Buffed AI for BNW

Yes agree chumchu, will make that the next AI improvement on the list if one of the others on Civfanatics is not already doing that.

Dekraa, in this mod siege get cover-I which they don't get in the base game. It helps the AI and that is why I added it. There is nothing the AI has to know about that promotion, there is no particular skill required to use it. The AI has trouble positioning it's siege and I want the AI to be able to capture cities better. It also helps humans a bit as well, since we often need siege to capture cities, but our siege are weaker relative to the AI.
 
glider1 said:
Yes agree chumchu, will make that the next AI improvement on the list if one of the others on Civfanatics is not already doing that.

Dekraa, in this mod siege get cover-I which they don't get in the base game. It helps the AI and that is why I added it. There is nothing the AI has to know about that promotion, there is no particular skill required to use it. The AI has trouble positioning it's siege and I want the AI to be able to capture cities better. It also helps humans a bit as well, since we often need siege to capture cities, but our siege are weaker relative to the AI.
Click to expand...
I definitely think that the cover promotion for all siege units makes for a better game. I could never imagine going back to playing without it. I think it's balanced for both the player and the AI.
 
Just out of interest, does anyone think I should change this mod like this?

  1. Return game back to default handicaps except:
  2. Leave the human combat nerfs and AI comat bonuses
  3. Leave the AI's science per era buffs it get's from the renaissance onwards
  4. Leave in the fix for the AI being too willing to engage in a third party war
  5. Leave in the cover-promotion for siege

It is just an idea because although it gives back to the AI all the happiness, free tech and other bonuses, it would have these advantages:
  • It would definitely cause players to go down at least one level
  • It would definitely make the late game victory race more competitive
  • It would simplify the mod, so that balance is essentially the same as the core game across all difficulties
  • It would mean the AI has more starting units, and therefore would protect it's workers and settlers a bit better

Just thinking out loud. I don't like the idea on balance, but thought I would put it out there.

EDIT: When I think about it, if I did this change, it would mean that for example, immortal players might even drop back down to king and emperor players might drop to king or even prince. On king, players would have more variety, be able to build wonders and the AI would spam less cities anyway making for a better game experience?
 
glider1 said:
Just out of interest, does anyone think I should change this mod like this?

  1. Return game back to default handicaps except:
  2. Leave the human combat nerfs and AI comat bonuses
  3. Leave the AI's science per era buffs it get's from the renaissance onwards
  4. Leave in the fix for the AI being too willing to engage in a third party war

It is just an idea because although it gives back to the AI all the happiness, free tech and other bonuses, it would have these advantages:
  • It would definitely cause players to go down at least one level
  • It would definitely make the late game victory race more competitive
  • It would simplify the mod, so that balance is essentially the same as vanilla
  • It would mean the AI has more starting units, and therefore would protect it's workers and settlers a bit better.

Just thinking out loud. I don't like the idea on balance, but thought I would put it out there.

EDIT: When I think about it, if I did this change, it would mean that for example, immortal players might even drop back down to king for example. On king, immortal players would have more variety and the AI spams less cities etc anyway....
Click to expand...
Would reverting back to default handicaps result in the game being harder in the early eras, but potentially easier at the end (compared to how it is now)? Why would it make the late game victory race more competitive? Would it be more about catching-up to the AI? especially in tech? Thanks in advance.
 
Thanks for taking an interest Marshall :)

Going back to default handicaps except combat would basically set up a new equilibrium on a lower difficulty level where humans would be doing a bit more catch up than how the mod is set up now, and once the AI get's it's per-era-science bonuses kicking in, they might never catch up at all.

That's the theory.
And that could already be happening even without changing anything in the mod!

So, I think I've got a simple solution.
I could release two versions of this mod. One could be called:
  1. Buffed AI-NO FREE TECHS (which is how the mod is now)
  2. Buffed AI

Basically, all that has to happen is to delete the handicap.sql file inside the mod and that would leave the AI with combat bonuses, science bonuses and cover promotion, and everything else would default to the core game. It is incredibly simple to do.

