Announcement

**Abdul Alhazred** · December 20, 2004, 15:44

Read

**Tamerlin** · December 20, 2004, 20:10

Good comments Maquiladora (at least I think so). Though working on these aspects would not help solving one of the main problems of the AI (the military part of it) it could largely improve the way the AI is managing its empires.

**Martin Gühmann** · December 22, 2004, 17:22

Originally posted by child of thor in Playtest II
I've been trying out a few things this morning.

Playtest build: 14/09/2004 - with the new Apolyton title screen(the brown one).

Well this build isn't the one provided by calvitix and it doesn't contain his changes that turns the AI into a naval moron with the need to calculate everything thousands of times without any results.

-Martin

**Lou Wigman** · December 26, 2004, 19:50

Can we have a new "build" without the experimental naval AI?

I'd like to playtest but the 20/10 build is not satisfactory.

**Ekmek** · December 27, 2004, 02:29

A new build would be nice, especially if it has increased bombard range in it

**Maquiladora** · December 27, 2004, 03:27

Originally posted by Tamerlin
Though working on these aspects would not help solving one of the main problems of the AI (the military part of it) ...

Well AI movement is not my problem to fix

Another thing that might help the AI: taking riots completely out of the game. Since perhaps what cripples the AI the most (in terms of growth and development) is managing happiness. Instead, if there were no riots and much stronger crime (related to happiness levels as it is now) that would encourage you to increase happiness and/or lower crime. No riots would also create a much more flexible city limit because you can keep building over the limit but you create more unhappiness and crime.

I know riots are a standard feature for most civ games and it might be hard to imagine a civ game without them, but what do we get in gameplay terms right now with riots? Boring micromanagement of entertainers/empire sliders to make sure every city has at least 73 happiness, and we have to check every turn because cities grow and create more pollution and more unhappiness. How can a city even riot at 72 happiness but be eternally content at 73?

Well its an idea.

A new playtest would be great too, lots of new things to play with.

**Martin Gühmann** · December 27, 2004, 12:45

Originally posted by Lou Wigman
Can we have a new "build" without the experimental naval AI?

Well I can do the first part of your request, but the second part needs some serious time. And unfortunatly I have already spent to much much of my Chrismas holidays for vision. Now I have to do some other stuff. The problem is that I have to remove every single change of calvitix which I suspect to be the culprit and have to test it. And that needs some time. And by the way I don't thing that is experimental naval AI, I think this is rather experimental AI that damaged the naval aspect of the game. The slowdown is in since calvitix second pack, he should have seen it if he uses maps with 3-4 AIs per continent.

Originally posted by Lou Wigman
I'd like to playtest but the 20/10 build is not satisfactory.

I agree but without a automatic change handling system it is a task as said to figure out the problem.

Originally posted by Maquiladora
Another thing that might help the AI: taking riots completely out of the game. Since perhaps what cripples the AI the most (in terms of growth and development) is managing happiness. Instead, if there were no riots and much stronger crime (related to happiness levels as it is now) that would encourage you to increase happiness and/or lower crime. No riots would also create a much more flexible city limit because you can keep building over the limit but you create more unhappiness and crime.

I know riots are a standard feature for most civ games and it might be hard to imagine a civ game without them, but what do we get in gameplay terms right now with riots? Boring micromanagement of entertainers/empire sliders to make sure every city has at least 73 happiness, and we have to check every turn because cities grow and create more pollution and more unhappiness. How can a city even riot at 72 happiness but be eternally content at 73?

Oh yeah, removing riots its like fixing vision by just forcing the game into good mode, so the map and status bar doesn't show anymore anything you shouldn't see, because then they should show everything. Instead of one week of work I just would have needed a few seconds to do the fix. What a fix concerning riots, if the AI cannot handle it just remove it. Why not remove the rest of the game or better why not start playing civ3.

-Martin

**Maquiladora** · December 27, 2004, 17:04

Originally posted by Martin Gühmann
Oh yeah, removing riots its like fixing vision by just forcing the game into good mode, so the map and status bar doesn't show anymore anything you shouldn't see, because then they should show everything. Instead of one week of work I just would have needed a few seconds to do the fix. What a fix concerning riots, if the AI cannot handle it just remove it. Why not remove the rest of the game or better why not start playing civ3.

So what is your idea to help the AI avoid riots? and make avoiding riots less tedious?

As I said Martin, its not just removing it to improve the AI, its removing a horrible part of gameplay for the human too. Whats so fun about micromanaging entertainer specialists for pollution every turn? Taking out riots adds to gameplay (but not tedium), since you arent so tightly restricted to 'good' and 'bad' happiness, you have different levels of crime and you deal with it in your own time.

But with riots as soon as you hit 72 happiness and risk revolt, your city stops doing anything? Yeah right. If you say theres only a small chance of rioting well theres still a chance, and that makes risking any chance of a riot pointless.

Which makes having riots in the game pointless, since the human knows how damaging riots are and can play a whole game without their cities rioting once. So it becomes rather an 'avoidance of feature' for the human but the AI doesnt know how to play close to the line on happiness, so the AI gets hurt badly.

