AI Controls Workers Better

[cookie_monster]

I would like to see the AI control the workers better. Its criminal to see how undeveloped their land can be. Perhaps something as simple as a ratio of workers to amount of cities would be better. I think its a big problem for the game as the workers never road luxuries,resources with priority. Also they never seem to use railways very good, never having enough workers to do it quickly.

I hope you can find a solution for this.

 

[Rabid Pop Tart]

I think the AI is the biggest thing to be improved. A better job was certainly done in PtW, and I'm sure BreakAway (That's who's doing it now, right?) will tweak it some more. But yes, AI workers are ineffecient, and need a lot of improvement.

 

[dexters]

I'll just add my note here.

I've done a lot of observing on the AI worker management. Part of the problem here is that the AI treats its workers like any other unit. They disband "kill-off" workers when they don't have the income to keep them. Only to build a new worker a few turns later. It's what I call AI worker inefficiency. Most human players would keep workers around and take a few gold per turn hit on their treasury. This is not true for the AI. This needs to be fixed. If it keeps more of its workers around, it doesn't have to rebuild so many of them and they can have stronger improvements.

The reason is simple.

It basically cost the AI on two sides here. First, it loses several turns work from a disbanded worker, and when the AI rebuilds a worker, the city loses a population point and a turn of production
---------------------------------------------------------

Regarding linking luxury, the comment is very true. The AI workers work in the context of individual cities. They will link nearby resources to the nearest city quickly. But often, the AI doesn't see the big picture.

A city a few tiles away from the capital may have a luxury. The workers will be quick to link the luxury to the city closest to it. But it will stay that way for a long time. The capital may be a few tiles away and instead of linking it to the rest of the empire, it takes a long time for the workers to get around to do it. In the mean time, other cities are having to use entertainers when they could be using that population to work the land.

 

[Gengis Khan]

The worker AI is terrible!! Even late in the game I won't automate my workers(rather just skipping or fortifying them). I don't have PTW so I don't have the automate without change feature

 

[Roland Johansen]

The AI workers should irrigate grasland less. Sometimes they even irrigate grasland to go from population 28 to 29. Yes, another entertainer. That's usefull

 

[cookie_monster]

Maybe Im wrong to say this but as this thread has over 100 views why are there only 15 votes? If you dont agree fair enough! But our goal is to improve the game, I think. Hmm perhaps a selection of 10 million more units is more exciting!!
Please no one be offended. I think I have a fair point.

 

[dexters]

2 things. Views only counts the number of times a page is loaded.

A person who comes in here and posts creates at least 2 views. As the post forces a reload. I'm not sure how the voting thing works, but when I voted, I manually reloaded the page to see the results.

There's at least 10 hits from me on this page.
----------------------
Roland Johansen: The AI focus on growth has been fixed in PTW 1.21F as they will mine grassland before they get out of despotism. I still do find the AI puts a bit more emphasis on growth than production since their developed cities have more irrigated tiles than I care for.

To this end the AI can be tweaked I think to focus more on production by mining grassland when the citiy has surplus food. ie: It's wasteful to have 10 surplus food and get no shields as you grew (because the new pop points are turned into entertainers or work land with no shield points). It is much better to have 5 food surplus and a few extra shields.

 

[gozpel]

There are some things the AI workers do that is just crazy...like irrigating grasslands when in Despotism, where they probably will stay for the next couple of millenia anyways. The priorities should change with governments...so let those silly workers mine grasslands early to get better production.

And how often haven't you seen AI workers improvea city next to your borders, only to leave without even building a road to connect to your civ? I think that's silly...trade is more important than anything else, so why doesn't the AI see that way too? I don't want to improve their lands myself and I do not want to sit and wait forever until they finally find time to build a harbour.

 

[dexters]

There are some things the AI workers do that is just crazy...like irrigating grasslands when in Despotism, where they probably will stay for the next couple of millenia anyways. The priorities should change with governments...so let those silly workers mine grasslands early to get better production.

Not sure what version of Civ III you have or whether you have PTW at all, but I can confirm to you that if you own PTW and have1.21f patch, this problem is fixed.

Read my statement above as I had said precisely what I am just telling you.

 

[Roland Johansen]

Originally posted by dexters
The AI focus on growth has been fixed in PTW 1.21F as they will mine grassland before they get out of despotism. I still do find the AI puts a bit more emphasis on growth than production since their developed cities have more irrigated tiles than I care for.

