Workers stick to city

[Lathieza]

Come on.... i really want this feature back. Island maps and such are horrible with workers on automated function.

Let me see a vote.

PS... cant believe they didnt put this function in from the start.

[Fish98]

yeah, it can be really annoying to sea your workers head off to the other side of the map, especially over sea

[Lathieza]

Aaargh forgot the part when your worker comes across an enemy unit. It shouldnt just flee when automated but become ACTIVE so i can move it out myself.

[slowtarget]

I'd go further. Instead of being able to attach a worker to a city, I want to be able to attach a worker to an area. Should be almost the same thing so far as code is concerned. Give me a button that says: "Automate improvements in this area". When pressed, it records the current worker location and then the worker automates, but never travels (or looks for tiles to improve!) beyond, say, 4 tiles from that location. When there is nothing more to upgrade, the worker goes to "sleep" at the location selected. Bonus points if you can draw the range circle on the map when selecting it.

The best part is that this addresses a number of issues:


It allows me to assign a worker to upgrade not just one city, but perhaps a cluster of cities (or rather: the area in the middle of a cluster of cities)
It would allow me to have a worker upgrade a city, but I could control the boundaries of that automation, thus keeping a worker from straying into an area that is technically owned by the city, but unsafe due to its proximity to enemies.
It would allow a worker to have an "idle station" that is both selected by the user and not in a city (thus taking up a unit slot in the city)
It would put a limit on the analysis scope of the worker automation AI. Instead of scanning your entire empire (or world) for cities/tiles to improve (which may be over a thousand tiles in a Huge map), it only looks within a 4-tile radius (~60 tiles?), saving time at the end/beginning of turns.
You could still have the normal city-automation by simply moving a worker to the city and telling it to automate in the area.


While I like all of these, its really the granularity of control (I could tile workers across my empire, doubling up on areas that are important to me) and the AI processing time improvements that I'd love to see.

[Furball]

Aaargh forgot the part when your worker comes across an enemy unit. It shouldnt just flee when automated but become ACTIVE so i can move it out myself.

This sis going to show how much i use auomated workers but i would have assumed they would activate rather than run off by themselves as they activate when manually assigned and an enemy unit is in the vicinity.

I guess this is just another reason to add to my reasons not to bother with automation.

[slowtarget]

This sis going to show how much i use auomated workers but i would have assumed they would activate rather than run off by themselves as they activate when manually assigned and an enemy unit is in the vicinity.

I was a bit confused by that, too. I haven't used automated workers since the first month of release, but even then I thought they activated when threatened. Can anyone confirm that they run?


I guess this is just another reason to add to my reasons not to bother with automation.

Indeed. Perhaps its confirmation bias or simple coincidence, but my end-of-turn times were dramatically reduced when I stopped using automated workers.

[SamBC]

I thought the 'head off to other side of the map' was just referring to when they run out of things to do nearby, or there are more urgent things to do on the other side of the map, not to do with being threatened. When they're not automated, they certainly just wake when there's enemy nearby, but I haven't automated for some time either.

[bronzeager]

i never liked this feature in IV, because once the workers ran out of things to do, they'd just congregate in your captial, where it was easy to forget about them.

i'd rather the designers worry about other things, since automated workers aren't a major factor in the game.

[Podling]

Automated workers do flee when an enemy is nearby; non automated workers stop whatever task they're working on and wait for new orders. I would prefer to keep it that way, with the caveat that the code for determining when an enemy is close enough to trigger that response is seriously borked and needs to be looked at again.

[Rinnero]

I'd go further. Instead of being able to attach a worker to a city, I want to be able to attach a worker to an area. Should be almost the same thing so far as code is concerned. Give me a button that says: "Automate improvements in this area". When pressed, it records the current worker location and then the worker automates, but never travels (or looks for tiles to improve!) beyond, say, 4 tiles from that location. When there is nothing more to upgrade, the worker goes to "sleep" at the location selected. Bonus points if you can draw the range circle on the map when selecting it.

The best part is that this addresses a number of issues:


It allows me to assign a worker to upgrade not just one city, but perhaps a cluster of cities (or rather: the area in the middle of a cluster of cities)
It would allow me to have a worker upgrade a city, but I could control the boundaries of that automation, thus keeping a worker from straying into an area that is technically owned by the city, but unsafe due to its proximity to enemies.
It would allow a worker to have an "idle station" that is both selected by the user and not in a city (thus taking up a unit slot in the city)
It would put a limit on the analysis scope of the worker automation AI. Instead of scanning your entire empire (or world) for cities/tiles to improve (which may be over a thousand tiles in a Huge map), it only looks within a 4-tile radius (~60 tiles?), saving time at the end/beginning of turns.
You could still have the normal city-automation by simply moving a worker to the city and telling it to automate in the area.


