Civ6 Blog Post - Automation

[Acken]

I find that surprising. Considering the amount of people that automate workers up to Immortal level on Civ5 it certainly was a popular feature.

As much as I don't care myself because that's just playing poorly, I'm not so sure about it.

I find it also worrying considering the AI will still need automation. Is that a way to hid a crappy algorithm and not have to deal with people complaints ? Hmm...

What makes sense though is to avoid a situation where someone automates a builder and then goes "crap where is it I need it for X" which was not an issue with immortal workers.

 

[King Jason]

In civ4 and probably even civ5 you had certain toggles for work automation that included things like;

-Automated workers leave existing improvements
-Automated workers don't chop forests.

Those sorts of threads should be particularly easy to program. Further, since I can't imagine any reason that players won't always improve nearby resources - the simplest and easiest way to implement Builder automation on some level without messing with the landscape that you intend to use for districts and aimlessly wasting builds would be to simply have the option;

-Automated Builders only improve resources.

In that case, builders would only ever do the appropriate action; Building plantations where the plantations go, building quarries where the quarries go, etc.

The only thing that couldn't be optimized would be the specific order that you might want those improvements done. However, that's the kind of micro-management some people enjoy. Others, however, don't enjoy that. I don't automate workers myself, but I really don't see any reason builders shouldn't have an auto option. Especially if you added the option I just came up with.


Also - "convinced the devs are wrong"? Really? People can have opinions that conflict with the dev's designs and neither party is wrong or right. Case in point; the graphics debate. Some people like them, some people hate them. The Designers chose it. People who don't like them can still have their frustrations, people that do like them don't have to deal with any frustrations - and the designers aren't "right" or "wrong" for going with that style. It's just the style they wanted.

 

[TheMarshmallowBear]

I think there's difference between automating Builders who have chargers and Workers who don't actually disappear.

 

[Browd]

I must say I find this whole line of argument fairly pointless, amounting to argument for argument's sake.

You build or buy a builder. If you are able to, and choose to, automate it, it wanders off and plops down 3 improvements (presumably defaulting to whatever the UI would default to if you manually moved the builder to that tile and the improve tile screen popped up) and after 3, 6, 8 maybe 10 turns (taking into account travel time), the builder is gone. So you decide to splurge on builders -- you build or buy five (or ten) builders and automate them, and 3, 6, 8 or maybe 10 turns later, they are gone as well. In the meantime, for some inexplicable reason, your 6th city still has no improved tiles. Very satisfying.

I get that you might be annoyed by having to actually think about which tiles you want improved, much less direct your builders to those tiles and instruct them what to do, but I suspect that you're only going to have 1 or 2 builders who are "alive" at any given point in time. In Civ V, you might have 6, 8, 10 or more workers sitting around with nothing to do, costing you gpt maintenance, but you still need them to deal with city border expansion and to spam railroads late in the game. I may be wrong, but I don't see that as an issue in Civ VI.

Frankly, I would expect to see greater demand for automated warfare -- after you are DOWed by Japan, you pick 6 or 8 units and tell them "go kill Japan." The AI then does all the work of moving those units into position (called "tedious" by some), just like the AI does with its own units -- and then you can watch Japan kill your units one by one (just like you can do to the AI's units). Hmm, well, on reflection, maybe that would be a bad plan....

The fact is, the human player will always make better decisions than the AI. Some don't care about that -- just see the popularity of internet videos about AI-only death matches in observer mode. But it seems to me that, if you want to "play" a strategy game, part of "playing" the game is making many decisions, large and small, including decisions like, "Hmm, this builder has one more build action. Do I want another farm, or should he mine that hill, or improve that extra luxury?"

You might say, "I don't care about that decision. I find it annoying. I wish the game would just decide that for me." Fair enough, but others might say, "I find it confusing and annoying to have to decide what techs to research. Can't the game just decide that for me?". Others might say, "I find it annoying to have to decide what buildings to build, or what (and where) to place districts, or what military units to build, or when to upgrade them, or how my gold should be spent. Can't the game just do that for me?" And yet others might complain about diplomacy, or trading, or any other system in the game, and ask "Can't the game just do that for me?"

And the answer to all of that would be, well, the AI has been programmed to do all of that for itself, so in theory all of that could be ported over and automated for the human player. But if you did that, and players actually chose to automate everything, what's left of the game?

 

[gunnerxtr]

They couldn't come up with a better topic to highlight than worker automation for the second 'First Look' video?

 

[TheMeInTeam]

A cynic's view would be that they've disabled this ability for the player because they can tell that the AI is going to derp hard on this dimension and that they don't have enough time left before release to fix it. By disabling Builder automation, they make this AI flaw considerably less transparent to the player.

Did civ 5 worker automation work like AI movement of workers? That was definitely not the case in 4 (and annoyingly, human auto workers were strictly inferior, with comparable improvement logic but would not avoid suiciding as effectively).

It barely matters since it's always been so poor, hopefully the UI/ordering/input buffering is streamlined in compensation.

