Buffed AI for BNW

[Poruchiik]

No probs Poruchiik, here you go:

What map conditions were you playing in your king game?

It was just a random Earth map as Korea. 8 players random map, Spain went faith-buying and took out the Mayans and the Iroquois pretty much instantly. Portugal and China were to the north and south of me respectively (I was in the Middle East). Portugal went full on landgrab mode and kept on settling new cities everywhere. China pretty much set back and cultured up while Portugal built a massive army, and started taking Iroquis cities from the east and Spain took them out on the west side.

Overall I was just sitting back minding my own business until Spain came and bulldozed down my cities and I couldn't do much to stop it.

Does this mod work with customized civs? Or is it only for the stock standard BNW civs?

 

[dhialuck]

Hello Glider and all, this mod looks amazing, great job! I want to try it out, but I wanted to ask if the AI free starting techs are still there in higher difficulties? Also which difficulty should I choose? (I usually play emperor but pull ahead a lot in the late game)

 

[glider1]

Hi Dhialuck
This mod works differently to any other in terms of difficulty. No matter what difficulty, there is no difference in techs, the AI get's adjusted automatically during the game. You could even play on prince and get a competitive game for a while. If you are an emperor player, try another emperor game, *but make sure you build a proper army and military infrastructure* because the AI can be punishing. It depends how you set up the game. If you play on a large islands map, the AI cannot be punishing, because it cannot get to you as soon! Likewise, pangaea maps would be very challenging in this mod. The difference in the difficulty is what happens in the late game. At higher difficulties, the AI should have more chance at exceeding the player and winning the game. I'm still working on making the late game more competitive. Joncnunn knows about this better than I do.

Poruchiik, the beauty of this mod is that no matter what customisations you make, so long as it doesn't need a DLL, the AI will adjust better because even if it doesn't understand the customisation all that well, the AI will be adjusted to the players performance. It is not perfect yet, because the AI doesn't get adjusted for yields that the player gets from non-city sources, and so certain policies still give the human an advantage. I'm thinking about looking into this if Joncnunn or someone can give me some ideas.

The main weakness this mod has at the moment (V15-17) is it doesn't work in multiplayer, and the player can cripple their neighbour by declaring a perma-war on them, which in duel games means this mod is too easy, but should be fine in eight or more player games. I'm currently working on getting the AI to defend itself better with the two warrior start, but it is slow going because the AI is intellectually impaired.

EDIT:
For anyone who wants a really really challenging game, download this mod and delete the HANDICAP.SQL file in the mod and nothing else. That will mean that the handicaps will be identical to the base game, except that the AI will be smart mod AND the AI will be getting an adaptive buff where necessary. That will close all loop holes currently in this mod and make it an exceptional challenge (I think, but have not tried it).

Think of it like adding double hot chilli to your meal.

 

[joncnunn]

Hi Dhialuck
This mod works differently to any other in terms of difficulty. No matter what difficulty, there is no difference in techs, the AI get's adjusted automatically during the game. You could even play on prince and get a competitive game for a while. If you are an emperor player, try another emperor game, *but make sure you build a proper army and military infrastructure* because the AI can be punishing. It depends how you set up the game. If you play on a large islands map, the AI cannot be punishing, because it cannot get to you as soon! Likewise, pangaea maps would be very challenging in this mod. The difference in the difficulty is what happens in the late game. At higher difficulties, the AI should have more chance at exceeding the player and winning the game. I'm still working on making the late game more competitive. Joncnunn knows about this better than I do.

Poruchiik, the beauty of this mod is that no matter what customisations you make, so long as it doesn't need a DLL, the AI will adjust better because even if it doesn't understand the customisation all that well, the AI will be adjusted to the players performance. It is not perfect yet, because the AI doesn't get adjusted for yields that the player gets from non-city sources, and so certain policies still give the human an advantage. I'm thinking about looking into this if Joncnunn or someone can give me some ideas.

The main weakness this mod has at the moment (V15-17) is it doesn't work in multiplayer, and the player can cripple their neighbour by declaring a perma-war on them, which in duel games means this mod is too easy, but should be fine in eight or more player games. I'm currently working on getting the AI to defend itself better with the two warrior start, but it is slow going because the AI is intellectually impaired.