So do you spend time teaching the AI to use a feature that makes tedious gameplay and is avoided anyway and the human will always be better at managing? Or do you take out the feature and replace it with a more interesting crime/corruption based system that the AI can use well straight away and isnt instantly punished with riots for?

Besides you wont find many people more against civ3 here than me, just see the many 'PW vs worker' threads in civ3 general. My suggestion for removing the info on the status bar and right-click is not like this at all, the status bar problem has been mentioned by other people before. And also I was making only a suggestion, but perhaps we shouldnt suggest anything.

**Martin Gühmann** · December 27, 2004, 18:00

The problem is that it is very unrealistic that you can threat a city as bad as you like. You can torture and exploit your people and maybe you have a very good security so that you can controll the situation. But in fact in history the situation always exploded at some time or caused other serious problems.

Removing riots means also removing fall offs, that's also a very nice feature if you cannot manage your city and then suddenly it leaves your empire. At least without riots it would be suddenly. Of course entertainer and slider optimization could be done automaticly not only for the AI but also for the human player. Basicly this should be done when mayors are on. Of course this requires also some work to produce some code. This doesn't envolve only the entertainers but other specialists, too. And so far the AI has no idea how to use them. On the other hand more subtile problems like crime and corruption must be handled as well. That would be the next problems that hit the AI. And finally I am not so sure whether a happiness level of 73 in the default game is optimal at least I saw some drawbacks in Cradle were a happiness near the limit caused reduced science output. Our problem is that we consider a happiness level of 73 as optimal and maybe we aren't so wrong, actual we can move all three sliders to their extremes without any serious problems, I think this is not intended as you have some problems to do it in CTP1, at least the amount of entertainers needed is bigger then the gain.

-Martin

**Maquiladora** · December 27, 2004, 18:45

Originally posted by Martin Gühmann
The problem is that it is very unrealistic that you can threat a city as bad as you like. You can torture and exploit your people and maybe you have a very good security so that you can controll the situation. But in fact in history the situation always exploded at some time or caused other serious problems.

Well you can do this with the current model. You can have martial law, all the happiness buildings and entertainers, if you have a revolt then you most likely couldnt be bothered to micromanage your happiness, which is boring to do. So youre forcing the player to micromanage even more with the threat of riots. Riots only exist as a predictable and totally controllable threat, so whats the point in them existing at all if theyre totally predictable? Of course you can have unpredictable riots then, but this is even worse because the player doesnt have control.

But with increased crime and corruption you still get the problems of low happiness but you dont have to religeously micromanage every city to a precise happiness level, you give the player the tools to fix a highly corrupt city, but its not a headache to micromanage entertainers.

Removing riots means also removing fall offs, that's also a very nice feature if you cannot manage your city and then suddenly it leaves your empire. At least without riots it would be suddenly.

Well riots arent sudden, because we know what causes them instantly to riot and we know how to stop them forever.

But thats right the revolt has to be caused in some other way then. You could have many other ways, such as changing governments some of your cities revolt, of course this is extreme but its just the point that it doesnt have to be only tied to happiness. Also war discontent could cause riots or revolts, but not directly tied to happiness, because this is where the micromanagement of entertainers kills.

Of course entertainer and slider optimization could be done automaticly not only for the AI but also for the human player. Basicly this should be done when mayors are on. Of course this requires also some work to produce some code. This doesn't envolve only the entertainers but other specialists, too. And so far the AI has no idea how to use them.

Of course the AI will be able manage entertainers properly eventually, as well as other specialists, but if everyone automatically avoids riots then whats the point of riots at all?

And finally I am not so sure whether a happiness level of 73 in the default game is optimal at least I saw some drawbacks in Cradle were a happiness near the limit caused reduced science output. Our problem is that we consider a happiness level of 73 as optimal and maybe we aren't so wrong, actual we can move all three sliders to their extremes without any serious problems, I think this is not intended as you have some problems to do it in CTP1, at least the amount of entertainers needed is bigger then the gain.

Yeah definitely the slider elements need to force the player to make tougher choices. Particularly make the rations slider most expensive to maximize because its the most long term benefit, so it should constantly be a decision whether to move it for other benefits. Like literally a 'guns or butter' choice.

**Martin Gühmann** · December 27, 2004, 20:33

Originally posted by Maquiladora
But with increased crime and corruption you still get the problems of low happiness but you dont have to religeously micromanage every city to a precise happiness level, you give the player the tools to fix a highly corrupt city, but its not a headache to micromanage entertainers.

Well you prefere to have a soft limit, I think they try to achieve it by given a chance of less then 100% of rioting if the happiness is under 73. However there must be a limit to allow riots and revolts. The only problem is that it is here too obvious. But in the end it is always programmed so it is determined. However I don't look so closely each turn on the happiness level. I wait until some cities riots and then fix all the other as well.

Originally posted by Maquiladora
Well riots arent sudden, because we know what causes them instantly to riot and we know how to stop them forever.

