Come on Firaxis, where the hell is Alpha Centauri 2?

[Blake00]

Thought you guys might enjoy this new article over at PC Gamer.. wonder if Firaxis will get wind of it lol

"Come on Firaxis, where the hell is Alpha Centauri 2?"
https://www.pcgamer.com/come-on-firaxis-where-the-hell-is-alpha-centauri-2/

[bvanevery]

I'm mildly inspired to renew my efforts, but I'm currently lost in programming language design stuff.

Think of the drones.

[dino]

I don't want Alpha Centauri 2 from modern Firaxis.

Thankfully their publisher doesn't have rights.

[Alpha Centauri Bear]

Great topic, Blake00.
I have recently thought about the same exactly thing too. And here is what I think about it.

Most of the games are pretty limited/restricted/confined (call it whatever you want) in their ideas. They may have brilliant ideas but the game is shaped by just few of them and it has certain look and feel and, therefore, play style, etc. The absolutely great example is Civilization series continuation after 2. While 1 and 2 were more or less generic, anything beyond that was pretty much centric around certain ideas like strategical resources in Civ 3 which player uses to build units and buildings. Is this a super brilliant idea? Absolutely yes for that time. I myself played it for a year and couldn't satiate myself. Yet, at some point I got bored as I knew everything about resources and gameplay felt rigid and one sided to me. Same for sequel games. They just make a different game wrapping it around some other different idea. One would play it for 1-2 years and would wait for next version (addon). And round and round we go.

Don't get me wrong. This is not a mistake. This is by design to attract more and more customers with new waves of versions and addons. These games are dumbificated on purpose to last no longer than few years.
Planned obsolescence!

Looks like Brian Reynolds wasn't aware of the concept or he just didn't care about future marketing and Firaxis allowed him to do that. He went ahead and improved game is every aspect possible. He didn't break any base ideas from Civ 2 but added 10 times on top and then some. Seriously. If one start counting original ideas in games, SMACX would score about 10 times higher than Civ 2 which was already rich in concepts and 100 times more than Civ 3 which was released after SMACX. The level of customization and replayability is insane and unprecedent and unbelievable. It is a true masterpiece designed to shine for generations ahead and not to be replaced with follow ups of any sort. Only the process of discovering and trying different features can take literally years. At least it was for me for sure. I am still playing it since inception and don't mind to play more. No other game is that long lastingly addictive except maybe Master of Magic which also not surprisingly wasn't part of the sequel. I believe tons of players would tell the same about their experience.

In short. It was specifically designed in mind to not have sequels. The amount of features is so vast that any addition to it would be just beyond any normal player ability to remember everything. Any serious modification to it can only make it worse (as Civ sequel displays it).

Summarizing the above I can say that there are huge modding community to this game and I believe this is what you are looking for. You yourself can actively participate in development and forge the true image of SMACX 2 as you like to see it. I don't think we need anything more than that.

[bvanevery]

There's fully 3x more stuff in SMAC than is necessary for a commercially viable title of equal weight to SMAC.  Like you said, they just kept going.  That doesn't mean they "baked" all their ideas, it means they just kept going.  I'm inclined to assume there were several computer abstractionists working on the project.  This abstract capability, that abstract capability, another system over here... this does make for a lot of moddability.  It doesn't make for a point in various cases though.

Like, terrain modification.  Yeah, the idea of using simulated weather as a weapon against other factions, sounds neat in principle.  But the reality is, climate modification messes up your own empire, and is an expensive laborious way to try to mess up someone else's.  It's far more practical to send in the tanks.  This is a system that has possibility, but it doesn't have focus.  It wasn't iterated on, as a valid game design.

What do you really need from terrain modification?  A land bridge.  That's about it.

The heightmap does make illustration of nukes pretty good, but the game otherwise defeats you on making or using a lot of nukes.  The only really valid nuke finish, is to let out piles and piles of Quantum or Singularity Planet Busters in one turn, destroying everything on Planet.  That way, you don't have to face Planet's wrath.  You just win the game.  Tell me that's a planned game design?

No, they just made a lot of abstract systems and didn't spend the production time to work them all out.  Surely couldn't afford to.  Made more than they could polish, basically.  Makes AI play badly, when there are more systems than the AI can reason about.

A better project, learning the lessons of history, would aspire to complexity but not so many systems.  So that the ones that are provided, are polished, and can have AIs that competently leverage them.  Currently the AI has no idea at all what I'm going to do to it, when I send the terraformers to my coast....

You can see evidence that someone did spend a lot of time on the consequences of certain abstract systems.  Like the AI shifting the focus of its attacks from one enemy city to another.  Or doing the "5 infantry spread" against a city.  Someone spent a lot of time working on some of the combat mechanics, which is why it somewhat holds up even today.  Of course, there are blind spots, like not using air power effectively, nor being competent at amphibious assault.

[Alpha Centauri Bear]

No. I wasn't saying they did a good job at baking all the innovations. Big part of them is half baked. That's why we are modding nowadays.
My point was that they didn't restrict themselves on purpose and added so many concepts that made game interesting and replayable. Even this weather weapon in your example is stupid but interesting to try out and see what happens. That alone gives it few more game tries and then another feature and another and so on. All these unique SP, unit constructor, and on and on.
That was my answer on whether we should not await AC2. Not the answer why AC1 has so many balance holes in it.

