New Beta Version - August 16th (8/16)

[azum4roll]

I'm actually quite liking how this version plays (with the hotfix in particular). In the three games I've played I had Venice, Byzantium, and the Ottomans in the top spot. I've never seen any of those civs in the lead on the game settings I play before. I'm still a bit iffy about the resource distribution - some areas have seven or eight (mostly bonus) resources around a potential city location while others areas might have at most one or two. It doesn't seem balanced. Generally speaking I'm very happy with this patch though.

Nothing's changed with the way resources are distributed, only the frequency. So it really shouldn't be that much different from before.

I thought only production was removed from capital yields? Also an undocumented change is that all handicap multipliers were doubled, so every AI now starts by buying a tile.

 

[pineappledan]

Finished my game. DV at 377

Spoiler The Dominion of Canada :

Previous comments:

continents, standard speed, King.

• I’ve never had the AI give me a run for my money at this difficulty. The logistical AI is much better, and there are 3 civs that are beating me in tech. There were 4, until I took the Celtic capital.
• my other 2 continental neighbours, Inca and Zulu, were in perpetual war with me since late classical (it is now the renaissance). They will not let up, and I haven’t been able to beat them back decisively enough to get them to peace out. I haven’t seen such an extremely bellicose game in a long time.
• it’s not quite as intense as emperor used to be, but I feel like this version is a half-step up in difficulty

• Waking units up does not prompt "A Unit Needs Orders", so you will try to keep hitting "next turn", but nothing will happen. You have to go through each of your units to find the unit with moves left. Usually it is one that was set on alert, and got woken up by a nearby military unit, or something finished healing.
• At some point, Military Industrial Complex was changed so Forts no longer get the +3 (only military training buildings, citadels, and UTIs now). This makes me sad, and I would have spoken out against this change if I had known it was there.
• AI really likes swapping world map. I unlocked the tech, and every AI prompted me to trade for it on the same turn.
• AI isn't doing great on upgrading its units. I tore Shaka's navy apart in Renaissance even though he had more ships than me, because he was using a mix of frigates/galleass/dromons

• Overall, difficulty is higher now, at least on King. Even at this low difficulty, I couldn't catch up to the tech and policy leader on the other continent (Sumer, dark blue on minimap).
• Once I got through Renaissance, the AI let up on the constant wars. I was only in 1 other war after that, after I finished Global Wargames and pummeled Zulu into dust.
• After that, I had a healthy lead in soldiers and infrastructure
• My 3 neighbours were Celts, Zulu, and Inca.

• • Celts picked Ogma, the Science pantheon, and catapulted into a big tech lead in the early game until I crushed her.
  • Zulu were neck-and-neck with me on techs and only slightly behind in policies. The Zulu were a constant danger to me until Renaissance, where I could clearly break out with naval supremacy. They had Himeji Castle on top of all their other infantry bonuses, so I couldn't break their ground advantage until I combined naval supremacy with 25% Global Wargames attack, my own UU, and Lebensraum. In any other war this would have been overkill, but I think the Zulu are so overtuned at this point that I had no other opportunities.
  • The Inca got a a mountainous start and maintained a fairly strong presence on the continent, only slightly trailing in techs and policies. They posed a threat early, but I was buffered against them by a vassalized Celts. The most recent buffs to them didn't feel unfair, I think they are in a fairly good place; my own preference to see their mountain yields changed, so they don't have a science advantage rest on historical/cultural grounds, but for game balance they seem fine.
• Decolonization is nuts. It's a way for a DV leader to hammer down any 2nd or 3rd place civ. I passed Decolonization 3 times on other civs, and only had it declared against me once (didn't pass). The proposal only serves to reinforce an existing winner's lead, so I don't think it's a fair or good proposal.
• Not being the wonder, tech, policy, or wonder leader in this game actually worked really well for me this game. A lot of negative attention was focused on Sumer, who got sanctioned and decolonized twice. It's nice to see the AI use the WC to hammer down on their perceived greatest threat, rather than just gang up on the human.
• I maintained a very strong military presence throughout the game (One of Canada's main perks), and having high military scores really did ensure a peaceful play in the latter half.
• The AI was making fair deals, and was prompting for trades often, which is a huge improvement over the old system of having so many strategics just sit there unless I put the time in myself to sell them. Big quality of life improvement, well done.
• The AI consistently just asked me to hand them cities with nothing offered in return. It's a stupid trade, and a nuisance to have to tell them no that often.
• Pathfinding AI is definitely borked, but I wasn't playing with the latest hotfix, so that might have been addressed.

