Fall patch discussion thread

[ThERat]

I would wish some exciting wars, where the AI can actually mount a challenge. Currently, the issue is that if the AI would actually put your cities in danger, they start to pillage your land and after the war, it takes forever to repair districts. That is nothing but game breaking annoying. I am sure if the AI were to be a real threat, the pillage penalties have to be a lot less punishing.

 

[dexters]

I would wish some exciting wars, where the AI can actually mount a challenge. Currently, the issue is that if the AI would actually put your cities in danger, they start to pillage your land and after the war, it takes forever to repair districts. That is nothing but game breaking annoying. I am sure if the AI were to be a real threat, the pillage penalties have to be a lot less punishing.

You want the AI to challenge you but don't like the pillaging?

If anything, the AI missed many opportunities to pillage and cripple me in my 1st and only post patch game on Trajan/Immortal

they did attack early and take my 2nd city before I took it back. In the late game, I also saw Spain surround and demolish Japanese cities with corps. artillery barrage. There's noticable improvements for sure, but more work is needed.

Overall the patch is much more stable now; running on DX12. Not getting the game crashing on exit bug anymore.
This was obviously a hot fix patch with some balance stuff thrown in. A lot of the bigger issues and requests haven't even been addressed yet.

Still waiting on a log.

 

[ThERat]

Look, I want the AI to attempt to take over a city. I do the same and never pillage as I plan to make the city mine and it should be productive. If all you achieve is to pillage, then what is there to take over? A city that you spend ages on repairing? That can't be the way.

I wouldn't mind them trying to pillage key resources, that makes sense.

 

[bladex]

been playing planet coaster. i start a new game of civ VI today with the patch installed now and than i hear this sound like a chain rattle but i don't see anything or there are no notifications anyone know what this sound does? and no it's not the sound from the trade route. it sort of goes like "chink" for about half a sec.

 

[dexters]

What did they change with the Dan Quayle rankings?

I'm still 'winning' with extremely low scores. There's no score bonus for winning apparently, so you're just ranked with your ending score, which means low scores on higher difficulty levels.

 

[cakes]

been playing planet coaster. i start a new game of civ VI today with the patch installed now and than i hear this sound like a chain rattle but i don't see anything or there are no notifications anyone know what this sound does? and no it's not the sound from the trade route. it sort of goes like "chink" for about half a sec.

sounds like your copy of the game is haunted

 

[nokmirt]

After finishing my first post-patch game I've decided to now wait for the second patch. The AI is still so ludicrously terrible that I just can't bring myself to play another game.

I played one game of Civ 6 so far. I spent 64 bucks on the digital edition, not a good investment so far. Fix the AI my friends!

 

[Zuizgond]

This is utterly pathetic.

25 years of the franchise and the 6th iteration in the series (with a patch that specifically mentions it) and the game doesn't have a functioning AI that can escort a settler.

How does anyone expect this game will ever have an AI that challenges the player on any difficulty at any stage of the game after 50 turns. Or expect to ever have a fun, exciting war with an AI that can't do the very basic of things like escort a settler or upgrade its units.

Waiting for another patch is a waste of time. The only time this game will have a functioning AI is when the community work on it. Not the Devs.

Looks like they start from scratch every time they release a new iteration in the series...

 

[Europa Victrix]

This is utterly pathetic.
25 years of the franchise and the 6th iteration in the series (with a patch that specifically mentions it) and the game doesn't have a functioning AI that can escort a settler.

Not just escorting, they obviously can't even capture barbarian settlers back. I've seen lots of random barbarian workers and settlers in my games that just muck about unmolested by the other AIs.

 

[topresch]

"Ice Age" isn't an exact period because we've had multiple cold periods on Earth I suppose. But if we're taking the last one (Pleistocene) = roughly 2.6 million to 11,700 years ago.

They should've just left out the quote about the "great beasts" and put in something about the reign of the dinosaurs instead. Pretty significant period of Earth's history, anyway..

This is not a very precise representation of the Pleistocene or "ice ages". You gotta choose from one of two descriptions:

1. The last period of ice age cycles (Pleistocene) started for roughly 2.6 Mya, and did not end 11,700 years ago, but is still ongoing. What did end in the latter was the last ice age cycle of alternating interglacials and ice ages.
2. The last period defined as an ice age started after the previous interglacial which ended c. 110,000 years ago and lasted until c. 12,000 (or 11,700) years ago, when the present interglacial began, the one we call the Holocene. However, there is no indication that this interglacial will not end in a new glacial period (unless we mess up so much that the whole system schewes).