EDIT:
I think the two play styles both have their merits with subtle differences before the renaissance and possibly big differences at the end of the game, not sure. What I really don't like about the game's default handicaps, is that the AI's happiness is a totally manufactured. The AI doesn't really need to trade luxuries with you at all...
 
glider1 said:
What I really don't like about the game's default handicaps, is that the AI's happiness is a totally manufactured. The AI doesn't really need to trade luxuries with you at all...
Click to expand...
That's also my biggest problem with default. I'd imagine it also makes certain ideological tenants, like the Order one which gives you a cultural bonus against civs less happy than your own, almost useless. not to mention negating the usefulness of certain World Congress proposals, such as banning luxuries.

Of these two possibilities, which one do you believe allows for the most strategic flexibility? The main complain against default was that, to be challenged, you had to play on immortal (at least); but doing so also meant that you had to use every trick in the book to keep up with the AI in technology. This imposed great strategic limitations upon the player.

By the way, I've been thinking of some possible additional features for this mod. They're mostly minor and don't have anything to do directly with this topic of how AI handicaps should be applied; so I'll talk about them more later. My main idea revolves around further nerfing the advantage of range three artillery. and also a few ideas to discourage most of the world from choosing Order almost every game.

I haven't played without this mod in so long that I almost don't remember default. If not for this mod, I probably would have shelved CiV a while ago. So my ability to judge the difference of AI handicaps effects in this mod vs. default probably isn't as good as some of the players who have recently switched to this mod.

Thanks again for taking the time and effort to further improve this mod!
 
Well, the thing is that an immortal player would not be able to play on immortal any more because the AI's combat buffs and human nerfs would push them down in difficulty, which then would engage more strategic flexibility....

The big question is, which version of the mod would produce the most strategic flexibility?

Next game I play (after I finish the current BuffedAI v12 game) will be to delete handicap.sql in this mod and play another game under very similar conditions. There is no need for me to release a new mod for now.
 
Good point Dekraa. I'll release two versions, and people can take their pick according to their play style. Without handicap.sql, people will probably have to drop one level and still have a tough time in the late game. With handicap.sql, I think you can stay on the same level but have more fun.

Another thing I really like about this mod is that there is a rough balance between units humans loose verses units the AI looses, when the militaries are roughly equal and stalemated. This means in a war of attrition, you can roughly say that if you kill an AI unit and loose an AI unit in one turn, that the military strength has not swung either way. If you loose a unit but the AI does not, in this mod, you know you have suffered a real and meaningful loss.

EDIT:
Another thing I really like in this mod is that humans really have to use their ranged units wisely. They are so weak and vulnerable, yet if used wisely, they can stay with the player throughout the whole game and accumulate a lot of experience. When I loose a melee unit in this mod, I really don't care too much. If I loose an experienced ranged unit, I care a lot. Not so much with siege because they have to be expended sometimes when capturing a city, but since ranged units are too weak against cities a lot of the time, they become precious for other subtle reasons. For me, cavalry are also fairly expendable units as I use them for a variety of roles that are normally risky liking raiding, flanking, pillaging, spotting, and charging.

None of the subtle ways of using units matters that much in the base game against the AI because it is so weak. And when you fight a deity army on the core game, it is about whittling down AI numbers in whatever gamey exploitative way you can, rather than thinking about it as simulating actual combat as it happened in history.
 
The problem is that with the two games I've played now with V12 is that I've been isolated on both. And I didn't accept an embassy with any civ as well (which means they have to found my capital by their own scout) so I have not had any wars. In my first I was 6 techs ahead of the second highest nation approaching Industrial era. I quit that game, so I didn't find out if the AI would have caught me with their increased bonus. But the same thing is now, I'm ahead of three techs, and we're in Reinessiance. This is not something I'm used of when playing on emperor, and without wars it's easy to be ahead in science, although the AI is very competent of building wonders, as always :) But I have not played enough to feel when the AI is accelerating with science bonus going into later eras. So maybe I should give it a shot.

Edit: But I think of the AI had also started with their bonus techs, happiness as natural, I would likely be behind at this stage. And with a war or two, I would be a loss cause, maybe :) But on the other hand, I'm not an expert in Civ, I always find things where I can improve. So I like an AI that is pushing me, sometimes I think losing gives you a lot more.
 
If you don't mind playing isolated you should delete the handicap.sql file before starting starting a new game. The way the mod is now, it is setup for players to have at least two other AI's as neighbours.

I think I didn't make that clear enough to players, they need to set up this mod so that the AI can attack them. If it can't attack, there is no point.

So yeah, delete the handicap.sql file and go for it!