EDIT:
I will reiterate that +3science from Autocracy and that is baloney. Bring back Fort yields!

 

[JamesNinelives]

Nothing's changed with the way resources are distributed, only the frequency. So it really shouldn't be that much different from before.

Sorry, I don't know the technical way to describe what I'm seeing. I didn't try to understand it until it changed. It really is different though! Not always - not everywhere or in every game. That's why I've been running multiple games to see how it feels over time. Capital locations feel consistently good for example. I like bananas on marshes, horses near oases, and stone in the snow. But other parts do not feel right. For example, if one city location has 5 bonus resources within a 2-tile radius, but other potential cities in the same general area have 1-3. It's not just an area with lots of bonus resources either - that would be fine. It's luxuries, strategics, and lots of bonus resources all in one place. I feel as if don't have a choice because one option is so much stronger than the others. At the same time I almost feel like I'm cheating if my rivals haven't had the same luck.

Spoiler :

I thought only production was removed from capital yields?

You're right, that's what the patch notes say. I thought it was everything based on what I remember from the discussions. Now I'm not sure.

• The proposal only serves to reinforce an existing winner's lead, so I don't think it's a fair or good proposal.

• Not being the wonder, tech, policy, or wonder leader in this game actually worked really well for me this game. A lot of negative attention was focused on Sumer, who got sanctioned and decolonized twice.

Edit: These two points seem at odds with one another?

I like Decolonization a lot myself, it feels like my only peaceful tool against late-game Austria and other diplo-heavy civs. Particularly on higher difficulties. Granted I like it partly because it works for me more often than it works against me. And I could probably use Spheres of Influence and Open Doors more often.

 

[tothePAIN]

The AI is building level 4 upgrades on units. Its amazing.

Also, Britain is the strongest naval civilization right now, because they have access to logistics units again.

 

[Kim Dong Un]

The AI is building level 4 upgrades on units. Its amazing.

Also, Britain is the strongest naval civilization right now, because they have access to logistics units again.

My last game saw me lose 2 cities to Lizzy's fleet containing SotL with logistics & bombardment. On King difficulty... Scary!

 

[Stalker0]

I like Decolonization a lot myself, it feels like my only peaceful tool against late-game Austria and other diplo-heavy civs. Particularly on higher difficulties. Granted I like it partly because it works for me more often than it works against me. And I could probably use Spheres of Influence and Open Doors more often.

I also think Decolonization can be a useful weapon against the "diplo king". The trick is to recognize the need early and deploy it before its too late. Otherwise it does become a weapon to kick down the little guy, just as PAD described.

 

[pineappledan]

Edit: These two points seem at odds with one another?

I like Decolonization a lot myself, it feels like my only peaceful tool against late-game Austria and other diplo-heavy civs. Particularly on higher difficulties. Granted I like it partly because it works for me more often than it works against me. And I could probably use Spheres of Influence and Open Doors more often.

I was the Diplo and military leader, but Sumer led in everything else. The AI valued the lead Sumer had in techs, policies, wonders, and tourism as more of a threat than my control of the WC. I think the AI is right to value those things higher, but it worked in my favor, because I ranged between 3 and 5 times as many votes as Sumer, and more than 2x the next diplo leader (Greece).

So, the conditions were right, and maybe my situation is a corner case, but I don't think it is. Especially not at higher difficulties when you are forced to focus on 1 victory condition. And in my case, I was able to abuse decolonization to kick Sumer down twice and Greece down once. Open door policies and spheres of influence are already a compounding force for a diplo leader (and I stole all of Inca's allies using that tactic). It just felt 20% too dirty.