EDIT:
For anyone who wants a really really challenging game, download this mod and delete the HANDICAP.SQL file in the mod and nothing else. That will mean that the handicaps will be identical to the base game, except that the AI will be smart mod AND the AI will be getting an adaptive buff where necessary. That will close all loop holes currently in this mod and make it an exceptional challenge (I think, but have not tried it).

Think of it like adding double hot chilli to your meal.

Note that Barbs are also punishing, so even if you have an isolated start you need quite a few units to clear local barb camps with the mod. (As an example if your starting warrior hits an upgrade hut and becomes a Spearman it still loses more hit points to AI brute when both started fully healthy.)
It's also important on water maps with the mod to keep your navy up to date. The ones with high navy flavors will build a fleet.

Immortal with the handicap.sql removed would be similar to standard Deity.
AIPerEraModifier & AIDeclareWarProb would be back to normal but more than offset by AI starting with base games Immortal level techs, AI growth bonus, AI large happiness bonus, etc. (At least the dynamic portion is likely to turn itself off for mature AI cities, but their brand new cities would get it)

Deity without handicap.sql would be similar to if base game had created a new difficulty level higher than Deity. (Sid level). Same reasons as above, only the base games bonuses are even bigger at this level.

Going the other direction, Emperor without handicap.sql wouldn't be up to standard Immortal yet, in large part because removing handicap.sql would also cancel the mods starting Worker the AI would get.

Down to King (and lower) the base games bonuses are so small that the mod might actually be easier without handicap.sql (again due to AI not getting worker)

I'm currently trying Polynesia on V17 on Tiny Islands as an example civ who gets a cultural bonus instead of a science one.

 

[glider1]

Thats a really good test setup Joncnunn. Polynesia would allow you to protect yourself and test the endgame without non-city science bonuses.

Big milestone for this mod today. The AI has new unit intelligence and will guard its workers, so the worker steal option is gone, and the AI protects its settlers better as well. City states however still have the worker steal option because they don't get any start units to protect them with. Should I look into that? I don' really want to.

 

[joncnunn]

Thats a really good test setup Joncnunn. Polynesia would allow you to protect yourself and test the endgame without non-city science bonuses.

Big milestone for this mod today. The AI has new unit intelligence and will guard its workers, so the worker steal option is gone, and the AI protects its settlers better as well. City states however still have the worker steal option because they don't get any start units to protect them with. Should I look into that? I don' really want to.

World wonder completion times are way down on this map (Tiny Islands) compared to normal. This is because hammer production sucks big time on this map. (There's little hammer leak with the human also having bad production)
Also, income is terrible on tiny islands pre Banking. The 2 AIs that are at war are losing a lot of money each turn.

The Celts who were first to religion first follower belief amounted to denial to the human (they took Monasteries and I had several nearby wine sources, they didn't have either wine or incense near them)

Polynesia's UI works well on this environment, and is providing more culture than the French UI, but Brazil's UI is better.

City state military behavior in base game is really dumb compared to the major AIs. (Major AIs keep a unit in their capital, the city states don't; the city states try to move every one of their units assigned defensive roles instead of fortifiing. While I wouldn't change the city states to grant them any additional units, I think the city states AI should be changed to be more like the major AIs.

 

[glider1]

Thanks, what you are writing here on these forums could all come in handy down the track for people trying to improve the game. I'd say that the CS AI is probably historically dumbed down by Firaxis because of performance reasons? What I think I might do is at least teach the CS to guard its workers. Their is also a problem with 1UPT traffic jams and CS, which probably made Firaxis keep the units moving around so that gaps open up. Ironically this decreases game performance!

So, what do you think if I were to include the no-more-traffic jam mod in this mod, which actually helps the AI a lot and means that CS armies could be fortified, because there is no traffic jam problem, since civs can get their peaceful units through to the CS. I personally found the strict adherence to 1UPT by Firaxis over the years bordering on dogmatic.