You might say, "I don't care about that decision. I find it annoying. I wish the game would just decide that for me." Fair enough, but others might say, "I find it confusing and annoying to have to decide what techs to research. Can't the game just decide that for me?". Others might say, "I find it annoying to have to decide what buildings to build, or what (and where) to place districts, or what military units to build, or when to upgrade them, or how my gold should be spent. Can't the game just do that for me?" And yet others might complain about diplomacy, or trading, or any other system in the game, and ask "Can't the game just do that for me?"

Ultimately if decisions are almost certainly not going to influence the outcome, the game should be over or very near over. The tedium in the game originates from being prompted to do heaps of inputs that have almost no impact, worse still when you out-speed the UI on such orders, either due to a fast player or on recommended (which is still-slow for many game settings) specs.

 

[chaosprophet]

They couldn't come up with a better topic to highlight than worker automation for the second 'First Look' video?

It was about the change between builders and the old workers, and the choice is likely they trying to first point some of the big changes for the new game, with unstacking cities and now builders. Governments / civics / civics tree may come up next.

 

[Ryika]

You build or buy a builder. If you are able to, and choose to, automate it, it wanders off and plops down 3 improvements (presumably defaulting to whatever the UI would default to if you manually moved the builder to that tile and the improve tile screen popped up) and after 3, 6, 8 maybe 10 turns (taking into account travel time), the builder is gone. So you decide to splurge on builders -- you build or buy five (or ten) builders and automate them, and 3, 6, 8 or maybe 10 turns later, they are gone as well. In the meantime, for some inexplicable reason, your 6th city still has no improved tiles. Very satisfying.

So I build 20 additional workers, problem solved.

They'll be doing something useful, somewhere. If they leave one city untouched, great, that shows me that I need more workers.

Fair enough, but others might say, "I find it confusing and annoying to have to decide what techs to research. Can't the game just decide that for me?". Others might say, "I find it annoying to have to decide what buildings to build, or what (and where) to place districts, or what military units to build, or when to upgrade them, or how my gold should be spent. Can't the game just do that for me?" And yet others might complain about diplomacy, or trading, or any other system in the game, and ask "Can't the game just do that for me?"

And the answer to all of that would be, well, the AI has been programmed to do all of that for itself, so in theory all of that could be ported over and automated for the human player. But if you did that, and players actually chose to automate everything, what's left of the game?

Not much, but why would you care? Even if automation is available for everything, even if there were an Autoplay-Button built into the game, how would that be a problem?

As I said in a previous post... things like the tech tree don't even need "manual automation", if you're annoyed with having to choose techs you can always just click future tech or whatever your mid-term goal is and let the game do the rest. In a way that is actually automation, just not with any logic behind it.

Automated Cities... mhh... I could swear I had seen that somewhere before.

 

[KrikkitTwo]

Even if there was automation, you would still need to keep an eye on those things because of the worker charges. You built one and automated it and soon it would be used up so you had to manually build another to leave it automated. And it also would be a problem when you want to build the improvements that aren't done automatically (like forts) just to find the builder you had kept around was used up a while ago.

Then with the districts system, how are the automated workers supposed to know where to not built improvements because you're planning to build a district / wonder there soon and building an improvement there now would be basically a waste of a builder charge? It wouldn't be much of an automation if you had to fill it up with inputs like marking tiles they shouldn't work on and so on.

Well the (you want to build something else there) v. (you want Something there is a real issue) is important.

The issue is When you make the decision.
(This is why a "paint" would be good) just like they should have building queues with buildings you can't currently build. (Since you need to make decisions like barracks v. Stables)

I could see one of the 'lenses' being an " unimproved tiles" lens...so you could build some builders and send them to the right area of the map.

 

[isau]

I think the real test will be what happens when we invade an AI city. Is the conquered city going to have been competently laid out by the AI? The AI will need automated Builders even if players don't get them, and we'll see the results when we take their cities.

My guess is the AI will make many choices a player wouldn't (not being mean here, just realistic--I understand the significant challenge faced by the programmers).

 

[TheMeInTeam]

Automated Cities... mhh... I could swear I had seen that somewhere before.

Civ IV had a (not so obvious) way to automate both production choices *and* whips/rush buy, usage depending on city focus no less.

Ironic, since it was the least input-intensive entry in the main-line series when it came to giving cities orders. By miles.

 

[AriochIV]

Late game bloat is a serious issue in 4X games, but I think automation is the least satisfactory way of reducing it. If a decision isn't interesting enough to make yourself, the game shouldn't be asking the player to make it in the first place. I realize that there are folks who enjoy watching the AI play itself, but I think it stops being a game at that point.

Civ VI has done a lot to make two of the most tedious late-game tasks -- Worker control and production choices -- a lot more interesting. It might be nice to be able to queue Builder commands in the same way you can choose production choices, but I don't see any need for automation. If the game reaches a point where you can't be bothered to make these choices, perhaps you should do what my father does: end the game and start a new one. Or maybe play on a smaller map.

It's amusing that people clamor for larger maps and more cities, and then belly ache because there's no automation to help them manage them.

 

[qwerty25]

I think the real test will be what happens when we invade an AI city. Is the conquered city going to have been competently laid out by the AI? The AI will need automated Builders even if players don't get them, and we'll see the results when we take their cities.