Maybe if Decolonization reduced influence to 100 max, rather than 50? At 50, and with the scaling influence decay in later eras, that city-state is only your friend for 5 more turns. At 100, if you weren't competing in the 1000s for a handful of city-states, it could potentially not even cost you allied status.

One more point I forgot to mention: The AI is putting a Very low priority on WC proposals right now. Way too low; I should almost post a github ticket because it might be a bug.

• I won every single WC proposal. Every time, there was only 1 other civ that competed. Every other civ gave 0
• In my last playthrough, International Games and UN were passed in the same WC, and I won them both. I shouldn't be able to do that.
• I posted in an earlier post, even the militarist civs completely ignored Global Wargames. I doubt I could have won my continental wars if any other civ on my continent had gotten to lvl 2 on that competition.

 

[Voremonger]

I played another game as the US, Immortal, standard speed, huge Communitu_79a with Terra.
My plan was to grab and hold as much land as possible early on and to then expand to the second continent later on.
This was somewhat complicated by Songhai who managed to take over the territories of Austria, Siam, and parts of Byzantium, making them the strongest player in the Medieval and Renaissance Era.
At the height of their power Songhai controlled 21 Cities; it was pretty stressful to fend off their attacks, especially with a River leading into my territory.
However, since I for once wasn't the most hated player in the game I was able to sanction Songhai through the World Congress; because Songhai had almost no extra votes they never got to appeal the resolution.
Eventually I was able to outscale Songhai and I killed them with Landships, Artillery, and Lebensraum.
Everyone now hated me but with ~40 Cities under my control the game was basically over.
I eventually achieved a Culture Victory when I killed the only player that I wasn't yet influential with.

Some thoughts:

1. The AI has seen significant improvements on the tactical level. I think a good way to complement this improvement would be to also work on long-term strategic planning; There is no point in being good at fighting battles if you're fighting battles where you're already at a disadvantage. In particular the logic for City placement could be improved. I think the AI currently does not do a good job at evaluating geographic barriers. They frequently settle Cities that are basically impossible to defend because it is separated by sea from their other Cities. I think the AI should prefer founding Cities on Tiles where there are geographic barriers pointing towards the Cities of other players and no geographic barriers pointing towards their own Cities. Also I think the AI are still not building enough Roads for military purposes. I think the optimal strategy would be to just build a Road on every Tile near a border to an enemy, but even just more connections between Cities would be a significant improvement. Currently the AI Road networks are something like a line or a star, a mesh would be better.
2. When I formed Defensive Pacts against Songhai a weak Tradition Civ that I had made sure would like me so that I would have a trading partner suddenly declared war on one of the Civs that I had a Defensive Pact with. I think this was a really bad decision on their part because based on my own judgement of the diplomacy logic they would have never dared to declare war on just me. Case in point: they died.
   
3. The AI seems to more frequently kill City States if they are affected by Open Door (and maybe Sphere of Influence?). I like this. It makes the AI feel like they are dynamically adjusting to the situation.
4. The AI seems to have become better at using bombers. A few patches ago they would use them mostly to attack Cities, now they use them to focus down Units. The AI have also become better at intercepting bombers with fighters. What the AI are still struggling with are Carriers. They place them too close to the front lines which allows human players to easily focus them down and kill 4 Units at once.
5. The AI is bad at utilizing Spies as Thieves. On huge maps where the number of Spies per Era is doubled this is particularly noticeable. The first problem seems to be that the AI do not seem to move their Great Works to their Capital where they cannot be stolen. The second problem seems to be that the AI either do not send enough Spies as Thieves or do not keep them in Cities long enough to steal a GW. Especially AI that are trying to achieve a Cultural Victory should be trying to steal more GWs I think.
   
6. I didn't get constantly DOWed from the other side of the world but I also had a huge military so I don't know how things would have been if I had tried to get by with less Units.
7. The Inquisition Enhancer Belief is pretty good. I usually take Zealotry for the extra Unit production but more Gold/less Unhappiness is also very useful.
8. The AI has gotten pretty good at placing their Citadels.

