Worker Automation

[Spatula]

You know the situation: you build a worker for every city you have and automate them. Then you find yourself reaching 1750 AD and every city has a nice little ring of roads around it, a few mines and irrigations, the odd rail track, and there are massive empty spaces around your empire.

Now you could just say "Well that's your fault for being lazy." Perhaps it is. But there are so many things going I just don't want to waste all my time moving workers around.

So this is what I propose: A menu for each worker, telling them exactly what to do, like the 'queue' function of cities mixed with the governor system. The finer details will have to be worked out, but what do you think about the idea in general?

 

[Jon Shafer]

It's a good idea, though it still doesn't alleviate the micromanagement much. Instead of babysitting the Workers every turn, you have to babysit them every 8 turns (or whatever). You then have to remember what you've told your other Workers to do so that they don't overlap, and so on. However, I think an map overlay interface which shows Workers, their tasks and so on would make that more manageable.

And yes, I also agree that a better automated Worker AI would be nice.

 

[eromrab]

i agree with both of yah... thats why i like the idea of abstracted workers... i think it would relieve alot of problems if implemented properly...

i wish they could make those computers that were in Terminator that made everyone all scared about computers taking over the world some day... stupid computers... why can't we have smart AI?

 

[Yuri2356]

On auto-worker concept that I've found works involves creating a standard "Work order" for each type of terrain. This order is the tied to an "Enhance" or "Work" command that will make the worker perform the set tasks for that type of terrain. The work could be interrupted at any time if you have other uses for them, and the normal commands could still be there for special situations.
Example:
Setting grassland to:
Road-Mine-Railroad
Place a worker of grassland and say "Enhance" and it will do just that!

 

[fishpimp]

On auto-worker concept that I've found works involves creating a standard "Work order" for each type of terrain. This order is the tied to an "Enhance" or "Work" command that will make the worker perform the set tasks for that type of terrain. The work could be interrupted at any time if you have other uses for them, and the normal commands could still be there for special situations.
Example:
Setting grassland to:
Road-Mine-Railroad
Place a worker of grassland and say "Enhance" and it will do just that!

I support this idea. But how about adding this: a function that allows the player to click& drag to highlight an area to be improved, like selecting a block of cells in Excel or highlighting an object in a graphics utility, and having the selected workers improve that area - like a corridor between cities, a large swath of jungle, or recently captured territory that needs to be rapidly improved.

 

[Yuri2356]

I support this idea. But how about adding this: a function that allows the player to click& drag to highlight an area to be improved, like selecting a block of cells in Excel or highlighting an object in a graphics utility, and having the selected workers improve that area - like a corridor between cities, a large swath of jungle, or recently captured territory that needs to be rapidly improved.

Excellent!
Now, Should the Worker AI be programed to spread out it's units, or focus on one tile at time?

 

[Licentia]

On auto-worker concept that I've found works involves creating a standard "Work order" for each type of terrain. This order is the tied to an "Enhance" or "Work" command that will make the worker perform the set tasks for that type of terrain. The work could be interrupted at any time if you have other uses for them, and the normal commands could still be there for special situations.
Example:
Setting grassland to:
Road-Mine-Railroad
Place a worker of grassland and say "Enhance" and it will do just that!

Very nice. It will have defaults of course, then you can modify what you want. That would be fanfreakintastic

 

[dh_epic]

Why even have worker units?

How does work get done?

You highlight an area for improvement and it is improved over time. Not unlike fishpimp's suggestion.

But what about capturing units?

You still need to produce a worker unit from a city. But it is added to a collective pool.

But what about the amount of time?

The bandwidth of work you can do in any one turn is determined by the size of your worker pool.

But what about the cost?

Idle workers don't cost anything. Each improvement costs 1 gold per turn of work. 8 turns to build a mine costs 8 gold -- not unlike now, except that cost is attached to the worker maintainance.

But what about capturing units?

Workers remain on the last tile they performed work on.

But what about movement?

No more movement, no more pathfinding. They essentially teleport.

But if workers are flying all over the place, how do you keep them protected?

Have the foresight to not try to improve land near an unprotected border. Or send troops at the end of the turn if you feel nervous about a sneak attack.

But then what's to stop a worker from teleporting around the world?

Workers can only make improvements within your borders.

But then how do you build forts in enemy territory?

Create a battle engineer unit. Or let the user convert workers into battle engineers for important conquests.

