CIV VII: 1UPT, Stack of Doom or Carpet of Doom. What's your prefs?

[Leucarum]

That is an impressive attempt at apologia for a terrible AI

It isn’t a question of the AI being good or not good enough

The Civ6 AI is awful, especially at combat, which is pretty damn important

No. Just an explanation for why I doubt it's firaxis' top priority. Like it or not, we aren't their target audience here at civfanatics.

 

[GeneralZift]

I don't think it's 1UpT fault that the AI is bad. You can design AI for 1UPT boards. It's just much easier if there is no restrictions, if course.

Also, fixing the problems with 1UPT does not mean doing a half measure between it and stacks. It could also mean finding alternative solutions to the problems that 1UPT has (eg. Moving groups)

 

[pokiehl]

Also I suspect some of the AI issues are an intentional choice from firaxis. Civ being the definitive gateway game for 4X you probably don't want to pour quite as much effort into making an insurmountable AI, and even at higher difficulties you want the AI to be flawed so let players have the fun of moving up difficulty levels. In a world where firaxis had infinite money they could satisfy everyone, but they have to prioritize where they drop funds... and I don't think AI investment is where it's at.

It’s not about making an insurmountable AI—there’s certainly no consensus on whether the Civ AI should be difficult and competitive (consciously trying to win a game) or moreso immersive (“roleplaying”, playing a convincing role in the world for the player).

It’s ultimately about making an AI that can actually utilize the game’s systems, which we simply don’t have in Civ 6. I think basic AI competence with the gameplay is something all players can get behind.

 

[Leucarum]

It’s not about making an insurmountable AI—there’s certainly no consensus on whether the Civ AI should be difficult and competitive (consciously trying to win a game) or moreso immersive (“roleplaying”, playing a convincing role in the world for the player).

It’s ultimately about making an AI that can actually utilize the game’s systems, which we simply don’t have in Civ 6. I think basic AI competence with the gameplay is something all players can get behind.

I can definitely agree with that. The AI definitely got less able to engage with systems the more were added to Civ6. It would be nice to see the AI fully engage with whatever the player could do. Even if it doesn't do it as an 'unstoppable force'

Though this is a little disconnected from the 1UPT argument, I think? The line tends to be that 1UPT is bad because it gives more options for the player to stunt on the AI. But is that true to any relevant extent? And if it is true, then from firaxis' point of view is it a bug or a feature that the AI can be gamed?

I'll play civ7 regardless of whether it has 1UPT or stacks, but stacking is often presented as a panacaea to everything bad in Civ6 and I don't really buy that argument. I don't think it'll change much.

 

[aieeegrunt]

I can definitely agree with that. The AI definitely got less able to engage with systems the more were added to Civ6. It would be nice to see the AI fully engage with whatever the player could do. Even if it doesn't do it as an 'unstoppable force'

Though this is a little disconnected from the 1UPT argument, I think? The line tends to be that 1UPT is bad because it gives more options for the player to stunt on the AI. But is that true to any relevant extent? And if it is true, then from firaxis' point of view is it a bug or a feature that the AI can be gamed?

I'll play civ7 regardless of whether it has 1UPT or stacks, but stacking is often presented as a panacaea to everything bad in Civ6 and I don't really buy that argument. I don't think it'll change much.

I don’t think anyone is making that claim. 1UPT is certainly a pretty big issue

You can add mods to give limited stacking to Civ5 and 6, but this doesn’t fix things like terrible city and district placement, picking the worst times (for them) to go to war with you, NOT going to war with you when they should.

Even simple things like building and then blithely sending off unescorted settlers while at and losing a war

 

[Leucarum]

I don’t think anyone is making that claim.

Who says it isn’t broke and it’s only 20%?

It most certainly broke the AI.

It most certainly means a ridiculous amount of tedium solving a sliding tile puzzle just to move your units

This is a design problem that was already solved way back in the 70’s by SPI/Avalon Hill games for stacking and hex grid; it quickly iterated to three units per stack primarily to keep stack management to a minimum.

There are mods for Civ5 and Civ6 that give you this, and they work well.

 

[Patine]

Also, fixing the problems with 1UPT does not mean doing a half measure between it and stacks. It could also mean finding alternative solutions to the problems that 1UPT has (eg. Moving groups)

But some compromise between the two extremes I feel is essential and necessary for a better system, all around.

 

[Voidwalkin]

That is an impressive attempt at apologia for a terrible AI

It isn’t a question of the AI being good or not good enough

The Civ6 AI is awful, especially at combat, which is pretty damn important

Yeahhhhhhhhhhh, I have to agree. It's too clumsy. It doesn't move up tight or in any particular order and can't effectively leverage its numbers.