My guess is the AI will make many choices a player wouldn't (not being mean here, just realistic--I understand the significant challenge faced by the programmers).

Ehh honestly i dont see coding a buiilder ai to be that hard if youre okay with subpar results. They already have a suggested tile algorithm.

Then they just need for pop of city have X amount of tiles improved. If X isnt reaches produce another builder. And improve tiles until X reached. Obviously resources first and everything. Really it isnt that hard to tell the AI to just build stuff.

The harder issue is the Ai deciding when to place districts and when to shift its economy. Then it actually has to decide what kind of tiles to place. (Like low on gold place gold tiles or need faith place holy stuff)

 

[Ryika]

Late game bloat is a serious issue in 4X games, but I think automation is the least satisfactory way of reducing it. If a decision isn't interesting enough to make yourself, the game shouldn't be asking the player to make it in the first place.

I agree, but how would you solve it? The only way I can of is to make it so that, after a certain point, certain Improvements are simply built automatically on Tiles that have no Improvements themselves.

Might even work, if basic Improvements have no cost to them.

 

[m15a]

Edit: I missed the last page of this discussion before I posted. Some of this was already said by others I think. Oh well.

People will expect to be able to automate and be really frustrated with what they get (no automation at all).

It's a really stupid argument, no matter how often you repeat it. "Better give them nothing than give them a system that somewhat works, that will make them happy!"

Except that people that haven't played Civ VI don't know what will and won't make them happy (and even after playing Civ V, they don't necessarily know what makes them happy when they play - they're not experts on designing the game). Of course there will be some people that are frustrated with everything in the game that doesn't work the way they want. Every change can be interpreted as a loss. But designers have to look at the game as a whole and decide what people would most enjoy overall.

Automation wasn't put in the game because players love to have things automated. The designers have to take a step back and see what players actually like to do and try to serve those goals. People use automation because they want go focus on other more important decisions, so the developers tried to make builders make important decisions without taking up too much mental space (by only having them last a few turns). And they took away the need to have workers keep building unimportant improvements in the late game just to give the units something to do.

Maybe the designers didn't do enough and the late game improvements will still be uninteresting, but just tacking automation on as a fix won't help that. If you don't care about those late game improvements, you shouldn't have to build them at all. That *may* be a valid complaint that the designers will consider, or maybe already have considered.

 

[isau]

Ehh honestly i dont see coding a buiilder ai to be that hard if youre okay with subpar results. They already have a suggested tile algorithm.

Then they just need for pop of city have X amount of tiles improved. If X isnt reaches produce another builder. And improve tiles until X reached. Obviously resources first and everything. Really it isnt that hard to tell the AI to just build stuff.

The harder issue is the Ai deciding when to place districts and when to shift its economy. Then it actually has to decide what kind of tiles to place. (Like low on gold place gold tiles or need faith place holy stuff)

I don't think it's quite that easy. In order to use Builders intelligently, the AI needs to plan where its Districts might eventually go, so it doesn't waste Builder charges on stuff they'll just end up building over. This is why Builder AI will be challenging. The AI has to be able to make long term plans and avoid expending its resources. It can't just think in the moment.

 

[AriochIV]

I agree, but how would you solve it? The only way I can of is to make it so that, after a certain point, certain Improvements are simply built automatically on Tiles that have no Improvements themselves.

Well, they've done the first important thing, which is to make the decisions impactful. If the outcome of the game is still in the balance, you're going to want to make those decisions yourself.

If it's late in the game and you've essentially won and yet you're still somehow acquiring or founding new cities, and you don't care how they're developed, then it really doesn't matter if you develop them at all. So don't.

But the larger point is that if you've essentially won, the game should be over. So this is where victory conditions can help. There should a "mercy rule" victory condition, which was the Galactic Council in MOO2 and the Domination victory prior to Civ V, which gives you the win once you reach some overwhelming percentage of territory and population instead of making you slog through an anti-climactic artificial endgame. So I hope they bring back the Civ IV version of the Domination victory mechanics.

 

[qwerty25]

Perhaps slightly more disconcerting is the loss of continuous worker queues from civ 4.

I hope theres a way to just flag what you want the tile to turn into ahead of time and have nearby builders just automatically build them.



Sent from my LG-H850 using Tapatalk

 

[isau]

So I hope they bring back the Civ IV version of the Domination victory mechanics.

I agree. I loved that win condition. It meant if someone was threatening to win you couldn't just turtle up your capital. This is a world domination game, and if someone conquers most of the world, they should win.

If you don't want them to win, you have to stop them, actively.

 

[m15a]

Perhaps slightly more disconcerting is the loss of continuous worker queues from civ 4.

I hope theres a way to just flag what you want the tile to turn into ahead of time and have nearby builders just automatically build them.



Sent from my LG-H850 using Tapatalk

Yeah that would be good. Or since theoretically, you shouldn't have many builders standing around, it'd probably make more sense to queue the builder actions on a per-unit basis.

That's a different kind of automation, though. It's automation in the sense that you tell a unit to move 8 tiles away (still in the game) but not in the sense of worker automation, where improvement decisions are made for you.