But I love pathfinding strategy! And dodging captures! And exploiting those who can't do it!

You can teach pathfinding to a monkey, but you'll never be able to teach a computer how to pathfind as well as a human being. It's micromanagement. Sure you can master it, but wouldn't you rather master something more meaningful?

What do you have against workers?

I just think we can reduce the micromanagement involved in pathfinding and replace that "free time" with some other meaningful decision. It could be religion or civics, or it could involve resource management, improved culture, improved espionage, advanced diplomacy, or new military tactics. Anything really.

It's like trading a bag full of pennies for a bag full of dimes -- they might weight basically the same, but I'll take the dimes, thanks.

 

[Ivan the Kulak]

Well, workers are useful for other things. You can sell them for gold, for one thing. Also, don't forget the nationality aspect of it. I will add slave workers from my vanquished opponents into cities that I have recently conquered, or some of my own, if I have many extra. If I have a starving metropolis I will often build a worker or settler from it, if I have the luxury of doing so.

That said, I would like to see less tedious building of every needed tile improvement by workers. Changing the RR system or having rail laid by means other than using workers to build them would go a long way towards resolving the problem. Having RR built by PW, and having workers build the roads in ancient times, plus build improvements out of the city radius, like resource mines, oil derricks and luxury farms, would be a good solution to the problem.

 

[The Last Conformist]

The first thing that needs to be done to workers are stack commands; "Stack Irrigate" and so on. That alone would remove most of the tedium of late-game Worker control.

 

[fishpimp]

DH_Epic - don't forget how great it is to have workers join slow growing cities! Although I think yours is a very cleaver and creative idea, I, personally, that is, as a person, prefer workers to be represented in unit form.

 

[dh_epic]

But is moving population between cities via workers really a strategy that's central to the game? Does it feel historically incorrect without it? I could personally live without using workers as an immigration tool. Especially if they did something else to model, say, actual immigration?

They're not gonna model anything new unless you simplify something old.

But be that as it may, the worker pool could feed back into cities, with some level of constraint (e.g.: no more than one worker joining a city per turn). It's definitely not opposed to the model i proposed. I personally don't care either way.

 

[frekk]

I doubt very much that they are going to replace workers as a unit in the game. Workers (or, at least, settlers that can do terrain improvement) have been in all 3 games and it's my guess that most constants like that aren't going to see much change.

So, we're probably better off talking about how to improve automation.

 

[Arathorn]

You can teach pathfinding to a monkey, but you'll never be able to teach a computer how to pathfind as well as a human being.

Anybody else find this paradox amusing? Sure, a monkey can learn trivially, as taught by a human. But a computer can't. No way. It's too complicated for a computer. <Shaking head>

Arathorn

 

[Wycoff]

The first thing that needs to be done to workers are stack commands; "Stack Irrigate" and so on. That alone would remove most of the tedium of late-game Worker control.

I think that that's probably the simplest, best solution. Keep stacked unit movement from Civ 3, add stacked unit commands. That would cut worker tedium significantly.

 

[Spatula]

I don't see why workers having been in the game for all 3 Civs is a reason to keep them. AI stupidity has always been in, and yet most of us want to see it dead.

 

[dh_epic]

Arathorn, as someone who studied AI for four years and has a big shiny degree that says "Cognitive Science and Artificial Intelligence" on it, I'm not saying I know more than anyone in the whole world, but I mean to say that I'm not pulling that out of my ass.

Pathfinding has ALWAYS been disasterous in games. Human pathfinding makes so many fuzzy judgements and snap decisions, they kind of feel their way through it. The smartest human players are simply better pathfinders than AI for any game. Even chess, if you extend "pathfinding" to "plotting a path through a continuous sequence of gameboards", you'll find that it's hard to make a chess AI that can beat an expert.

The only point I'm trying to make is that automating workers will never be as good as if you control them yourself. So you're screwed either way -- you either automate workers and suffer a disadvantage, or you manually control workers and waste your time with something that's basically monkey work.

Wouldn't you rather be a visionary for your empire than a glorified tour guide?

 

[ThePrankMonkey]

i want a "stay within borders" button for automation, so they dont go across other civs territory to get to a city that isnt connected to my main empire. i can creatre worker in that isolated city, no need to have 40+ workers try and make a mad dash to that city and going through other civs territory i might add.

i dont like ROP and i RARELY make ROP deals. even with a MPP.