It's defending a city being attacked. It'll move its present archer one tile to the right of the city, rather than inside it, probably because another unit is slated to arrive in the city(but is 2 turn away). Your 3 archers 1 shot it in one turn. Never gets off a volley. Slated archer arrives, but uses 2/2 MP and ends its turn... 1 tile outside the city. Move your 3 archers up, 1 shot it in 1 turn. No volley launched.

4 warriors show up. Engage. Can only surround your triangle 3 places. 4 is a forest/hill. You shoot the 2 warriors that move up in the briefest way, and one falls, on approach. The 3rd is taking an extra long way to get around the other two and the 4th is behind forest/hill. End turn. Shoot the 2nd warrior, down, end turn. Take an attack from the 3rd long way warrior. Shoot it, rotate. Take an attack from the warrior that's used 2 turns from taking the forest hill. Shoot it down after another rotation.

In the end, you've taken 3 attacks, launched 18(!!!!!), defeated a force twice your size, and taking the city is a done deal. And this is routine as can be.


Some AI problems are an easy fix with 1 line of code(WTH DO YOU KEEP BUILDING ANTI TANK FOR? IM CRIPPLING YOUR WHOLE EMPIRE WITH 2 BOMBERS ON A CARRIER!). But problems with 1upt are fundamentally about the 10k variables in terrain and troops. Much harder and probably not solvable with this level of available GPU power.

 

[aieeegrunt]

Ya I am not seeing “panacea” in anything I said

 

[GeneralZift]

Ya I am not seeing “panacea” in anything I said

I like reading these forums just for these new words

But some compromise between the two extremes I feel is essential and necessary for a better system, all around.

Yes, perhaps, I agree. Didn't they already do a compromise from Civ5 to Civ6 - by letting you stack Religious units and Support units and Military units?
For Civ7 I could forsee them splitting Military into Melee and Ranged:
So you could stack Religious, Support, Melee, Ranged.

Anymore than that you really are risking a lot of clutter on one tile.

 

[Arcaian]

Some AI problems are an easy fix with 1 line of code(WTH DO YOU KEEP BUILDING ANTI TANK FOR? IM CRIPPLING YOUR WHOLE EMPIRE WITH 2 BOMBERS ON A CARRIER!). But problems with 1upt are fundamentally about the 10k variables in terrain and troops. Much harder and probably not solvable with this level of available GPU power.

Speaking from first hand experience, you'd be surprised at how many more lines of code are required to fix that problem than your estimate of 1 line

 

[Patine]

Yes, perhaps, I agree. Didn't they already do a compromise from Civ5 to Civ6 - by letting you stack Religious units and Support units and Military units?
For Civ7 I could forsee them splitting Military into Melee and Ranged:
So you could stack Religious, Support, Melee, Ranged.

Anymore than that you really are risking a lot of clutter on one tile.

I do believe you're approaching this from too simplistic and rigid an angle. And, this is not a physiccal chit-and-hex game where you have to manually sort the units, and pick up and move the stacks by hand. If the game is designed competently, it's UI will be able to manage any stacking scheme decided on, and relay it intuitively to the player.

 

[aieeegrunt]

I like reading these forums just for these new words



Yes, perhaps, I agree. Didn't they already do a compromise from Civ5 to Civ6 - by letting you stack Religious units and Support units and Military units?
For Civ7 I could forsee them splitting Military into Melee and Ranged:
So you could stack Religious, Support, Melee, Ranged.

Anymore than that you really are risking a lot of clutter on one tile.

Just warning you, the odds of me spelling those fancy words correctly is low

I do believe you're approaching this from too simplistic and rigid an angle. And, this is not a physiccal chit-and-hex game where you have to manually sort the units, and pick up and move the stacks by hand. If the game is designed competently, it's UI will be able to manage any stacking scheme decided on, and relay it intuitively to the player.

Look I have literally played prototypes of 3 UPT, in the form of mods for Civ5 and 6 and it works. The shape of hexagons kinda lends to being able to see and select your units without any stack management malarky

I tried to revisit Civ4, and while just about every mechanic and system ih that game is superior to Civ6, the stack malarky was just too much busywork.

Also the graphics aged like strawberries. Civ3 aged far better.

 

[Roland Johansen]

It's defending a city being attacked. It'll move its present archer one tile to the right of the city, rather than inside it, probably because another unit is slated to arrive in the city(but is 2 turn away). Your 3 archers 1 shot it in one turn. Never gets off a volley. Slated archer arrives, but uses 2/2 MP and ends its turn... 1 tile outside the city. Move your 3 archers up, 1 shot it in 1 turn. No volley launched.