 

[arthurF]

Elaborate?

well, there are a number of factors obviously that the vassal AI should be considering, like tech level of its military vs tech level of the enemy, and size of military vs size of enemy ( and these should be land vs land , and water vs water comparisons), but a war is a golden opportunity (if these comparisons look favorable) to claim cities and land and glory. Maybe eventually get out from under the boot, too.
Instead, i see some vassals just puttering around, doing nothing militarily. Moving a ship here or there, defending vs an attack, sure, but not mounting an offensive.
I am sure it happens, but not any where near enough.

In a separate vassal issue, why a strong nation that has not had its capital taken voluntarily submit to vassalization ? I just saw the swedes voluntarily become vassals of persia, and the thing is, Persia and I are the two most powerful nations in the game currently. Sweden had just beaten off my attack with aplomb. Therefore, it could probably beat off a Persian attack ( well, maybe...those AI militarys are crazy huge often times). It still had its capital (the swedes did) and plenty of other cities, so why become a vassal of Persia?

 

[arthurF]

I played another game as the US, Immortal, standard speed, huge Communitu_79a with Terra.
My plan was to grab and hold as much land as possible early on and to then expand to the second continent later on.
This was somewhat complicated by Songhai who managed to take over the territories of Austria, Siam, and parts of Byzantium, making them the strongest player in the Medieval and Renaissance Era.
At the height of their power Songhai controlled 21 Cities; it was pretty stressful to fend off their attacks, especially with a River leading into my territory.
However, since I for once wasn't the most hated player in the game I was able to sanction Songhai through the World Congress; because Songhai had almost no extra votes they never got to appeal the resolution.
Eventually I was able to outscale Songhai and I killed them with Landships, Artillery, and Lebensraum.
Everyone now hated me but with ~40 Cities under my control the game was basically over.
I eventually achieved a Culture Victory when I killed the only player that I wasn't yet influential with.

Some thoughts:

1. The AI has seen significant improvements on the tactical level. I think a good way to complement this improvement would be to also work on long-term strategic planning; There is no point in being good at fighting battles if you're fighting battles where you're already at a disadvantage. In particular the logic for City placement could be improved. I think the AI currently does not do a good job at evaluating geographic barriers. They frequently settle Cities that are basically impossible to defend because it is separated by sea from their other Cities. I think the AI should prefer founding Cities on Tiles where there are geographic barriers pointing towards the Cities of other players and no geographic barriers pointing towards their own Cities. Also I think the AI are still not building enough Roads for military purposes. I think the optimal strategy would be to just build a Road on every Tile near a border to an enemy, but even just more connections between Cities would be a significant improvement. Currently the AI Road networks are something like a line or a star, a mesh would be better.
2. When I formed Defensive Pacts against Songhai a weak Tradition Civ that I had made sure would like me so that I would have a trading partner suddenly declared war on one of the Civs that I had a Defensive Pact with. I think this was a really bad decision on their part because based on my own judgement of the diplomacy logic they would have never dared to declare war on just me. Case in point: they died.
   
3. The AI seems to more frequently kill City States if they are affected by Open Door (and maybe Sphere of Influence?). I like this. It makes the AI feel like they are dynamically adjusting to the situation.
4. The AI seems to have become better at using bombers. A few patches ago they would use them mostly to attack Cities, now they use them to focus down Units. The AI have also become better at intercepting bombers with fighters. What the AI are still struggling with are Carriers. They place them too close to the front lines which allows human players to easily focus them down and kill 4 Units at once.
5. The AI is bad at utilizing Spies as Thieves. On huge maps where the number of Spies per Era is doubled this is particularly noticeable. The first problem seems to be that the AI do not seem to move their Great Works to their Capital where they cannot be stolen. The second problem seems to be that the AI either do not send enough Spies as Thieves or do not keep them in Cities long enough to steal a GW. Especially AI that are trying to achieve a Cultural Victory should be trying to steal more GWs I think.
   
