatly favor the player, but on a large map, that advantage would be minimized, because of the distances involved...

I believe the player will handle it better (he can do a mass disband one turn before he switches to get a lot of his production back), but with the tendency for the AI to send units into the player's land, it will make military expansion and defense for the player even harder...which is not a bad thing in my mind.

The player will have to invest more heavily into ranged/flanker units as the backbone of his armies. Right now, what I have noticed is that the player generally does a better job of balancing his stack composition vs the AI, and this would level the playing field.

This will make taking cities even more costly.

[Maquiladora]

Originally posted by hexagonian
In the game, where does the AI get it's priorities for terrain improvements???

ImprovementLists.txt doesn't seem to alter the priority too much, just gives what the AI needs to build if it is in a growth, production mindset

In strategies.txt there are these:

// bonus for building production tile improvements
ImproveProductionBonus 120

// bonus for building growth tile improvements
ImproveGrowthBonus 130

and some other options amongst those, to improve different cities first. They seem quite effective, but I guess it'll need some trial and error. I just use the default values and work with the AI's habits by changing the tile imps themselves, like you do. But it's good for making different strategies for different AI's.

I had noticed in the past that in Cradle, the AI focused on Farms and Trading Posts almost exclusively. When playing, I had usually created a lot of Mines on everything, and slaved heavily to get my pop growth.

That is a good idea. Actually I guess I'm doing that now, but I don't find any of the early tile imps (except roads and watch towers) worth the cost.

Mines and mine shafts take too long to pay back, and you're already getting growth from slaves so you don't need much food, especially with granaries.

Commerce improvements seem even more expensive, and you can get lots of natural commerce early from forests and sea anyway. Later perhaps you need to build them for science, because scientist specialists are pretty weak now.

Even better if you get a food good or a river to settle on/near you really don't need improvements for the first 250 turns at least.

If I found myself in some really peaceful game or easily defensible area of the map, where I can keep my PW high, I'll build more tile imps, but even then maybe I'd rather have some wonders or buildings instead and turn my PW off.

Perhaps the weak tile imps are actually helping the human more. Since the AI is always generating PW to build them, but the human can skip them and still maintain good growth in other ways.

The player will have to invest more heavily into ranged/flanker units as the backbone of his armies. Right now, what I have noticed is that the player generally does a better job of balancing his stack composition vs the AI, and this would level the playing field.

I've noticed that as well. A lot of AI armies I've seen so far have used mostly flanking units along the front line, which is not a good idea when I have 5 defensive/offensive and 2 flankers. Perhaps it's just units they found in ruins, and formed the army on the way. There seems to be quite a lot of Chariots and Javelin Cavalry popped from ruins early in the game for example.

[hexagonian]

Originally posted by Maquiladora
In strategies.txt there are these...

Cradle uses it own strategies.txt, so there should not be a conflict between it and Apolyton Pack.

Originally posted by Maquiladora
Perhaps the weak tile imps are actually helping the human more. Since the AI is always generating PW to build them, but the human can skip them and still maintain good growth in other ways.

Depends on the situation, since slaving goes to the city closest to the capturing force, there are cities that will not get a lot of them. Plus some cities need to be augmented by farms because of the terrain (Hills and Mountains)

But yes, I also make roads a priority in the early game.

[Maquiladora]

Originally posted by hexagonian

Cradle uses it own strategies.txt, so there should not be a conflict between it and Apolyton Pack.

When I mentioned "Apolyton Edition (AE)" I was talking about the source code project version of CtP2, not Apolyton Pack the mod. I got Cradle working with the AE, which improves the AI at source code level. Things like water exploration and island settling are improved quite a lot, plus many other little things. It's also a lot better for testing/watching the AI, which is why I wanted to get it working with the AE.

Depends on the situation, since slaving goes to the city closest to the capturing force, there are cities that will not get a lot of them. Plus some cities need to be augmented by farms because of the terrain (Hills and Mountains)

You can remove garrison from the cities closest to the slaving unit so the slave goes to the city you want it to. Although it requires a bit of micromanagement.

In the case of my slave city "Huiztlan" above I built it up big because it had lots of forests around it for commerce, and my other cities weren't too heavy on commerce. Now at size 24 with a ziggurat it generates 1/3rd of all my science, not bad out of 11 cities.

[Martin Gühmann]

Originally posted by hexagonian
ImprovementLists.txt doesn't seem to alter the priority too much, just gives what the AI needs to build if it is in a growth, production mindset

To be precise ImprovementLists.txt doesn't do anything at all.

Originally posted by Maquiladora
When I mentioned "Apolyton Edition (AE)" I was talking about the source code project version of CtP2, not Apolyton Pack the mod. I got Cradle working with the AE, which improves the AI at source code level. Things like water exploration and island settling are improved quite a lot, plus many other little things. It's also a lot better for testing/watching the AI, which is why I wanted to get it working with the AE.