4 warriors show up. Engage. Can only surround your triangle 3 places. 4 is a forest/hill. You shoot the 2 warriors that move up in the briefest way, and one falls, on approach. The 3rd is taking an extra long way to get around the other two and the 4th is behind forest/hill. End turn. Shoot the 2nd warrior, down, end turn. Take an attack from the 3rd long way warrior. Shoot it, rotate. Take an attack from the warrior that's used 2 turns from taking the forest hill. Shoot it down after another rotation.

In the end, you've taken 3 attacks, launched 18(!!!!!), defeated a force twice your size, and taking the city is a done deal. And this is routine as can be.

Nice description of the incompetence of the AI with the 1UPT system.

This 1UPT system is used in pure war games, with many differences:
- Everything is pre-placed, where the AI is given a highly advantageous position.
- The maps are mostly empty of units. Most tiles don't have a unit, making it far easier to move units relative to each other, with less blocking
- The movement speeds are far far higher, making it far easier to move units relative to each other, with less blocking.
- It is the only focus of these games, so they can build the terrain features around this combat.
- It is the only focus of these games, so they can work on building a decent AI for this form if warfare.

And even then, you typically fight against an army twice your size controlled by the AI and you barely lose units in the battle.

So forcing such a system on civilization with much denser units on the map and far lower movement speed is doomed to result in an AI incapable of forming a challenge in warfare.

Luckily many combat systems have already been developed around limited units per tile which offer a more interesting warfare, where the composition of your stack, the combined arms effects, the fact that some units are the counter to other units, are all taken into account.

 

[Patine]

Look I have literally played prototypes of 3 UPT, in the form of mods for Civ5 and 6 and it works. The shape of hexagons kinda lends to being able to see and select your units without any stack management malarky

I tried to revisit Civ4, and while just about every mechanic and system ih that game is superior to Civ6, the stack malarky was just too much busywork.

Also the graphics aged like strawberries. Civ3 aged far better.

It may work, but that doesn't mean the envelope can't be pushed even further to create a comprehensive and innvative true compromise between the two extremes. Why settle for just, "it works?"

 

[Evie]

When people suggest limited stacking, it sounds like it's just a request for more flexibility in armies/corps to some extent... So why not just allow that?

Allow up to N ( max number could be limited by tech) of the same unit to stack, and allow the stacks to be broken apart (Only at full health to avoid the question of how damage would get split and prevent gaming the system somewhat). You don't have to worry about working out what happens when you merge different types of unit, the stack still functions the same as any other unit. And it's not much different to an army?

Most suggestions for limited stacking I’ve seen emphasize that it would be about stacking units of *different* types. Which is pretty much the exact opposite of what you’re suggesting.

 

[Leucarum]

Most suggestions for limited stacking I’ve seen emphasize that it would be about stacking units of *different* types. Which is pretty much the exact opposite of what you’re suggesting.

Fair point, and it can be done that way - it has been in the past after all. Though only stacking the same units probably makes it easier to keep units' abilities differentiated from one another.

 

[Roland Johansen]

All the interesting aspects of combined arms combat only emerge when you combine multiple units of different types. You can have a situation of 5 groups of medieval swordsmen charging the enemy front line of 2 pikemen supported by the gunfire of one arquebussier, while those 2 pikemen are getting shot at from range by an early cannon disrupting their ranks. Meanwhile the small front line of 2 pikemen see 2 groups of knights emerging from their flanks charging the enemy medieval swordsmen flanks.

Who will win, 5 medieval swordsmen plus a cannon versus 2 pikemen, one arquebussier and 2 knights? It's 6 against 5, but the effect of combined arms on both sides and the technological differences may be decisive. And of course the luck of the dice and the terrain. There will likely be losses on both sides.

It may be a fight of 6 against 5 because of one side having a technological advantage in stacking or a leader combat trait or a general attached to the army.

 

[Disgustipated]

Limited stacking for me. Kind of ridiculous to have 20 or so units occupying a single tile. And 1upt takes decades just to move your army anywhere. Yeah that's realistic. Or ideally just move combat off the main map. It really doesn't work having archers shoot across the English channel.

 

[GeneralZift]

Limited stacking for me. Kind of ridiculous to have 20 or so units occupying a single tile. And 1upt takes decades just to move your army anywhere. Yeah that's realistic. Or ideally just move combat off the main map. It really doesn't work having archers shoot across the English channel.

You know, it's a game, and the realism is not all that matters. If it came to pointing out every unrealistic thing, there would be no game left...