6. I didn't get constantly DOWed from the other side of the world but I also had a huge military so I don't know how things would have been if I had tried to get by with less Units.
7. The Inquisition Enhancer Belief is pretty good. I usually take Zealotry for the extra Unit production but more Gold/less Unhappiness is also very useful.
8. The AI has gotten pretty good at placing their Citadels.

I second all this, and also some weak AI will still declare war on much stronger nations, even while losing a war with a second civ. Really terrible strategy.

 

[arthurF]

Decolonization is nuts. It's a way for a DV leader to hammer down any 2nd or 3rd place civ. I passed Decolonization 3 times on other civs, and only had it declared against me once (didn't pass). The proposal only serves to reinforce an existing winner's lead, so I don't think it's a fair or good proposal.

Yeah, but it is also the only way to fight against a civ that took statecraft. Statecraft makes dominance of city states nearly inevitable, with large bonuses.

 

[Recursive]

well, there are a number of factors obviously that the vassal AI should be considering, like tech level of its military vs tech level of the enemy, and size of military vs size of enemy ( and these should be land vs land , and water vs water comparisons), but a war is a golden opportunity (if these comparisons look favorable) to claim cities and land and glory. Maybe eventually get out from under the boot, too.
Instead, i see some vassals just puttering around, doing nothing militarily. Moving a ship here or there, defending vs an attack, sure, but not mounting an offensive.
I am sure it happens, but not any where near enough.

In a separate vassal issue, why a strong nation that has not had its capital taken voluntarily submit to vassalization ? I just saw the swedes voluntarily become vassals of persia, and the thing is, Persia and I are the two most powerful nations in the game currently. Sweden had just beaten off my attack with aplomb. Therefore, it could probably beat off a Persian attack ( well, maybe...those AI militarys are crazy huge often times). It still had its capital (the swedes did) and plenty of other cities, so why become a vassal of Persia?

Military decisions are ilteroi's domain.

Deciding whether to be voluntarily vassalized is part of the diplomacy AI, but it also affects the deal AI. Gazebo changed this recently so would be best to direct such comments to him, since I'm away for the weekend (on Github, preferably ).

I second all this, and also some weak AI will still declare war on much stronger nations, even while losing a war with a second civ. Really terrible strategy.

This should be fixed for next version, I improved war logic. See my post here: https://forums.civfanatics.com/threads/diplomacy-ai-development.655040/page-31#post-15891921

 

[BiteInTheMark]

Had only time for one game so far with the newest version (3), my opinion.... the tactical AI seems be improved greatly. I saw a lot of units coming together on my border and the attack was very obvious, a good mix of melee, cavalry and artillery. Unfortunatly for them, I was fielding much less but overall 1 level higher units and repelled the attack. Iam still impressed by the improvement.
The going to war decision making still needs some tweaks, I was very weak military wise in the early game but never saw any war declaration from my closest neighbors Portugal and Greece which both picked authority. Going authority should really increase the war probability of those, else its a bit like facing civs with only a half full policy tree while you self have a full tree. Otherwise, the wars AI vs AI seems to be much more dynamic and active, I saw several cities swapping its owner in my game. The aztecs were able to conqer 3 cities from 2 other civs and brought the siamese capital down to 50% till they peaced out. I also got a war declaration from them, even Iam out of their range and worse, they lost a research treatement 2 turns before its expiring. That was a bit irrational too, I think they were bribed.
Not sure if the bonuses for AIs are already that high at emperor or if they are now much better at planning, but even with a lot of wonders and GPs (which trigger my Arabian UA), other AIs are on par with me cultural wise and my lead on techs is with only 2 techs very small.

Overall a great improvement.
My most disagreement in changes is still the pantheon change for Ancestor. Its already a pantheon with only faith as reward and not easy to found with. On the other side, God of the open sky is still imba and able to give you up to 4 times more yields than Ancestor is now able to.

 

[Jiggly]

Playing as Zulu. Cant undo Raze City after pressing button, even after creating peace. Any ideas how I can fix this? Can i manually undo it in the code of a game or something?

 

[Recursive]

Playing as Zulu. Cant undo Raze City after pressing button, even after creating peace. Any ideas how I can fix this? Can i manually undo it in the code of a game or something?