To this end the AI can be tweaked I think to focus more on production by mining grassland when the citiy has surplus food. ie: It's wasteful to have 10 surplus food and get no shields as you grew (because the new pop points are turned into entertainers or work land with no shield points). It is much better to have 5 food surplus and a few extra shields.

I didn't know they changed this in PTW 1.21F. I'm a European, so no patch 1.21F for me. (There's a thread about this in the main section, the patch for Europeans is in the making and should be finished around now).

In my games I usually try to grow as fast as possible to cityfase 6/12/16-20 and then produce as much as I can (I play a little more subtle than that but that's far more difficult to explain). Could the computer player not be programmed to do that perfectly. It's not that difficult, Grow fast until maximum and then Produce maximum. Ok, you need a lot of workers to perform this changing from irrigation to mines and back. (I'm not a programmer, I admit)

 

[dexters]

1.21f PTW for Europe is out.

Check Civfanatics front page.

As for getting the AI to play a certain style to counter your style, I think that's a bit too much and would require extensive learning AI additions.

What I think cookie monsters and the others who have made their thoughts known is basically to say that the AI is not playing as EFFICIENTLY as it should be. Basically, still a lack of focus on production when growth is available, even if this growth means the AI would have to whip out those entertainers. It adds absolutely nothing to their game. Their score may be helped (a little) but 4X games like these it's all about production.

Growth is important... naturally, as larger cities can work more tiles. But as I've said, it is kind of pointless to have a large population working a bunch of irrigated tiles with no shield as a size 12 city may have no significant production advantages over a size 8 city, but at size 12, the AI has happiness problems, needs more luxuries, and still can't build units and improvements any faster.

 

[Roland Johansen]

Originally posted by dexters
1.21f PTW for Europe is out.

Check Civfanatics front page.

Thanks. I was just looking over there.

As for getting the AI to play a certain style to counter your style, I think that's a bit too much and would require extensive learning AI additions.

I don't think the AI is capable of countering my game style
I just think it is a good strategy to first grow hard and then turn those growth squares into producing squares with my large workforce.

Growth is important... naturally, as larger cities can work more tiles. But as I've said, it is kind of pointless to have a large population working a bunch of irrigated tiles with no shield as a size 12 city may have no significant production advantages over a size 8 city, but at size 12, the AI has happiness problems, needs more luxuries, and still can't build units and improvements any faster.

You're right. If the AI can't change the large fast growing cities into producing cities once they reach fase 12, because it has happiness problems, then this strategy doesn't work. I've read somewhere that the AI is very reluctant to use the luxuryslider. I think that is what is behind this problem. If it could use the slider, then it could change the large foodproducing cities into large production cities.
Another problem of the AI is that it doesn't produces enough workers. After the initial landgrab fase it is wise to upgrade your land very fast with a lot of workers (or go conquer some other civs). The AI doesn't build them enough. I upgrade my land faster and I keep enough workers so that I can change the land to whatever I like.

 

[dexters]

AI will not use luxury slider in peace time. This is confirmed by the AI programmer Soren.

It will use it up to 20% in wartime.

As a general base rate, I keep my luxury slider at 10%, even with all 8 luxuries in my cities. This is because it gives me more flexibility under democracy as war weariness and unhapinnes sinks in much later. And for less developed cities, I don't have to waste population in entertainers. It is worth it for me and I think the AI should be able to keep theirs at 10% in peacetime as well.

 

[Roland Johansen]

You're right, the AI should be allowed to use the luxury slider in peacetime.

Computers are very good at counting. So the computer should just count the number of food/production/gold it wastes by using entertainers and compare that to the amount of gold it loses by using the slider. That's what I basically do. I just don't exactly count it, I make estimates. (And sometimes I of course make a different choise based on other motives, but that's not important in this discussion).

 

[skylined]

What do you think about the idea of having "public work shields" instead of workers as in CTP2 game which has no worker unit at all ??

 

[Roland Johansen]

I've played CTP too and I liked the public works idea. At first. Later I realised that there was a disadvantage too. The workers can't be killed anymore so a city can improve its surrounding terrain while enemy units are walking around the city. Also the possibility to cripple the enemy's terrain improving abilities by killing his workers is lost.
Next to that, some minor strategy deploying your workforce is lost because a city that is founded far away from anything you own can be improved just as easy if you use the public works idea from CTP. In CIV3 you have to bring some workers to that city and protect them with some units.

Maybe someone can think of a mix of the two systems incorporating the advantages of both systems (advantage of public works is of course that it is more easy in its use).