Buffed AI for BNW

[glider1]

Great to hear pondloso, that is the intention of this mod. Instead of increasing the quantity of new things in a mod to keep the interest up, make it so that what is already built into the game, becomes more useable. So things like combat in this mod is more useable because the AI really puts up a fight. Same for religion etc. What I love about this mod is that it is set up for roughly 1:1 losses. In a battle with the AI the player looses a unit, the AI looses a unit and it comes down to who looses the most units over the battle duration.

The other thing that this mod brings out is the importance of unit promotions. The AI doesn't get any promotion bonuses in this mod and so you get to see the real thing. It doesn't matter in the base game because the AI gets huge experience bonuses but still easily dies. In this mod, even without any promotions, the AI still puts up a show because of its combat buffs. I'm trying to get out of the habit of years of playing the broken combat in Civ5, of just ignoring my promotions and the opponents promotions, because the AI was so bad it didn't matter.

EDIT: give a lot of the credit to improvements in the quality of the AI to Ninakoru and Smart mod. Yesterday I saw the effect of the range-3 penalty extending to battleships as well as artillery. Any unit humans attack at range-3 will be penalised (NOT CITIES). I liked the way I had to decide between a pot luck range-3 attack or moving in closer to range-2 to ensure the hit works.

 

[glider1]

If we get desperate and want to improve the late game performance of the AI and can't think of any way to do it, we have one last resort. I've noticed that the adaptive buffs in this mod really start kicking in come the late mid-game to endgame phase. So it is basically a buff for the late game already. So if I just give the adaptive buff more of a boost, say 20% extra on science, that will fix late game performance for sure without affecting the early, mid-game phases as much. Even with the boost, it will still switch off if the AI's yield is better than the player.

 

[glider1]

I'm pretty sure that the opener to Rationalism doesn't leak to the AI. The flat 10% bonus to science is calculated after all of the city yields. So can't see how it would leak to the AI. Problem with that opener is that it is always a good choice because it adds 10% to a players research without feeding to the AI. I think that should change. The rest are probably ok.

 

[joncnunn]

I'm pretty sure that the opener to Rationalism doesn't leak to the AI. The flat 10% bonus to science is calculated after all of the city yields. So can't see how it would leak to the AI. Problem with that opener is that it is always a good choice because it adds 10% to a players research without feeding to the AI. I think that should change. The rest are probably ok.

How easy is it to check to see if the human does have that policy? And if so leak 10% more than otherwise?
If it's not easy to change, you could change the Rationalism opener to a policy that grants 10% more science to cities with a Monument, which would be almost the same thing.

It we start tweaking policies it should be in a separate file, but I'd

1. Downsize the Tradition policy that both makes garrison units maintenance free and gives cities a bombard bonus to only making the garrison units maintenance free, and relocate the city bombard bonus to one of the Honor Policies.

2. Change the Liberty finisher bonus from the free great person to new cities starts with size 3 population (current Resettlement tenet)

3. Rename that tenet to something else, and have it grant a free great person.

 

[glider1]

Great suggestions Joncnunn, keen to try those as well. I think what I will do is release a seperate mod just for game changes eventually. It could be used in tandem with this mod.

TODO: BuffedAI game changes that incorporates Joncnunns national wonder and policy changes

Meanwhile however, here is the latest build of BuffedAIv18 beta. I've made one change that is going to have a noticeable impact on the difficulty of the game. After watching the AI handling settler escorts for a while, I've decided rather than writing heaps of code to catch every last way that the AI stuffs up settler movements, I've done this instead with a few lines of code:

• Barbarians will not try to capture AI settlers (they will try to capture human settlers). They will still try for any workers (but the AI knows how to guard them now). Other major civ AI's will still try to capture settlers, and humans can still try to capture settlers too. Think of it that the barbs are a bit less barbarian because they spare the AI women and children.