[MilesBeyond]

Hot take: I think the eventual completion of one of the various open versions of SMAC is not only more likely than a sequel, but far preferable. In the last year between Daggerfall and Caesar 3 I've seen the immense potential open source ports have for breathing new life into old games and it's incredible. Even if Firaxis were at their prime I don't think they'd be able to do any sort of sequel that would compete (quality-wise, at least) with an open SMAC where modders have easy access to every little bit of the game and its engine.

I've also arrived at the conclusion that at this point in time it's just not feasible for a game studio to develop a competent AI for a game's release. It's not just a matter of it being a task that requires more hours than the devs can throw at it, but it's also a matter of understanding the game's mechanics and the ways people play it in a way that playtesting just doesn't cover. In order for an AI to be strong it needs to be able to evolve - to understand the metagame, understand which options are strong and which options are traps, and to respond to the ways players try to manipulate and cheese it. In other words, I think that the only reasonable way to a competent AI is to continually update and tinker with the AI after a game's release - and that's a task that will virtually always end up falling on modders rather than developers.

[bvanevery]

Even if Firaxis were at their prime I don't think they'd be able to do any sort of sequel that would compete (quality-wise, at least) with an open SMAC where modders have easy access to every little bit of the game and its engine.

Well it's a money, focus, and sustainability thing.  I just blew 1 person-year over 2.5 calendar years on the low hanging fruit, not even anything under the hood.  I'm not doing that again until I have a chance of getting $$$$$.  I've been around $0 open source for a long time and the fact is, there are limits to what those communities can produce.

I've also arrived at the conclusion that at this point in time it's just not feasible for a game studio to develop a competent AI for a game's release.

You have to be a crazy loner like me.  The minute you have multiple stakeholders, someone will carp about costs and shut down the agenda.  I don't have to answer to anybody.  If I think fewer features so the AI can be competent is the answer, nobody can gainsay me.

In other media, I think Game of Thrones taught us that if your production contains too much stuff (in their case, characters with storylines) you won't finish well.  Things like "good authors" won't necessarily execute well or finish their books.  A good backbone for a production is a finite resource.  You can't count on it to stretch indefinitely into the future, it will buckle.  When you get the lesser writers finishing it up without direction, the endgame sucks.  Too many characters to tie off, not enough quality.  It's just like AI, loaded late in a game production.

Another problem you might be underestimating in modding communities, is that of standards.  For instance, despite all the improvements I made in the tech tree pacing and balance etc., I'm not aware of any great number of people deciding that my work should be followed as any kind of a standard from now on.  I'm pretty sure the multiplayer crowd is completely ignoring me, because a few years ago we didn't like each other.

So, you end up with modders pulling in different directions, and not spending much time playing each others' work.  That's not that different from the realities of team commercial production.

[MilesBeyond]

You have to be a crazy loner like me.  The minute you have multiple stakeholders, someone will carp about costs and shut down the agenda.  I don't have to answer to anybody.  If I think fewer features so the AI can be competent is the answer, nobody can gainsay me.

I'm reminded of the old (and probably still ongoing) flame wars between fans of older Civ games and fans of Civ 5. And the defense a lot of Civ 5 fans gave for 1UPT was fascinating to me - their primary issue with Stacks of Doom was how frustrating it could be when you're just playing away and then suddenly the AI declares war and drops a doomstack in your territory that you weren't prepared for and you were screwed. They preferred 1UPT not in spite of but actually because of the fact that it made the AI less competent at warring. The point I'm getting at with this is that even if it were possible for a developer to come up with a deadly competent AI, it sometimes seems to me that the people who would want that are heavily outnumbered not just by the people who wouldn't care, but by the people who actively wouldn't want it.

And I think to an extent that's just part-and-parcel with the 4X genre. These are strategy games, but they aren't 100% strategy games - they often wind up having some element of simulation to them as well. A big part of the appeal for many players is in building and developing an empire your way, to suit a particular vision or idea you've had, rather than in finding the most efficient way possible to achieve victory. So it creates a challenge in the sense that it gives you a split playerbase. It's not like a game like, say, Starcraft 2 where it's fine if the AI is trying to stomp you as aggressively as possible because that's the whole point of the game and if you don't like it there's no point in playing in the first place. In a game like SMAC, however, some players get engrossed in trying to make their empire into a utopia or a perfectly-controlled police state with not a single drone in it or a lush, beautifully terraformed paradise and find that the AI's attempts to win actually detract from their enjoyment.

Another problem you might be underestimating in modding communities, is that of standards.  For instance, despite all the improvements I made in the tech tree pacing and balance etc., I'm not aware of any great number of people deciding that my work should be followed as any kind of a standard from now on.  I'm pretty sure the multiplayer crowd is completely ignoring me, because a few years ago we didn't like each other.

So, you end up with modders pulling in different directions, and not spending much time playing each others' work.  That's not that different from the realities of team commercial production.