 

[joncnunn]

So, what do you think if I were to include the no-more-traffic jam mod in this mod, which actually helps the AI a lot and means that CS armies could be fortified, because there is no traffic jam problem, since civs can get their peaceful units through to the CS. I personally found the strict adherence to 1UPT by Firaxis over the years bordering on dogmatic.

No more traffic jam as is civilian workers can pass thru other civs civilian & military units when not at war with them? That's an excellent idea that's long overdue.

Yes, most of the tactical AIs problems on 1 UPT wouldn't have occurred at 2 UPT where you could always escort a siege unit with a melee one. (Which is why SSI went that way.) In addition, at 2 UPT, you'd have seen archery, early siege, and city bombard at 1 hex range instead of 2 and most land units with 1 fewer movement point.

 

[kaspergm]

No more traffic jam as is civilian workers can pass thru other civs civilian & military units when not at war with them? That's an excellent idea that's long overdue.

"No more traffic jam" might just be the greatest single mod out there in terms of improvement it has on gameplay issue. Only objection I have with it is that I can't catch enemy Great Prophets when I play with it, which was otherwise a favorite pasttime of mine.

Spoiler :

 

[joncnunn]

"No more traffic jam" might just be the greatest single mod out there in terms of improvement it has on gameplay issue. Only objection I have with it is that I can't catch enemy Great Prophets when I play with it, which was otherwise a favorite pasttime of mine.

Spoiler :

It's looking like you had more than 6 units, so in this case just parking six of them right next to the city instead of having some 2 hexes out would have blockaded out the unwanted Great Prophet.

Also dealing with unwanted Great Prophets, I found it was extremely easy to mod Great Prophets to give them the same "unwelcome evangelist" penalty that Missionaries have. (If you want to be more extreme you can remove the promotion that allows Great Prophets to enter territory without open borders in the first place)

 

[kaspergm]

Also dealing with unwanted Great Prophets, I found it was extremely easy to mod Great Prophets to give them the same "unwelcome evangelist" penalty that Missionaries have. (If you want to be more extreme you can remove the promotion that allows Great Prophets to enter territory without open borders in the first place)

Obviously, if I just wanted to block the prophets, I could just use an Inquisitor. But indeed, I also made a mod that gives a small attrition of -50 FPT on Prophets, which is why I wanted to trap them.

 

[joncnunn]

Obviously, if I just wanted to block the prophets, I could just use an Inquisitor. But indeed, I also made a mod that gives a small attrition of -50 FPT on Prophets, which is why I wanted to trap them.

The bad news is that the removing of foreign religion is fully functional on a GP no matter how weak it's gotten, (e.g. A GP down to the minimum possible strength without being killed is functionally an inquisitor that still works in foreign lands.)

 

[glider1]

Obviously, if I just wanted to block the prophets, I could just use an Inquisitor. But indeed, I also made a mod that gives a small attrition of -50 FPT on Prophets, which is why I wanted to trap them.

Suggest you play this mod. In this mod AI missionary and prophet spam is reduced and so people don't need radical tactics to keep foreign religion out. In the base game it is ridiculous (but it has to be). That is one of the reasons I built this mod.

 

[thelemon402]

What exactly is the difference between V15 and V17? Been testing out V17.

 

[glider1]

Hi ya thele, thanks for testing it. Difference is just that V17 includes the players trade route yields into the calculation of the adaptive buff that the AI may get. If the player has no trade routes, there will be no difference between V15 and V17. I haven't officially released it because Poruchiik (see top of page 12 this page) saw something odd with the income civs were getting. I can't think of how it is possible but who knows.

EDIT:
In theory it means that the AI should get a bit more of a buff in gold or science if it needs it.

 

[joncnunn]

Suggest you play this mod. In this mod AI missionary and prophet spam is reduced and so people don't need radical tactics to keep foreign religion out. In the base game it is ridiculous (but it has to be). That is one of the reasons I built this mod.

The difference in behavior is that in base game, usually the AIs try to convert your capital.
In this mod they send their Great Prophets to a different target. (Either one with some followers already or else one that has some passive influence over; even way would explain behavior I see in this mod and they often amount to the same thing.)
In addition they are much smarter than in base game about targets that are blocked by an Inquisitor. (In base game, they tend to stay around 4 or 5 hexes out pacing back and forth waiting for you to get the Inquisitor out of their way, in this one they pick a different target.)