What this means in practice, is that instead of 10-20% settlers being lost by AI incompetence, almost 100% are able to settle. What it also means is that the AI WILL successfully colonize terra-nullius continents. This is going to help wide playing AI's. It also means that the adaptive buff will probably kick in sooner too. It also means that the AI is going to have a bigger population sooner, and will benefit from buffing leak more as well. It also means that wide playing aggressive AI's are going to be able to attack sooner and be more successful as well, because they will be able to build armies quicker. So watch out!

I've tested it and I know it works. It is going to be fun to see how much of an impact it is going to have on difficulty. I suspect it is going to be a noticeable effect.

save game compatible with previous V18 beta

 

[glider1]

How easy is it to check to see if the human does have that policy? And if so leak 10% more than otherwise?

Done

I'll just run through a game with the rationalism change before I upload. It could have a big effect on the endgame. If humans choose rationalism, 10% extra science is leaked to all AI's on a per city basis if the AI qualifies

Rationalism is still important despite this serious nerf. The point to remember is that with the adaptive buff, if the player has a bigger population than competing AI's, even with yield leak to the AI, they still will be yielding higher than the competition. This encourages players to engage in increasing their population either peacefully or through conquest.

The adaptive buff really hurts the player when the AI's population is significantly larger than the players, and the AI's yield/population ratio on a per-city basis is low. That is when the adaptive buff really kicks in.

It'll be interesting to try.

 

[joncnunn]

Update on the Netherlands V18 + National Wonder change game on Immortal
I just entered the Industrial Era (Turn 205) via Industrialization (timing worked better to avoid unwanted Great Prophets than Scientific Theory would have)

The AI Indonesian offensive against AI Russia stalled out after Russia was down to only their capital. AI Indonesia is actually somewhat on the defensive now, trying to fend off an AI Russia counter offensive.

I was actually able to successfully build Notre Dame. I've also built Leaning Tower (using it for GE for Globe). AI India built Sistine.
The unknown civ was AI Austria, they have several city states that are allies but so far haven't married any of them.

AI Sweden wanted a DOF with me, which since it gives a Great Person spawn bonus I was happy to take. However, a few turns ago, AI Sweden, apparently out of the blue, backstabed me by denoucing me. Of course I denounced them right back and started shifting my forces for war time.

When I saw their galley fleet, I used 1000 gold to buy Sweden's only CS ally and signed Defensive Pacts with both AIs that I had active DOFs with.
So far, my only loses are that half my cargo ship routes will need rebuilt post war. Sweden though lost a cargo ship in turn and their first wave. (They have a second wave) Sweden obviously doesn't have Astronomy yet so it's at a point in the game where Galleys are very fragile but can be replaced quickly.

Post war I'm going to need to self found another city. Only one city state has Coal, it's only a 3 source, and they are allied to Austria so I can't rule out Austria marrying them. (Sweden doesn't have coal either)

AIs in science, most of the AIs are within 3 techs of me, one actually has the same number, but Indonesia and Russia are both really far behind now due to their war.

 

[glider1]

Sounds like a whirlwind of a good game going there Joncnunn. How is the tech rate of the AI's?

EDIT:
If you wanted to try the Rationalism nerf, the new beta V18 I posted above is save game compatible. You could save your game just before selecting rationalism, just in case there is some problem in my code (I doubt it).

 

[glider1]

I found a small improvement in the code today that was causing AI workers to flee from danger unnecessarily. I've fixed that and when I officially release V18, the AI will be that little bit more productive in the early game yet again. It might even have a large effect because in the base game, workers would always flee from barbarians. Now, if they are being guarded, instead of leaving a pillaged tile, they will repair it, meaning that it could be the difference between an AI being unhappy for many turns, or only unhappy for one.

 

[glider1]

Yeah I think that might be it. I play on King difficulty but it seemed really weird that I was getting 2k+ gold per turn. Some AI's were getting 256-500 GPT, some of the luxury resources weren't improved and they had two trade routes going towards me.

Well done Poruchiik I finally got exactly the same bug as you. It is because I cannot include trade route yield in the adaptive buff. You cannot miss the bug because all of a sudden you get ridiculous gold income. It can affect V17 which was never officially released, and also V18 beta too. V15 is fine.