Also, there were a lot of great beasts during the glacial period across the globe, such as the mammoths, cave bears, whooly rhinoceroces, giant lemurs, giant sloths, saber-toothed tigers, Elasmotherium, giant elks, Aepyornis etc. The dinosaurs would be WAAAAAY too weird to put in, being separated by at best (with Homo erectus) 64-65 million years. The great beasts of the ice age and a bit further fits perfectly in the story of humans.

 

[Siesta Guru]

For those curious, a large part of the AI problems they have is because they decided to do a lot of decision making through what's called a 'behaviortree'. It's basically a collection of types of behaviors (move / attack / settle/ collect units/ upgrade / declare war/plan a units construction/ etc) that are done in some succession. Some of these are assigned to operations, which is basically a team of units doing some behavior together.
The idea may have originally been to make these types of behaviors easy to edit and even moddable. But I think it just ended up flopping horribly. The problem with it is mostly that these behaviortrees are almost completely blind to anything, there's very very few conditionals in there that would allow anything even resembling smart decision making. Units get constantly locked inside of behavior nodes, such as having them move to somewhere, without any way of getting them out of there again.

An example of a conditional that is missing is that there's nothing there that allows you to check whether your settler is on the same tile as a unit. You can tell the behaviortree to form formations, but it only seems to do that if the units are already on the same tile, otherwise itll just pass the 'formation making node' and continue on as if nothing went wrong.

Another example is that when units are locked into an operation, you cant have them switch over to another task. Capturing civilians is done through a different operation/behaviortree than say attacking cities is. So if a unit locked to 'attacking a city' sees an unescorted civilian settler, its gonna run right past it.

And when attacking a city, there is nothing there that allows you to check the health of the enemy city. If humans see that a city is at 5 health, they'll attack it quickly before its too late. The AI can't see it, and will be distracted by anything it would ordinarily be distracted by. Move a unit out of the city a turn before its captured, and its rather likely that all of the enemy units will decide to chase after that unit. You buy a turn, get some city health back, find some reinforcements, and the war is won.

Unless they either abandon or significantly improve on this system, no big AI steps can take place. Sure, small improvements here and there can be done (check out my mod AI+ for some improvements), but nothing that will actually resemble smart behavior, especially in those areas most visible, like combat.

 

[c4c6]

Biggest problem is they require iron/niter to upgrade. Both seem very rare and require a tech to reveal so you are stuck with regular warriors until infantry if you don't get either one.

I like a lot the civ4 approach: use copper as strategic resource! A real alternative to iron - corresponding unit (generic Axeman) is earlier available, but weaker.

(I'm aware, that this won't solve the current problem of the AI being hesitant in developing the resources, but it would smooth the principle problem of having NO iron at all.)

 

[bladex]

AI TUNING]

• Adjusted AI victory condition focus to increase their competitiveness in Science and Tourism.
• Increased AI competitiveness in building a more advanced military.
• Increased AI value of upgrading units.
• Tuned AI city and unit build planning.
• Improved the ability of city-states to maintain a strong military.

NONE of this happens in my games. think i'll stick to planet coaster till the next patch.

 

[CaiusDrewart]

For those curious, a large part of the AI problems they have is because they decided to do a lot of decision making through what's called a 'behaviortree'. It's basically a collection of types of behaviors (move / attack / settle/ collect units/ upgrade / declare war/plan a units construction/ etc) that are done in some succession. Some of these are assigned to operations, which is basically a team of units doing some behavior together.
The idea may have originally been to make these types of behaviors easy to edit and even moddable. But I think it just ended up flopping horribly. The problem with it is mostly that these behaviortrees are almost completely blind to anything, there's very very few conditionals in there that would allow anything even resembling smart decision making. Units get constantly locked inside of behavior nodes, such as having them move to somewhere, without any way of getting them out of there again.

An example of a conditional that is missing is that there's nothing there that allows you to check whether your settler is on the same tile as a unit. You can tell the behaviortree to form formations, but it only seems to do that if the units are already on the same tile, otherwise itll just pass the 'formation making node' and continue on as if nothing went wrong.

This is so interesting. I just could not fathom how Firaxis failed to get the AI to escort its settlers, given that the AI has been easily able to do this in past Civ games and that Firaxis even created a fancy new escort feature for Civ VI. But your explanation makes sense.

I got about 10 games out of Civ VI, which is a pretty fair amount. But I think the AI is so bad that I can't play it much more. There are so many cool new features and so much potential. But the combat part of the game isn't fun if the AI mills its units around and doesn't attack, and the expansion part of the game isn't fun if half the AIs can't get past three tiny cities on Deity. The fall patch didn't really change these things. So probably back to Civ IV for a while for me.