What I normally do is always play on large continents maps because you will mostly get a few civs on the same continent. If I am alone on a continent I re-start.

So.....deleting the handicap.sql file would be a way of dealing with all players in all situations. The problem then is that players will start out isolated and think everything is normal, and be totally under-prepared for aggressive AI's if and when they do come.
 
If you have not already seen it then this might be of interest regarding changes in the AI-coding. See also this thread. I will comment further when I have the time to properly test V. 12
 
Dekraa said:
What if keep the all original bonus the AI gets, plus BuffedAIv12, but skip the happiness bonus the AI gets?

.... but after reading this three
http://forums.civfanatics.com/showthread.php?t=407981
maybe the AI have to have its happniness bonus after all..
Click to expand...

That's a very old thread from 2011 in Vanilla when there were a lot fewer sources of happiness.

The current state is there are two times in base game when the AI has problems:

1. AIs with high expansion flavors don't really care what their current happiness is; if they see an empty city spot, they'll take it.

2. The AI does not appear to pay much attention to its happiness or those that are influential over them when selecting an ideology. In addition, it appears to be extremely slow to take corrective action from this.

Meanwhile the ones that don't run into either of these end up in base game looking like they are going after the Suma Tablets for all achievement on high difficulty levels. (This has more to do starting in BNW with the base games AI bonuses from human being on high difficulty level stacking with the AI Default handicap one than just the AI Default handicap)

I think a good solution in a mod for this would be for it to enable AI cash buying of happiness buildings, along with making a priority for it dependent upon current happiness.

Also, the AI goes thru policy trees at a decent rate even without the base games large handicap bonus left over from Vanilla and G&K for AIs seeking to win via Old Cultural victory.

PS: I just now finished the game I started under either V9 or V10. AI fell way behind in the second half of the game, but I see that in the mean time V12 was released that moved up the timing of the AI getting the science bonuses.
 
joncnunn - suggest you delete handicap.sql in V12 (and only that file), start a new game, stay on the same level or drop a level, but ensure that you build your army up and keep it prepared. Build walls and castles too because in this mod they can be very important! Don't try to capture a city without experienced units. It is possible that BuffedAI might be a better mod without the handicap.sql file. Without that file, the AI will still fight harder by a long way and it still get's the per-era-science bonuses that are apart of this mod!

I'll release a version without it as soon as I can.

EDIT: good suggestions about what to do with happiness, let's hope someone or myself does something about it.
 
glider1 said:
joncnunn - suggest you delete handicap.sql in V12 (and only that file), start a new game, stay on the same level or drop a level, but ensure that you build your army up and keep it prepared. Build walls and castles too because in this mod they can be very important! Don't try to capture a city without experienced units. It is possible that BuffedAI might be a better mod without the handicap.sql file. Without that file, the AI will still fight harder by a long way and it still get's the per-era-science bonuses that are apart of this mod!

I'll release a version without it as soon as I can.

EDIT: good suggestions about what to do with happiness, let's hope someone or myself does something about it.
Click to expand...

It's kind of obvious why the AI would fight way harder without that file at a glance:
It's actually mostly this line and less so the happiness (and other stuff) in this file:
UPDATE HandicapInfos SET AIGrowthPercent = 100;
(Population is science, gold, production, etc.)

Also, on the military part you might want to consider having the human attack penalties only apply against the AIs and not against the barbs.
 
Good point on the problem of levelling the growth bonus.

However, it is a lot of fun to play with it the way that it is (up until the industrial), because we can see the AI growing according to what we would expect, as in the natural bounty of the country they find themselves in. I love actually seeing myself fall behind in population because I'm on the plains and the Indians are next to fertile rivers.

So, I rely on the AI(s) attacking me, to cause me to fall behind the AI and have a competitive late game.....

But it breaks the endgame as it stands for people that are not getting attacked by the AI.

That is why I recommend to delete handicap.sql, to have a competitive endgame for everybody....

EDIT:
I've got an interesting idea for the next version that will hopefully work for everybody....
 
I tested V. 12 today. Played fractal, standard, standard, immortal. Rolled Arabia with a nice desert start. Btw, it is really hard to find starts with multiple close enemies if you do not play pangea and end up in the middle. You might avoid balancing it too much around that as it eliminates many maps and starts.