While I like all of these, its really the granularity of control (I could tile workers across my empire, doubling up on areas that are important to me) and the AI processing time improvements that I'd love to see.

Completely agree. These issues are real and your idea solves them. The only problem your "extended" improvement to worker automation compared to my "ultra" (i refer it as ultra not because its very very good :) , that means that it solves almost all possible problems, BUT requires several major changes in game, for example separate screen for priority management). That will not allow you to tell your workers to build farms on river tiles instead of trading posts.

But dont worry, any "ultra" or even "extended" solution (solves more problems than "simple" solution, requires one major change, for example coding of how to draw range circle) is unlikely to be implemented. We may hope only for the most "simpliest" solutions that are easy to implement and do not introduce anything more than one button.

[slowtarget]

Completely agree. These issues are real and your idea solves them. The only problem your "extended" improvement to worker automation compared to my "ultra" (i refer it as ultra not because its very very good :) , that means that it solves almost all possible problems, BUT requires several major changes in game, for example separate screen for priority management). That will not allow you to tell your workers to build farms on river tiles instead of trading posts.

Yeah, I was ignoring the issue of automated workers choosing to make improvements that few players would actually choose. That's sort of standard with any sort of automation, but the Civ V worker AI has some sort of obsession with trading posts. Sure, it's probably just expressing the realization of the devs that in Civ V, there is no problem that can't be solved with more gold. However, that doesn't mean that gold is always the fastest or most efficient solution. Sometimes (often when I first start improving tiles) I'm in more need of growth than gold.

Oh well. It's always something. The Civ IV improvement AI wasn't great, either. At release, I seem to remember that it loved to build farms even when they weren't helping anymore. Hopefully the Civ V AI will be improved with future patches. Of course, even if it is, it will never match what a human can do.

[SamBC]

Yeah, I was ignoring the issue of automated workers choosing to make improvements that few players would actually choose. That's sort of standard with any sort of automation, but the Civ V worker AI has some sort of obsession with trading posts. Sure, it's probably just expressing the realization of the devs that in Civ V, there is no problem that can't be solved with more gold. However, that doesn't mean that gold is always the fastest or most efficient solution. Sometimes (often when I first start improving tiles) I'm in more need of growth than gold.
I have a suspicion that the decision is based on which improvement, on a given tile, will increase the 'value' of that tile (probably using the same maths as the city-location code) most; for me, it rarely seems to suggest TPs, except during a GA, when the value of it is higher at that time because it will increase most tiles by two (or three) gold rather than one (or two). Similarly, the recommendations change when improvements get buffed by tech. I think it's likely that the auto-workers choose what to build on a given tile based on the same principles as the recommendations given for non-auto-workers.

[slowtarget]

I have a suspicion that the decision is based on which improvement, on a given tile, will increase the 'value' of that tile (probably using the same maths as the city-location code) most;

I suspect you're right, and I'll admit that I'm not an authority here. I stopped using automated workers a long time ago. Things may have (probably have) changed since then. When the game first came out, the bonus for Trading Posts (+2 gold) was probably perceived as better than the bonus for Farms (+1 food) in the general case. The result was that I was seeing loads of TPs around cities that were growing slower than a worker could improve even three cities.

So, it's not doing that now?

[SamBC]

I suspect you're right, and I'll admit that I'm not an authority here. I stopped using automated workers a long time ago. Things may have (probably have) changed since then. When the game first came out, the bonus for Trading Posts (+2 gold) was probably perceived as better than the bonus for Farms (+1 food) in the general case. The result was that I was seeing loads of TPs around cities that were growing slower than a worker could improve even three cities.

So, it's not doing that now?
Not as badly. Recommending plenty of mines, farms, the occasional lumber-mill, some TPs (especially once you have the rationalism policy that gives them science), and fairly-sensibly suggests Moais and Terrace Farms when applicable as well. The biggest problem to me is how the recommendations change during GAs. If an auto- or AI worker were improving a city during a golden age, maybe a whole team of them, it will have loads of TPs.