I actually see that as mostly a positive thing because mods are, well, modular. For example, I think that for not just SMAC but nearly every 4X, what works for balancing singleplayer and what works for balancing multiplayer are going to be different things, so I think it's a good thing to have different projects. Similarly, it allows for different approaches: Some might prefer a more minimalistic approach that tweaks balance but avoids drastic changes, others might want a more comprehensive overhaul that addresses balance more thoroughly by adding or removing mechanics.

That's the difference between modders and team commercial production; the former gives you a bunch of different horses, some fast, some slow; the latter gives you a camel.

[Alpha Centauri Bear]

Gaming industry as any other industry aim to satisfy customers to increase the return. People buy games for entertainment which means different thing to different people. Big part of engineers and other technical worker born earlier used to read books and solve problems with their mind. That's why they entertain exercising their brain by fighting more advanced opponent. New generation people born into computers and video games delivered into their crib are entertained by moving pictures. These are people who believed that 30 strength unit fighting against 20 strength unit should have higher chances than 3 against 2. They don't need challenge. They need story. And that what industry gives them. Civ series games become more and more sophisticated in playing but much more forgiving in AI tearing unprepared player apart. All the district placement mechanics doesn't seem like a strategic game anymore. It moves toward SimCity - a city building simulator.
I won't a bit be surprised if they add road network and/or zoning and/or ability to actually erect buildings within districts in next game.

[bvanevery]

their primary issue with Stacks of Doom was how frustrating it could be when you're just playing away and then suddenly the AI declares war and drops a doomstack in your territory that you weren't prepared for and you were screwed.

Nobody likes a BS crap sandwich from out of nowhere.  The issue is shock vs. suspense.  Alfred Hitchcock said, shock is a bomb going off.  Suspense is a bomb ticking under the table.  It is important to telegraph causality to players, so that  they have an ability to plan and react.  Many things in 4X design can push the game into long term planning exercises that don't have a lot of short term relevance to what's going on in the game.

For instance, why do I hand terraform every single square of my empire, on a Huge map no less?  Because the AI sucks at it, and I think all that terraforming gives me an advantage.  But does it really?  How much "click waste" do I engage in, before something of importance like an invasion actually happens in the game?  The player time expenditure is very much disconnected from the key events of importance.

It's not enough to make an AI that can beat players.  One must also make a game where the player can readily see cause and effect, in the various game mechanical systems.

Intelligence gathering and subterfuge should probably be more a part of 4X than it typically is.  A surprise like D-Day should be an achievement, not a seemingly random ass pull "doom stack" showing up on the edge of your empire.  I'm sure the Nazis spent a lot of time monitoring Allied communications and movements, to try to figure out where the blow was coming from.  They knew they were in for it though.  That wasn't a mystery.

Again, I'm the lone wolf dev and can integrate these AI competence and player perceptual / narrative issues.  I don't have to answer to anyone who would derail the project.  The tradeoff is there's only 1 of me.

[bvanevery]

These are people who believed that 30 strength unit fighting against 20 strength unit should have higher chances than 3 against 2.

On the small scale, they aren't wrong.  If you put 30 men in a bar room brawl, against 20 men, and all participants are somewhat but not greatly competent fighters, I would expect the 20 men to be beaten up as badly as the 30 men care to do.  It is not some linear odds spread.  The advantage of the extra manpower is decisive in a situation where the combatants are not terribly skilled at multiple opponent fighting.

Now if you give everyone lethal weapons and the combat is to the death, then the results are more volatile.  It would depend on whether those 20 men are able to cut down a lot of the extra 30 men in the opening moments of the combat.  The multi-person killing power of the weapons is important.  Does everyone merely have a sharp knife, or is everyone using a submachine gun in close quarters?  Are combatants bunched up so they can be killed many at once, like in the case of a grenade or artillery shell going off?  Or do they have good dispersion, cover, and smokeless powder?

"The kind of combat assumed" is an important game design goal to communicate.  "3 to 2" is actually an underspecification of combat.  It does not authorize or promote "60% chance of victory for 1 side, 40% chance of victory for the other".  Nor 100% chance for one side, with some variable number of injuries and deaths, as in my bar room brawl example.  What matters, is getting the game designer and the players on the same page.  Stating force proportions and pretending that those inherently mean something, is a poor way to do that.

In practice, I think when the combat relationships are underspecified, players just learn how the system actually behaves through osmosis.  I don't know exactly how Musketeers perform in Civ III, like what their combat formulas are.  I do know what they could and couldn't pull off against real combat barriers.  I seem to recall that waiting for the much more offensively powerful Cannons was a good idea.  I played many games where I spewed out tons of Musketeers anyways, trying to overcome walled cities by brute force.  Unless I had truly ridiculous quantities of them, it didn't go so well.  And the ridiculous quantities, represented a lot of productivity that could have been put into some other project.  Generally seemed cheaper to wait a certain number of years for Cannons to become available, then clean out those enemy strongholds.

Similarly, there was almost no point doing long distance naval invasions.  Units were too wimpy by the time they sailed to target.