But thats right the revolt has to be caused in some other way then. You could have many other ways, such as changing governments some of your cities revolt, of course this is extreme but its just the point that it doesnt have to be only tied to happiness. Also war discontent could cause riots or revolts, but not directly tied to happiness, because this is where the micromanagement of entertainers kills.

Well I also would like to see other ways for riots as well, I think the only idea I don't like about it to get rid of them at all. So the only solution I see is that the costs must be so high to be completely free of riots that it seriously damages you, but after a certain limit of happiness you should be have a very slow chance, but it shouldn't be so low that you don't notice it.

OK let's come to the other part of this posts, I compiled a new playtest build based on the all pack uploaded, yesterday. As I already have said I have no time to fix the most serious bugs. Therefore those two serious problems remain:

- A massive slow down of gamee on maps with AIs on more then one continent.
- No AI naval activity

Unfortunatly the first problem makes the game unplayable on maps with more then one continent. But I think this is better then to release nothing.

Here is the change log since the last playtest version from the readme file:

2004-12-28
Added: An attemp to fix range bombardment, but it does not seems to work.
Fixed: Vision
- The staus bar only shows what it should show. (Terrain status of the last visit)
- Help tile window shows terrain statisics and improvements from the last visit.
- The map only shows terrain and improvements and owners and city radii from the
last visit.
- If fog of war is toggled of status bar, help tile window and map show everything.
Changed: The algorithm used in SP games is now the same as that used in MP games.
Fixed: Feat accomplishment propagation from server to client.
Fixed: Propagate city defensive bonus.
Fixed: Mouse and keyboard events are also handled after buffer overflow.
Fixed: Some crashes concerning the Alexander scenario.
Added: Localized strings for the new civs to the other languages.
Fixed: Arena and Aqueduct prerequisites
Restored: Compatibility for Diplomod.
Added: A Quit Button at the multiplayer screen for hotseat games.
Fixed: Memory leaks concerning vision and the resourcemap.
Improved: The music screen unconfirmed user changes are not reseted when returning to
the music screen after visiting the music track selection screen and removed memory
leaks.
Fix: Unblock the client user interface when the server has handled the city production.
Restored: Database compatibility of constDB for MP testing.
Untested fix: Make cities grow and make science handled well in PBEM.
Updated: Some player slic built-ins so that they can be not only used in string
replacement but also in optaining integer values.
Repaired: Some invalid links in the English Great Libary of the Alexander scneraio.
Updated: GAMEID to match the value of the new anet.inf
Fixed: Some crashes in concerning diplomat data and the sp new tribe screen.
Fixed: Movement point propagation for settlers from disbanded cities in network games
with 3+ players
Corrected: Return type of IsVeteran function.

And here is the link to the new playtest build:

Edit: Link removed, there is allready a never version of the playtest.

-Martin

**Maquiladora** · December 27, 2004, 20:50

Well I also would like to see other ways for riots as well, I think the only idea I don't like about it to get rid of them at all. So the only solution I see is that the costs must be so high to be completely free of riots that it seriously damages you, but after a certain limit of happiness you should be have a very slow chance, but it shouldn't be so low that you don't notice it.

No I dont want to get rid of riots/revolts either, but dont have it directly decided by happiness. I think War Discontent, (severe) Pollution and radical (Republic to Fascism) Government changes are good options, with these kinds of effects you have to take more realistic action to prevent a city from rioting/revolting, so rather than building a shrine to stop it rioting (which is a bit silly) you bring some units home or lower your pollution etc. This at least eliminates the problem of having to micromanage entertainers too much, but you still have crime related to happiness, so there is still a reason to increase 'general' happiness.

Thanks for compiling the playtest

**mapfi** · December 28, 2004, 06:49

Just pitching in to support Maq's proposition - I also believe happiness/riots needs rethinking. I wouldn't go that far though, rather abolish the barrier 73 and let slowly increase the chance of a riot from 80 or so downwards. Then similary have chance of returning to peace from rioting, so e.g. a 78-city that riots is likely to be peaceful again next turn.

**Martin Gühmann** · December 28, 2004, 12:47

Originally posted by Maquiladora
No I dont want to get rid of riots/revolts either, but dont have it directly decided by happiness.

Obviously this was the misunderstanding.

In fact it is true that it is very easy to prevent a city from rioting revolting. It depends all on factors you can influence, maybe some extra factors that are happiness independent should influence, like conquest distress should increase the riot change independently from happiness.

Originally posted by Maquiladora
I think War Discontent, (severe) Pollution and radical (Republic to Fascism) Government changes are good options, with these kinds of effects you have to take more realistic action to prevent a city from rioting/revolting, so rather than building a shrine to stop it rioting (which is a bit silly) you bring some units home or lower your pollution etc. This at least eliminates the problem of having to micromanage entertainers too much, but you still have crime related to happiness, so there is still a reason to increase 'general' happiness.

However low happiness should affect the efficence of a city, a city with low happiness should have a low output of science production and food. This means of course a city can starve on low happiness.

One think to lower unhappiness through pollution is engaging entertainers as this reduces the number of workers and thus less pollution is caused.

-Martin

Announcement

PROJECT: Playtest Thread III

PROJECT: Playtest Thread III

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment