Page 1 of 1
All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-21 08:51pm
by ray245
Civilization VI: Rise and Fall wants to solve a problem. That problem is perpetual growth, and it plagues many 4X games. Whether your aim is world conquest or cultural hegemony, victory in Civilization and many of its cohorts depends on domination. However peacefully you try to play, you’re often straight-jacketed into a utilitarian-psychotic view where all resources and people are just raw material to be assimilated, Borg-like, until the whole map is monochrome.
But as the early excitement of exploration and expansion ebbs to late game stagnation, the fun runs out. Historically, stagnating empires tend to fragment and collapse. But in Civilization VI, like many games, you’re the star of the show – and there’s nowhere to go but up.
Through the Civilization lens, raw economic strength is success. Population is power. Stateless barbarians and rival empires are the only existential threat. Only growth matters. Invest, improve, reap the rewards, invest again.
That means that sooner or later, in every Civ game, you’ll reach a point where the challenge is gone but there’s still a long grind before you reach the point at which you have enough capital cities, culture points, rocket launches or religious conversions to win the game.
Rise and Fall, the first major expansion for Civ VI, makes bold moves to enliven the endgame, like introducing Dark Ages and “city loyalty” , which makes your cities more liable to defect but also gives you a chance of phoenix-like rebirth into a Heroic Age. The intent is to inject that early game dynamism into established empires, encouraging them to, well, rise and fall.
This problem of a stagnant endgame is well known to designers, and most modern 4X games attempt a solution. Stellaris breaks up its galactic-scale bloat in the mid-to-late game with robot uprisings and invasions from deep space. Endless Legend‘s long winters curtail growth. Crusader Kings II checks expansion with rebellious vassals, who have their own ambitions, independent of the player’s.
But all these attempts to make the player feel like David when they’re clearly Goliath – to give them a heroic arc – are ultimately unsatisfactory, because they’re trying to escape the fundamental principles of their genre. The 4X model is perpetual growth. In order to escape the usual flow, and stagnation in the design, these games would have to rethink their core mechanics completely.
The idea of perpetual growth underpins much of our society, but games seem uniquely committed to it as a medium. It can be seen everywhere from the chase for highscores to the consumerist dreams of The Sims, who buy better things in order to enjoy better lives. Perhaps gaming’s roots in the toy and consumer electronics industries are one reason for the emphasis on growth; the constant hankering for bigger, faster, more. We climb the tech tree, we level up, we collect bigger and bigger weapons because of a widespread assumption that growth is an inherent good.
In the traditional hero’s journey, the hero must surrender wealth, status, or power in order to attain what they really need (love, self-respect, spiritual enlightenment). But that can be hard to model in games – ‘press F for epiphany’ – without taking away player agency completely. It’s a narrative device that doesn’t easily translate into a mechanical choice or consequence. Often, the hero’s journey is reduced to material progression. Start with a wooden sword, end with a crystal sword. We get the first part of the story, but not the reversal; all rise and no fall.
There is a historical context for this modern myth of perpetual growth. It emerged from the Industrial Revolution, when incremental technological progress combined with the fruits of empire – a massive influx of natural resources and slave labour – produced unprecedented economic increase and a global population explosion from 1 billion in 1804 to 6 billion in 1999. Many of us act as if we believe this will continue forever, even to the stars themselves.
It’s worth emphasising how unusual this view is and how much it is tied to our specific historical moment. Whereas Civilization VI and its predecessors see history as a glorious upwards march, previous societies saw it as a cycle or decline. The Greeks believed in a descent from the Golden Age to the Iron Age, where life was hard and children were ungrateful. The Aztecs knew the four preceding worlds were destroyed by the gods, and theirs too was doomed to die. Medieval Europe looked back to the Romans as the pinnacle of civilisation; when the Romans ruled, the pyramids were already 2000 years old. Pre-industrial peoples lived among the ruins of ‘greater’ cultures, and took the message to heart: remember you are mortal.
We live in a brief historical blip where abundant resources and a few centuries of astounding progress have allowed many to believe that the good times will roll forever. The universal law is progress, from the wheel to the flying car, Magna Carta to Martin Luther King, and not the cycle of life and death that we observe in every other aspect of the universe. We live in a dream of immortality.
So it isn’t just a narrative problem that Civilization, and games in general, insist on perpetual growth. Yes, it produces the late game stagnation that Rise and Fall tries valiantly to shake up. But when games hew so closely to the perpetual growth model, they uncritically reinforce one of the most damaging myths of our time – a myth currently destroying our only biosphere.
In Civilization VI, climate change has been written out entirely, even as we live through planet-wide ecological collapse in the real world. The novel inclusion of natural beauty (as ‘Appeal’) only gives modifiers to growth. Famine is a minor inconvenience in your grand plan, as it was to empire-builders in Ireland or India. While older Civilization games included climate change mechanics (Alpha Centauri even set psychic death worms on polluters) Civilization VI is reluctant to take a side on ‘controversial issues’.
So however many farms you build, the soil won’t degrade. You can’t poison or overfish the seas, and those whales that provide you with luxuries will never go extinct. Your mines and factories won’t taint the air. The forests you fell will never erode the soil to the point where your cities flood and roads crumble. Even in Rise and Fall, your empire enters a Dark Age by failing to grow enough, rather than, say, through overconsumption.
However welcome an improvement, Rise and Fall can’t fix Civilization. The problem is too deeply embedded. But there are games that point to alternative models. The feuding characters of Crusader Kings II and the story events of Stellaris mitigate brute expansion. Pirates! ages your captain, eventually forcing him to retire; a similar process could afflict empires. The forthcoming At the Gates casts you as a barbarian working to burn the empire down. Fate of the World addresses climate change directly, and emphasises both the destruction it could cause, and the sheer difficulty of abandoning fossil fuels in a world so dependent on them.
Then there are games of decline. DEFCON is about minimising losses: can any of your civilians survive full-scale nuclear war? The browser-based Seedship penalises players who search for perfection; the longer you drift through space, the more likely you are to lose your precious human cargo. I can imagine a strategy game that focused on, instead of ruthless growth, the distribution of finite resources, or balancing the needs of different communities and the natural world. We have Sid Meier’s Civilization; what would Ursula Le Guin’s Civilization look like?
It’s vital we think of alternatives that challenge the myth of perpetual growth, instead of uncritically repeating it, not simply to create variety in strategy design, but also because that same myth is partly responsible for the destruction of our own world. Now more than ever, we need to imagine civilisation in new ways, and to imagine civilisations which don’t take as their underlying principles the drive to expand, exploit, and exterminate until the last lights go out.
https://www.rockpapershotgun.com/2018/0 ... ore-525079
As someone that spend a bit of time playing grand strategy games, I would agree with some of the views expressed in this article. Your faction or empire is nothing more than a problem to be solved. Your game ends when your empire is destroyed, and you never get to experience the world after your empire has collapsed.
How many times have a player build up an unstoppable empire that can take over the entire world because there's very little game mechanics that makes a world conquest difficult? A player simply need time and some decent skills to cover the whole globe.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-21 10:01pm
by Tribble
Part of the problem may be that it is difficult to program A.I. which can competently deal with the complexity and nuances involved. For example, the A.I.s in Civ 5 and Civ 6 simply cannot handle one-unit-per-tile restrictions effectively; the more the games progress and the more unit types become available the worse they got. By the time you reach the endgame they are almost laughably inept.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-22 04:59am
by AniThyng
Bear in mind that while players want a challenging AI, no one actually wants to lose - what a game like R&F needs to try to do is ensure that even the "fall" part feels to the player like"progress". Otherwise people will just give up and start a new game when they see that they are losing.
What i disagree with in the article is that not all games need to send a message of decay and failure - Civil is still really a very PG13 cheery game even if it has war and nuclear genocide. But it's not some grim hard nosed game that will leave you feeling depressed about the state of the world and the pointless loss of life in war...
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-22 11:36am
by TheFeniX
Isn't this the same series that pays homage to an old programming snafu by constantly making Gandhi a warmonger? This really reads more as a "wouldn't this be cool" idea thread from a fan. But the coding involved puts it out of the realm of most development.
But true on the AI part. There's not a lot of development there in general. Most of the big hitters who could make something worthwhile are dead or cashed out. And those were all shooter games. RTS in general has issues, 4X even more so, when you start talking about multiple AI instances (which AFAIK they don't even bother with) and giving them a win state other than "demolish the player."
This always turned me off to the series. The AI isn't trying to win, it's trying to make you lose. That works really well in a shooter, but looks psychotic in a 4X game. The AI tends to gang up up you as if controlled by the same player (it is) and relies much less on strategy than straight cheating. Because they have problems with overhead or just weird AI issues. And I get it's hard to make the computer play fair. It knows everything that is going on, telling it to forget that, or keeping certain info from it becomes an involved task. Also, the speed and precision it can bring to the table when it has a perfect understanding of the game rules.