Using the latest hotfix? If so, post on Github.

 

[Jiggly]

Using the latest hotfix? If so, post on Github.[/QUOTE

Yep latest hotfix

 

[InkAxis]

Playing as Zulu. Cant undo Raze City after pressing button, even after creating peace. Any ideas how I can fix this? Can i manually undo it in the code of a game or something?

Is it a holy city?

In a separate vassal issue, why a strong nation that has not had its capital taken voluntarily submit to vassalization ? I just saw the swedes voluntarily become vassals of persia, and the thing is, Persia and I are the two most powerful nations in the game currently.

Well, the thing is vassalization is technically protection. Perhaps Sweden was so scared of you it decided it needed protection from Persia.

 

[Jiggly]

Using the latest hotfix? If so, post on Github.

Do you guys finish games when you hit bugs (cant stop raze city) like this or do you restart? Issue is i dont know if its gonna happen to all cities I try to raze or if its just a one off?

 

[pineappledan]

It’s gonna happen to all cities. You haven’t been able to stop razing for a while.

 

[Recursive]

That razing bug keeps popping up...

Gonna post my war logic changelog here as well for greater visibility:

Removed tons of useless code and five unused memory variables (there was leftover code for "trade agreements" and other untradeable items, plus unused vanilla code, that was taking up around 2,500 lines of code)

AI should no longer agree to start additional wars if they are already doing badly in an existing war or in general (very unhappy, bankrupt, lost their capital)

AI now has an aggressor flag in memory that sets whether or not they WANTED to start the wars they're currently involved in. So if they declare war, it is set to true. If they were planning on it and you declare war on them first, it is also set to true.
- Aggressor flag decreases reluctance to conquer cities and makes the AI less likely to make peace

AI now has a function to identify "phony wars" that they have no interest in waging, especially if they were declared on rather than the other way around. If it considers itself to be in a phony war it will agree to make peace at the earliest opportunity.

Completely rewrote coop war logic for saner, cleaner code
- Removed exploits where you could chain coop wars to avoid war weariness: you are now capped at 15 turns where you cannot make peace, and when the lock is set your war duration is subtracted from those 15 turns.

- When a coop war agreement is made, you will automatically declare war after 10 turns has passed (no prompt to back out). If one of you goes to war before then, both of you will (regardless of whether you or your target declared). While this slightly sacrifices player choice, it's more consistent, much less exploitable and more strategic of a choice. The AI will benefit from this a lot.

- You can still back out of a coop war agreement by ending your DoF, denouncing or declaring war on your ally - or by making a DoF or DP with the target. All of these will be treated as breaking your promise and apply a minimum -50 penalty to diplomacy and a reluctance to agree to future coop wars. The person who broke the agreement gets the penalty, of course. If something makes the coop war invalid (like the target becomes a vassal) then it is cancelled automatically with no penalty.

- AIs now start a sneak attack operation against the target when they are preparing a coop war, and are more likely to request that others join their planned war, so you may see alliances of more than 2 civs declaring war at the same time

- Warmonger penalty for coop wars is fixed. No penalty for declaring, and penalty for conquering cities is reduced by 90%.

- You are no longer able to request that AIs declare war on their friends, DPs, or on vassals and other invalid players. The button is disabled (previously it was still enabled but asking would do nothing except potentially get you a penalty with the target AI). AIs also can't request this from you.

- Rewrote coop war desire evaluation. AIs should utilize coop wars more often, be more likely to start coop wars with players that have a track record of agreeing to their requests, and choose targets more wisely. Penalty for refusing to agree also scales based on the AI's Forgiveness flavor now (ranges from -10 to -30 to recent assistance)

Improvements to approach selection
- AI should be AFRAID less often

- AI will be far more reluctant to wage distant wars before they're able to cross the ocean, and more aggressive towards neighbors in the early game

- AIs that have agreed to a coop war or are planning an attack will now always have the WAR approach

- AIs that are NOT *currently* planning an attack will no longer have the WAR approach. This should result in saner, more strategic diplomacy.