By the way, how many slic event errors do you get. Actually, I added this stuff for slic debugging.

-Martin

[Maquiladora]

Originally posted by Martin Gühmann

By the way, how many slic event errors do you get. Actually, I added this stuff for slic debugging.

I had it turned off so none I forgot I still had it turned off from a long time ago when there were lots of popups, but I'll start using it again.

[hexagonian]

The option that makes all infantry units gov-specific is actually very cool to play. It's not as hard as I expected, but I also did not play it on a high difficulty level either, and my map was easily defendable to boot.

I did not build as many infantry as normal, but depended on other units like mounted to cover me when I was switching governments. (I tend to neglect this branch after getting Jav Cavalry in earlier Cradle games). Plus I stacked all my Wonderunits as my main force to tackle the enemy as the situation warranted. In the past, I normally split them up.

What does make it easy is disbanding units right before switching governments so you can recover some of the production. From what i can tell, the AI does not do this, so this actually works in the player's favor.

And I had an incoming stack disintegrate right before my eyes. Too bad the invading civ was so far away.

I'm giving it another go on a higher level, as well as using Maq's culture setup as part of the mix. (I did switch out some of the building with some Ancient/Medieval buildings, as well as used the Religion Feat techs.)

This next game will be a test to see how the Culture points pan out for the AIs... then I will adjust accordingly.

[Maquiladora]

Originally posted by hexagonian

And I had an incoming stack disintegrate right before my eyes. Too bad the invading civ was so far away.

Ouch. What do you get for disbanding a unit in a city, 50% of it's production value or something? That could be a huge bonus for the human if it's that much.

Is there anything you can do to help the AI in this area, perhaps simulate that disbanding somehow?

I'm giving it another go on a higher level, as well as using Maq's culture setup as part of the mix. (I did switch out some of the building with some Ancient/Medieval buildings, as well as used the Religion Feat techs.)

So you had no problem finding everything you had to change? Some of it is not as clear as it could be, especially the buildings per turn part. Fingers crossed it all works as advertised anyway.

[hexagonian]

The funny thing was that the stack promptly turned around and fled from my borders.

Originally posted by Maquiladora
Ouch. What do you get for disbanding a unit in a city, 50% of it's production value or something? That could be a huge bonus for the human if it's that much.

Is there anything you can do to help the AI in this area, perhaps simulate that disbanding somehow?

Actually I could not figure out what the disband bonus is. The manual says 'a small percentage'...

I cannot figure out how to get the AI to think along those lines of selectively disbanding units. It would have to anticipate going into a government switch, then pull it's units into cities to disband before actually making the switch.

The easy way is to pump up the production or PW cheats for the AI - the better way is to give it production based on unit cost whenever a unit is disbanded via a SLIC code, but the production would probably have to go into PW. (How can you designate production to go to a particular city?)

Originally posted by Maquiladora
So you had no problem finding everything you had to change? Some of it is not as clear as it could be, especially the buildings per turn part. Fingers crossed it all works as advertised anyway.

No problem that I can see so far...the code is firing based on the score screen.

but I already anticipate having to adjust which buildings I use, because the AI might not be savvy enough to get the number of buildings it needs to get the SLIC to fire.

I'm not exactly sure how to change the numbers, because it appears to be some sort of formula, rather than a straight number.

From what I've seen in the game, the AI will generally get the Religions bonuses and some of the Feats, and the player will outstrip the AI in the Wonders department. (Usually I have double the wonders as the next AI).

[Maquiladora]

Originally posted by hexagonian

The easy way is to pump up the production or PW cheats for the AI - the better way is to give it production based on unit cost whenever a unit is disbanded via a SLIC code, but the production would probably have to go into PW. (How can you designate production to go to a particular city?)

I can't think of an easy way to add production spent to a city.

You could create and immediately disband a unit in a city that has room, but it could turn into a headache to get working, and for not much effect. Perhaps Martin or someone else has a better idea?

I guess at least tell the AI to build a few more of the other types of units, like you do.

No problem that I can see so far...the code is firing based on the score screen.

I'm not exactly sure how to change the numbers, because it appears to be some sort of formula, rather than a straight number.

Yeah a lot of the formula is visible in the main code, rather than hidden in easy to change functions. I'll try to clean that up for a future version if need be.

Meanwhile if you're having trouble changing or adding anything in particular let me know and I can go into detail how to change it. Or you can tell me how you want it set up and I can do it.

[Aias]

Any chance you're going to be reposting on a download site so I can download Cradle 2.0?

Stay away from here for a little bit and see what you miss?

[Martin Gühmann]

Originally posted by Aias
Any chance you're going to be reposting on a download site so I can download Cradle 2.0?

Actually, I uploaded a version of the archive to Apolyton and removed some unnecessary files from it so that its file size is just 40 MB. Anyway I removed the invalid links from hexagonian's posts and added a link right at the bottom of the opening post.

-Martin