Now, having another AI (or multiple) running in your own faction that leads to this downfall is more overheard. So they usually just ignore it and focus on more scripted scenarios. It's a shame that at the time we have more to work with than a 486DX, AI seems to have gone the way of the Dodo.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-22 12:44pm
by ray245
TheFeniX wrote: ↑2018-03-22 11:36am
Isn't this the same series that pays homage to an old programming snafu by constantly making Gandhi a warmonger? This really reads more as a "wouldn't this be cool" idea thread from a fan. But the coding involved puts it out of the realm of most development.
But true on the AI part. There's not a lot of development there in general. Most of the big hitters who could make something worthwhile are dead or cashed out. And those were all shooter games. RTS in general has issues, 4X even more so, when you start talking about multiple AI instances (which AFAIK they don't even bother with) and giving them a win state other than "demolish the player."
This always turned me off to the series. The AI isn't trying to win, it's trying to make you lose. That works really well in a shooter, but looks psychotic in a 4X game. The AI tends to gang up up you as if controlled by the same player (it is) and relies much less on strategy than straight cheating. Because they have problems with overhead or just weird AI issues. And I get it's hard to make the computer play fair. It knows everything that is going on, telling it to forget that, or keeping certain info from it becomes an involved task. Also, the speed and precision it can bring to the table when it has a perfect understanding of the game rules.
Now, having another AI (or multiple) running in your own faction that leads to this downfall is more overheard. So they usually just ignore it and focus on more scripted scenarios. It's a shame that at the time we have more to work with than a 486DX, AI seems to have gone the way of the Dodo.
That's the problem with game design than the AI itself. How a game designer conceptualize the game will affect how they go about building the AI. Most game designers still think of grand strategy games with multiple factions as a challenge a player has to overcome, when it should more about learning to survive in a vast ecosystem of different factions with their own little objectives.
Grand strategy games should be thought of as a version of MMO-RPG, where the player faction is learning to survive in a large and hostile world that carry on with their lives regardless of the player. Sometime goals do conflict and hence the player must learn how to navigate in such a world and survive.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-22 01:22pm
by Tribble
TheFeniX wrote: ↑2018-03-22 11:36am
Isn't this the same series that pays homage to an old programming snafu by constantly making Gandhi a warmonger? This really reads more as a "wouldn't this be cool" idea thread from a fan. But the coding involved puts it out of the realm of most development.
Yep. In the original Civ Gandhi had an aggression rating of 1 (out of 10), and was supposed to be the most peaceful civ in the game. When a civ switched its government to democracy that would reduce its aggression by 2, and since Gandhi would inevitably switch to democracy this would cause his aggression rating to go to -1, which then rolled over to an aggression rating of 255... on a 10 point scale. This also tended to happen around the same time nukes were developed, so Gandhi would suddenly from peaceful builder to genocidal maniac WHOSE WORDS WERE BACKED BY NUCLEAR WEAPONS!
The programmers and players (including myself) found it hilarious, and it's turned into something of a tradition. More recent games give a separate ratings for warmongering and use of nukes, so Gandhi is now peaceful builder... who also happens to stockpile nukes and won't hesitate to use them. And he doesn't like people who declare war, but will happily cheer them on when they drop nukes on their opponents!
It's that kind of schizophrenic behaviour that I love most about the series; all of the AIs have their own unique brand of crazy
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-22 01:24pm
by TheFeniX
ray245 wrote: ↑2018-03-22 12:44pmThat's the problem with game design than the AI itself. How a game designer conceptualize the game will affect how they go about building the AI. Most game designers still think of grand strategy games with multiple factions as a challenge a player has to overcome, when it should more about learning to survive in a vast ecosystem of different factions with their own little objectives.
True, but the scope and desire for an evolving metagame is pretty hard to conceptualize for AI programming. As the amount of actors and possible outcomes increases, you need that much more decision making. It gets even crazier when the AI has motivations other than "kill player."
To take the MMO example, near all of your interactions with NPCs is scripted, there's no real AI functionality there. Even the combat AI can be simple because they have a very small decision tree in the "what to use to kill player." They might emphasize moving behind you, surrounding you and keeping X distance from another friendly unit (to stall unit stacking targeting issues). They might hit X% HP then enter a "flee state." Or HP<% on friendly unit, cast heal. These are so simple even a moron like me could logically lay them out for use in coding. But even combat AI with the only goal being "kill player" can get crazy.