If anybody wants the next V18 beta with this bug fixed, let me know.

 

[joncnunn]

Sounds like a whirlwind of a good game going there Joncnunn. How is the tech rate of the AI's?

EDIT:
If you wanted to try the Rationalism nerf, the new beta V18 I posted above is save game compatible. You could save your game just before selecting rationalism, just in case there is some problem in my code (I doubt it).

Most advanced AI was even in number of techs. Most AIs were within 3 techs. Then there is AI Indonesia & AI Russia far behind in 7th & 8th place with their war that has been going on most of the game. Russia only has their capital. While Indonesia has several cities, most of the other AIs denounced them.

I'm afraid I'd already opened Rationalism prior to that version of V18 beta. The Rean era got delayed so much that I was only one policy short of completing Aesthetics when I entered it. (Also I had built Oracle earlier.) So I had already opened Rationalism before the Industrial era began.

The Swedish war is now over. AI Sweden made reparations of half their GPT for peace.
I've rebuilt the destroyed cargo ships and have founded the coal city.
The Worlds Fair project is now under construction. I added Secularism.

RAs between the AIs only started to be signed about 10 turns ago, so none of them have completed yet.

Update:
Modern Era entered on turn 232 (Via the usual Oxford to Radio)

AI on AI RAs are only now just starting to complete.

I'm currently hosting the World's Fair (10 turns remaining)

The Indonesian - Russian war is still going on, Indonesia beat off the Russian counter offensive and is now knocked Moscow down to 20% health remaining.

About five turns ago, Austria married their first city state.

Two turns ago, the World Congress welcomed city states when half the world reached Industrial era.
As a result, Austria has become host.

My current science per turn output is 475 per turn.
My current income is 191 per turn (1 turn remaining in GA).
My culture is 489 per turn (includes both GA & World's Fair modifiers)
My tourism is 40 per turn.

I have full Tradition, Full Aesthetics, all but one 1 Rationalism policy. I'll have an ideology next turn.

China & India are both 1 tech ahead and 2 techs behind.
Austria is 3 techs behind.
Indonesia is now 1 tech ahead and 5 techs behind as it seems to solved its earlier happiness problems between some of the AI denouements against them expiring and having built a new city in a different landmass to get another unique luxury type.
Sweden is now 7 techs behind as it has severe happiness problems due to having been at war with 2 AIs + myself.
Celts are now 10 techs behind. I'm not sure why they are having problems getting AI to trade luxuries with them as they haven't been at war or denounced by anybody.
Russia is now 14 techs behind. I don't expect them to be in the game much longer. Update: As soon as I hit end turn, Moscow fell to Indonesia.

Austria has the most cultural influence for me to overcome. (Their allies of several cultural city states).

 

[glider1]

Thanks for the interesting read Joncnunn. What map are you playing I think from the pic it looks like small continents?

If you are running infoaddict, you could see what the effect of the rationalism nerf would have been because the AI science rate would be 10% higher than it is now. Question is how much of an effect that would have had on their tech rate. I think it would be quite a bit.

You wait till your next game, when the AI is going to be able to settle almost 100% of the time. It really helps the liberty AI's.

Since I have to kill the trade route income for the adaptive buff, perhaps it will balance up?

 

[joncnunn]

Thanks for the interesting read Joncnunn. What map are you playing I think from the pic it looks like small continents?

If you are running infoaddict, you could see what the effect of the rationalism nerf would have been because the AI science rate would be 10% higher than it is now. Question is how much of an effect that would have had on their tech rate. I think it would be quite a bit.

You wait till your next game, when the AI is going to be able to settle almost 100% of the time. It really helps the liberty AI's.

Since I have to kill the trade route income for the adaptive buff, perhaps it will balance up?

I actually don't have infoaddict.