So, if your neighbor(s) founded a religion you'll still need Inquisitors, but not as many (and in addition its different cities you want their home base to be)

 

[glider1]

Thanks Joncnunn. It could be just a fluke (like so many things about this AI) because I haven't changed the AI for religious units. It is possible that it is part of the smart AI mod which has (I haven't checked). It is possible that because the AI is no longer carpeting the map with religious unit spam, that is engaging other parts of the already existing code. Lets hope that the changes are consistent good. I sure hope they are!

EDIT:
Just an update on progress. I'm in the middle of measuring the turn time performance with V18, which greatly improves how the AI guards its workers and settlers. It should also fix that silly bug where AI workers walk around the map like drunks.

 

[joncnunn]

My Immortal Polynesia game on Tiny Islands with V 17 is now complete. Cultural Victory in 1902 AD.

Highlights are the everybody's production, food, and science rates were lower compared to normal. (I was in the Modern era researching Plastics, and all the AIs were in the Industrial era)
Gold was also a big problem for the first half of the game (until Banking)
These were mostly Tiny Islands induced (there's a lot less land tiles on this) but also in the case of the human is compounded by having Maoris.
My Cultural victory was primarily Maori + Hotel + Level 3 Freedom Tenet + Olympics driven but I also had completed Aesthetics and some national wonders in my capital and used faith based GEs for Leaning Tower (and chose Uffizini for the GE produced in turn), Broadway, and Eiffel Tower. In another I hand built the Louvre and was also hand building Statue of Liberty when the game ended.

Some AI Great People sanity checks are needed, one of the AIs wasted a Great Musician on a tour of me when it was only at the minimum 100 points, and quite a few also wasted low powered Great Musicians on me as well. The AI would be better off using on Opera Houses if it's already been born if the strength is less than 1000.

 

[glider1]

Thanks for pushing the extreme ends of this mod Joncnunn. If the player is pushing out high culture/tourism on relatively low core yields, that would bypass the adaptive buff. That combined with the AI's poor strategic play. Also playing on tiny islands probably doesn't engage the combat AI as much, which means bypassing the combat buffs the AI gets in this mod? (Is that true? I haven't played on these types of maps)

EDIT:
Could I ask a question? Should I turn off the adaptive buff for city states?

I think it was a mistake not to because it actually hurts the conquest civs and hurts their ability to expand. The reason is that currently city states get more yield to build a lot more troops because of the leak they get, making conquest of them that much harder. It also means a lot more unit spam and difficulty to get civilians into the city states because of the traffic jam.

So I think I should boost only the major civs, so that the conquest civs that capture a CS (which will be easier to do) will likely also get an adaptive buff, which would make their presence more of a challenge for players to deal with.

 

[Poruchiik]

I haven't been able to recreate the same effects that got me a ton of GPT in v17 within the early game.

I played a map that I modified based on the Extended Europe TSL with it starting with Greece and ending on the Western End of India. There were only 12 players on the map but the spawn locations were really close together. The customized civs were Ambrox's Phoenicians, Medes, Hittites and Tomatekh's Harappa and Sumer. From the base game Babylon, Assyria, Egypt, Greece, The Huns, Persia and Byzantine(only because I wanted the city of Constantinople in the game, don't know how to create a CS that is just Constantinople).

I've yet to incorporate alot of CS in it because there isn't very many in the base game that are Middle Eastern flavored. Kabul, Ormus, Jerusalem, Yerevan, Samarkand and Kiev are the only ones I could put in the game without some areas ie. the Eastern Mediterranean being too crowded. The map is still a bit in-progress.

Overall, I think it was the spawn locations of Phoenicia, Assyria, Babylon, Sumer, Persia being really close together as it was easier for me to get trade routes to the Phoenician capital as well as Babylon as soon as I hit Optics for another trade route with Assyria. It also could be that I'm not use to seeing the AI tech and build structures quicker than normal.