Imagine that, the game knows your defensives are on cooldown. But we don't want to cheat here, so we check to see if that enemy "saw" you cast your defensive. And then the required amount of time has passed for it to wear off, now they can cast "killinembomb spell." Now expand that decision making process to the 100s of scenarios with the 100s of unit you have and they have.
The AI decision making tree can get out of hand quickly when you start adding different goals to an NPC. They can begin to conflict and even override other goals. Such as the infamous "murder for a broom" incident with Oblivion AI where law-abiding NPCs where each given the tool the other needed to do a job, so one killed the other and looted the broom he needed to sweep the floor because the importance of sweeping outweighted his desire to follow the law.
Beth is bad at coding, but 4X can get insane. I recall Fireaxis having an issue with their "multiple win state" AI early in development of one Civ game where an AI on task to win via Culture (I don't play the game, I just read about AI) immediately swapped over to despotism, smashed all the "whatevers," and built an army in response to the player hitting a keynote era in their own tech. They couldn't solve the problem, so they basically just scrapped the whole thing (NOTE: if anyone can find that old developer article, I would like to have another copy).
And you can compound this issue immensely with enough AIs and decision trees. While CPUs are millions of times faster today, I recall old Chess AI losing it's shit as the game got more and more moves in. The sheer number of possible moves leading to more possible moves caused the AI to nearly stall. And that's just one AI controlling one rather large branching decision process on a game that has very clear rules and has been around for over 1,000 years. And that AI is just looking for a "best move" scenario to beat you while losing as few pieces as possible. Not.... being suicidal terrorists trying to break up your empire, but also waiting for the best time to move to inflict maximum damage for the amount they will lose.
Don't get me wrong, I'd love to see AI on this level. But man it would be an undertaking to make it feel real enough.
I hope one day we get something like "Havok Physics" but for AI: this procedural AI you just license and hammer into your game. But if something like that is even possible, I expect it to come out of something else besides video games as most developers are pressured to just drop it and rely on "let the AI just cheat to make it a challenge."
Hey, maybe we can get "Cloud AI" like EA did with Cim Sity! That worked out great, right?
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-22 01:46pm
by TheFeniX
Tribble wrote: ↑2018-03-22 01:22pmYep. In the original Civ Gandhi had an aggression rating of 1 (out of 10), and was supposed to be the most peaceful civ in the game. When a civ switched its government to democracy that would reduce its aggression by 2, and since Gandhi would inevitably switch to democracy this would cause his aggression rating to go to -1, which then rolled over to an aggression rating of 255... on a 10 point scale. This also tended to happen around the same time nukes were developed, so Gandhi would suddenly from peaceful builder to genocidal maniac WHOSE WORDS WERE BACKED BY NUCLEAR WEAPONS!
Yea, that's the one. That's not a buffer underflow, what do they call that? When a coder forgets to put a constraint on a variable? The original release of Star Control 2 had this, you could only have a max of like 10 landing craft, but you could sell them past 0, which would then have them wrap around to like.... whatever the integer cap was for a 16-bit game... like 65,000. Of which you could the sell those for resources: infinite money.
The programmers and players (including myself) found it hilarious, and it's turned into something of a tradition. More recent games give a separate ratings for warmongering and use of nukes, so Gandhi is now peaceful builder... who also happens to stockpile nukes and won't hesitate to use them. And he doesn't like people who declare war, but will happily cheer them on when they drop nukes on their opponents!
And I don't even want to know how to code an AI that can deal with the change in logistics as technology increases, especially when some opponents might be stone age level tech, literally, while you have tanks. It would be fun to try and put on paper though.