 

[Europa Victrix]

Unless they either abandon or significantly improve on this system, no big AI steps can take place. Sure, small improvements here and there can be done (check out my mod AI+ for some improvements), but nothing that will actually resemble smart behavior, especially in those areas most visible, like combat.

Out of curiosity, how did the AI work in Civ5 exactly?

 

[topresch]

I like a lot the civ4 approach: use copper as strategic resource! A real alternative to iron - corresponding unit (generic Axeman) is earlier available, but weaker.

Perhaps make things even more realistic with very limited amounts of highly valuable tin needed for Bronze Working units (c. 90 % copper and 10 % tin is needed to make bronze). During the European Bronze Age for instance, copper was found across the region, but tin was only found in Devon/Cornwall, Iberia, Brittany and the Czech Republic of today, making aristocratic trade networks extremely valuable and wide-reaching.

 

[Sascha77]

This is not a very precise representation of the Pleistocene or "ice ages". You gotta choose from one of two descriptions:

1. The last period of ice age cycles (Pleistocene) started for roughly 2.6 Mya, and did not end 11,700 years ago, but is still ongoing. What did end in the latter was the last ice age cycle of alternating interglacials and ice ages.
2. The last period defined as an ice age started after the previous interglacial which ended c. 110,000 years ago and lasted until c. 12,000 (or 11,700) years ago, when the present interglacial began, the one we call the Holocene. However, there is no indication that this interglacial will not end in a new glacial period (unless we mess up so much that the whole system schewes).

Also, there were a lot of great beasts during the glacial period across the globe, such as the mammoths, cave bears, whooly rhinoceroces, giant lemurs, giant sloths, saber-toothed tigers, Elasmotherium, giant elks, Aepyornis etc. The dinosaurs would be WAAAAAY too weird to put in, being separated by at best (with Homo erectus) 64-65 million years. The great beasts of the ice age and a bit further fits perfectly in the story of humans.

Err.. yes? So how does any of that make the quote less wrong?

2.6 million years ago is still way "younger" than 3.4 million years ago. So the order of "first stirrings of life -> ice age -> man taking his first upright steps" still doesn't make sense.

Or, in plain old numbers:
from 3.5 billion years ago ("first stirrings") to 3.4 million years ago ("stone age") - or "2.6 million years ago ("ice age"), to 4 million years ago ("first upright steps", best-case scenario).
That's like saying "from the reign of the Roman Empire to the horrors of World War 2 to the battles of the Napoleonic Wars".

Now, I know they wanted to end the quote on "man taking his first upright steps", but if you want to do that and still have the sentence make sense chronologically, you need an earlier event/period instead of "stone age" (one that occurred *before* man took his first upright steps).

Ok.. hair-splitting mode OFF..

S.

 

[topresch]

Err.. yes? So how does any of that make the quote less wrong?

2.6 million years ago is still way "younger" than 3.4 million years ago. So the order of "first stirrings of life -> ice age -> man taking his first upright steps" still doesn't make sense.
[...]
Ok.. hair-splitting mode OFF..

S.

It is a very hair-splitting post from me You wrote that the last ice age started 2.6 Mya and ended 11,700 years ago. However, I hold that that the last ice age started 110,000 years ago and ended 11,700 years ago, while the ice age Pleistocene CYCLE started some 2,6 Mya and has not ended (even though we label our geological period the Holocene). The seriality (stirrings etc.) you illustrate here though, I clearly agree with .

 

[Sascha77]

I hold that that the last ice age started 110,000 years ago and ended 11,700 years ago, while the ice age Pleistocene CYCLE started some 2,6 Mya and has not ended (even though we label our geological period the Holocene)

Interesting. I should really read up more on geological ages ... this stuff is fascinating.



S.

 

[ThERat]

For those curious, a large part of the AI problems they have is because they decided to do a lot of decision making through what's called a 'behaviortree'. It's basically a collection of types of behaviors (move / attack / settle/ collect units/ upgrade / declare war/plan a units construction/ etc) that are done in some succession. Some of these are assigned to operations, which is basically a team of units doing some behavior together.
The idea may have originally been to make these types of behaviors easy to edit and even moddable. But I think it just ended up flopping horribly. The problem with it is mostly that these behaviortrees are almost completely blind to anything, there's very very few conditionals in there that would allow anything even resembling smart decision making. Units get constantly locked inside of behavior nodes, such as having them move to somewhere, without any way of getting them out of there again.

Fascinating stuff. Would this mean that even if Firaxis releases the code, there is nothing much modders can do?