I ended up with one close neighbor, England (<10 tiles) and 4 distant neighbors (<25 tiles.) I stole worker and settler form my closest neighbor and killed its warriors. That seemed to break them and I could camp by the city and farm xp. They built no more military units until turn 100, instead building, workers, walls and caravans. When I finally captured them it felt like putting them out of their misery. I got 3 DoWs and some denounciations after I took England's only city but then I was already at chivalry and had high xp camel archers. Got sweet peace deals giving me cities, gold and luxuries for killing their units. I used that to focus on science and growth. Renaissance turn 130 and secularism soon after with great writers. Industrialism turn 165, but it took 15 turns and a great general to get coal. Reached modern era 190. By then most AIs were in medieval or renaissance despite stealing massively from me. The Incas were the only AI to reach industrial by turn 205 when I got plastics and ended the session and they were almost extinct due to an attack from me and Egypt.

If you want the AI to be competitive midgame you need to give the AI it more science earlier. Humans players can at the very least get national college by 85 and universities 120 even if you are building military. The AI do not seem to be near that.
 
chumchu said:
Humans players can at the very least get national college by 85 and universities 120 even if you are building military. The AI do not seem to be near that.
Click to expand...

The NC time (and therefore the rest) is very start specific in BNW.
How many AIs were you able to make contact with early?
Did you get horses at all?
Is the AI near you broke?
etc.
On starts where I'm been unable to sell due to few AIs near me and those are broke, NC is delayed to turn 95.
I do agree that the NC turn number isn't really impacted by building defensive military units. (Archers being dirt cheap in hammers; and can be upgraded)

The MemMod that's listed on the front page actually does have something that would slow a human down building NC.

Code: 
		<Update>
			<Set NumBuildingNeeded="3"/>
			<Where BuildingType="BUILDING_NATIONAL_COLLEGE"/>
		</Update>
As is actually causes you to need to build 6 Libraries to build NC. (There's a long standing bug that only affects mods where it doubles the number needed when not set to the default value of "-1"; I'm sure that modder intended you to only need 3; set it to 2 if you want them to need 4 as setting it to 1 to need a flat 2 in some cases would speed up NC)
 
Thanks

I'm in the middle of testing V13 that will fix a lot of these issues. It is a completely new but very simple concept - dynamic AI basic needs buffing. There is no need to adjust or balance anything, because the AI cannot fall too far behind in this system because it is automatically being fed the appropriate buffs to fulfil it's basic needs to keep up with the game - but no more. If the AI is exceeding it's basic needs, it's buffs are scaled back and it competes reasonably equally with the player. So it's a feedback system that has built in dampening. The problem is, I don't think it will be multiplayer compatible, it is a single player solution....

So you get the best of both worlds, an AI that often will be progressing at a realistic natural rate, but buffed when it falls behind a basic needs threshold. I think it will turn out that some of the AI's will actually not need the buff, because they will reach a stage where their own development snowballs (but not because of artificial handicap bonuses, but because it has been given a helping hand and the rest it does itself). Remember that humans are competing against multiple AI's and therefore face multiple threats.

It's looking good so far, hoping to have it out in the next couple of days. I can assure you that with this new system, the AI will not ever fall behind enough in science that it can be walked over. The other benefit of this approach will be that you will never have to re-roll ever again, because if you pick a modest difficulty level, the AI will not steam roll you, yet be unable to fall drastically behind you either. Every game will be competitive and because of the AI's combat buffs, exciting and threatening. If you choose a high difficulty level, all that happens is that the basic needs buffing doesn't automatically kick in nearly as much.

EDIT:
Because the system doesn't need manual balancing (it automatically balances and the buffing numbers are logically derived not chosen by trial and error), I should be able to get the mod out soon, it is just a question of testing which is time consuming. If you want to help test, I can release the prototype as a beta let me know.
 
You must log in or register to reply here.

Similar threads

Fergei
[C3C] In defence of culture (the hidden jewel of Civ 3)
Replies
10
Views
678
Fergei
Fergei
Taefin
Taefin's More AI Cheats
Replies
2
Views
763
Taefin
Taefin
T
Better way to implement AI difficulty bonuses
Replies
1
Views
619
Krikkit1
K
B
Is it possible to mod teams into the single player games and individually set each AI player's difficulty?
Replies
1
Views
408
coolhandluke11
coolhandluke11
Fergei
Ideas for game modes in multiplayer
Replies
0
Views
142
Fergei
Fergei
Back
Top Bottom