Off-topic: another fun little "oh shit" coding moment. Fans of UFO (X-Com) complained the game wasn't hard enough, even on Superhuman difficulty. This was because the game defaulted to Beginner after every load-transition. Even though you could manually up the difficulty after a mission load, the spawns and stats for the aliens were still set to what they were on the load. Not knowing this at the time, Microprose listened to the complaints and made the sequel (Terror from the Deep) even more sadistic (NOTE: even on Beginner, UFO was not a joke of a game). Terror from the Deep also had a bug:
It defaulted to Superhuman difficulty after every load transition.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-23 11:35pm
by GuppyShark
I really like the way Crusader Kings models it. There is a personality assigned to every 'block' of the empire, and it has opinions based on the traits of its superiors and events. An unpopular ruler can be brought down by internal factions - essentially, as the ruler's reach grows, more of the threats become internal rather than external. Because of this, in CK2 you can actively influence the internal politics of rival realms, work to rise to the top of your own when you are not top dog, secure support for your heirs, etc etc. It's a lot more nuanced than 'this region is yours, it generates X minerals per turn'.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-26 05:54am
by Lord Revan
I think part of the problem is programming AI that's properly "stupid" as to not make it "game it turn 10 and you loose!". To make it so that AI makes mistakes or errors in judgement that you can exploit that still aren't (seemingly) random but at the same time isn't too easy to exploit and your status as the top dog isn't secure.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-26 12:52pm
by Lagmonster
Now I kind of want a mod that allows each player to see every single unit, city, and policy in place for each opposing Civ, and then play one-on-one in a duel against an AI that has its difficulty set to 'Skynet'. I'm curious if, barring the ability to cheat, and giving the human the same advantage that the AI has of being able to see every piece on the board, it could actually beat me, or say, top human players.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-26 02:32pm
by TheFeniX
I doubt it. You'd probably stomp them. They're designed from the ground-up to rely on cheats, such as rushing units out as they need them. I have to wonder if the AI even understand half, if any, of the underlying game mechanics. It would be like taking away Infinite spawns and grenades from Veteran Call of Duty: it would be a cake walk.
I don't think 4X AI (at least at this point in time. I believe the hardware is there, but not the will to code it) has the ability to beat a 4X player that understands the concepts of the game and applies them correctly. The AI gets bogged down in variables as the gameplay expands and they can't keep tabs on everything going on like a human can. Once we understand the mechanics, we are just much better at multi-tasking than current game AI and dealing with emergent situations on the fly. Without their cheating, they'd pose no challenge because they have a hard time making decisions (other than seemingly random ones) without a solid amount of information, even with COMPLETE information. Information humans can "Best guess" on their way to victory. AI developers have had this problem for years.
So they rely on things like "triggers," someone does X, they find Y, they rely on a scripted response for their decision making. An easy way to see this is to play a couple games of HotS on the highest AI difficulty.
Even stupidly simple AI like in Heroes of the Storm is easy to confuse. Spitting at their core used to make the entire enemy team port back to base, even if it was just one little AI mob that got through and was about, and this is the hardest difficulty. Blizzard attempted to fix this, sometimes only one will do so, sometimes none. They seem to be able to determine if they have a possible "win state" if they are pushing an objective, they are willing to trade with you, but they routinely end up in positions where even the most basic HotS players could understand the current "trade" will lose them the game. But the AI refuses to give up the objective (such as continuing to collect skulls during an Event on haunted mines, so they get the strongest golem, all while their Core burns down).
Contrast (and related) to something like Godlike bots in Unreal Tournament. Even without their "wallhacks" (as they near always have your head/body in their reticle in a 1on1 match), on a clear and unobstructed map: there is no way possible you could win. Their reaction times are essentially 0 and they will never miss with a hitscan weapon. And so, FPS AI developers have to do the opposite of 4X developers: they have to hamstring their AI to make it remotely fair. A lot of this is just adding "wait times" to when the AI is allowed to shoot you once it has an unobstructed view. As the difficulty is reduced, the timer is increased.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-29 03:59pm
by Skywalker_T-65
When I play a Paradox game (I find them more fun than Civ. The inner historian in me cries when I play Civ V and have China in Germany or something) I tend to set myself goals and just build there. While letting the rest of the world develop along its own paths. I'm a plenty good enough player to world conquest if I felt like it, but that gets boring after the tenth time taking Trebizond on a conquering spree.
Much more fun to create a limited goal (say, forming the Holy Roman Empire) and then just watching the AI fumble around in the rest of the world.
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-29 10:15pm
by Tribble
I also really like the Empire / Empire Deluxe games. The computer is pretty good at it at the higher settings and can be quite aggressive. Now that I think of it, its been awhile since I've played...
Re: All rise and no fall: how Civilization reinforces a dangerous myth
Posted: 2018-03-30 08:44am
by Zixinus
At the risk of doing "hey, wouldn't be this cool idea?" post:
The problem can be "solved" if internal stability factors would be introduced. Not just "population happiness", but things like infrastructure, resource dependencies, etc. This would significantly change the game, because the player would have to focus on the internal construction of the empire as much as the external ones. Also, as the empire develops internal stability could evolved, like environmentalism being a new thing (to the player, it would be internally constructed) that the player has to research and account for.
The reason Civ games won't really do that is that they're a franchise, a winning formula they want to keep the essence of. The whole point of 4X games is to expand madly, to focus on the race-like strategy of expansion and competing for supremacy. They rely on the fallacy of infinite growth because infinite growth isn't the goal, it is battling competitors until the position of supremacy is archived.