This actually is a large island map, every civ did start on their own landmass, but this time the RNG creating the script placed several players primary island right next to each other, so it only required Optics.
Also, I'm at war again with Sweden, they preempted me by about 20 turns. At least I got most of my Gallasses upgraded to Frigates before this round started, but I wanted 20 turns to finish building my Dutchmen. (I have several in my build queue but have only completed one.) Sweden is also at war with my two defensive pact partners.

 

[glider1]

Do you use war to gain population and shrink the AI? In this mod, the player needs to be competitive in population to overcome the adaptive buff this mod gives the AI.

TIPS:
Capturing a big AI city is one way to increase players population and overcome the adaptive buff in this mod. It is not about boosting only the players population, but simultaneously shrinking the AI's. That is how to win this mod.

The difficulty of this mod scales progressively with map size, because with small maps, the proportion of the AI's population that is buried inside a fewer number of cities, means if the player captures an AI city on a small map the player gains more population and the AI looses more population as a proportion (relative to larger maps)

Here is the next version of BuffedAI V18beta. From what I am seeing, the AI is doing significantly better now that it settles 100% and its workers improve and repair better. It does better, because settling and improving tiles more effectively, boosts its population sooner, without applying any handicap to help it with population except the default work rate.

This version also fixes the trade route bug. It also includes the rationalism nerf (still untested)

ATTACHED: Save game compatible with any V18

 

[joncnunn]

Do you use war to gain population and shrink the AI? In this mod, the player needs to be competitive in population to overcome the adaptive buff this mod gives the AI.

In this case, because Swedish AI had declared war on me earlier, I was planning a war to punish them by taking one or more of their cities. As above they preempted me by about 20 turns while I was still building my Dutch Sea Beggars. (While I already had upgraded most of my fleet to Frigates)

The second war ended on turn 246 when my military was just short of conquering one of their cities but at the negotiating table they were willing to cede that city + all their gold on hand + over 30 gpt. (So by failing to take the city, I avoided "minor war monger", destroying any of the buildings, and population loses. Size 8 city, but its main value to me is both the unique local luxury resource and its nice strategic location value.)
During the course of the war, I had one food cargo ship route cut. The Swedes had several routes pillaged, plus a large number of ships, but they appeared to be building ships almost as fast as I could sink them.
My science per turn output is now up to 550. I went Freedom. My next policy came one turn earlier than I really wanted so it went to a level 2 Tenet instead of completing Rationalism.

Austria has married a second city state.
China completed PT.
AI Indonesia is considering AI Celts as its next target, (he asked me if I wanted to join them, I declined due to distance)

On population, I just run 3 food cargo routes to my capital, which keeps it taller than any of the AIs capitals.

 

[glider1]

Thanks for the update, internal food routes increase players population but don't decrease the AI's population, so they will have limited success at giving the player an advantage over the AI in this mod because of the way the adaptive buff works.

EDIT:
I worked out why the trade route bug was happening. You can't include the AI player you are trying to buff in the trade route calculation for the adaptive buff. Not sure why exactly but its probably because of feedback. So I've compromised the buff to exclude that condition (so the trade route buff can still work).

 

[glider1]

Here is a graphic that shows how the adaptive buff works in this mod, and also shows that the new Rationalism nerf in V18 beta is working:

Spoiler :

EDIT:
The same principal applies to gold and production.

 

[thelemon402]

Thanks for the great updates. I am not seeing the attachment for V18 tho...

 

[glider1]

Sorry thelemon, here is the latest and possibly the last V18 beta before I release later this week before the weekend. It is working well but I'm still a little concerned about late game AI science (really not sure, not enough games played to know).

Here is a list of things changed so far in latest attached:

• Better guarding
• Better worker improving
• AI chance of settling improved
• Rationalism nerfed
• AI research agreement buffs still untested
• Trade route buffing bug fixed

 

[glider1]

I think I'll delay V18, found an exciting way to ensure that the AI keeps up in the late game.

Got to try it out though properly though (let me know if you want to help).

Question about the AI. Does it rebase its aircraft from a city that it has lost or does the AI just loose those aircraft altogether?