Tentative Project - Part Three

[Peter Triggs]

I think that Daniel and Colorme are to some extent talking at cross purposes about those settings in DiffDB. For what it's worth here's my take on them. It's easiest to start with a specific example so consider gold adjustment on impossible level, but restricted to the ancient age:

Taken from DiffDB (Impossible)
----------------------------------------------------
AI_MIN_BEHIND_PERCENT 1.8
AI_MAX_BEHIND_PERCENT 0.8
AI_MIN_AHEAD_PERCENT 2.0
AI_MAX_AHEAD_PERCENT 3.0
AI_MIN_BEHIND_GOLD_ADJUSTMENT 1.0
AI_MAX_BEHIND_GOLD_ADJUSTMENT 1.5
AI_MIN_AHEAD_GOLD_ADJUSTMENT 1.0
AI_MAX_AHEAD_GOLD_ADJUSTMENT 0.9
----------------------------------------------------

If you graph this data I believe you get:


--------------------------------------------------------------------------------


max behind adj 1.5 I x0
I *
I *
I *
min behind adj 1 I x 1 - - - - - - - - x2
=min ahead adj I *
I *
max ahead adj .9 I x3 - - - - - - - -
I
Gold Adjustment I
Ancient Age I __________________________________________________ _____ _______
.8 1 .8 2 3
Factory Settings max_behind min_behind min_ahead max_ahead



--------------------------------------------------------------------------------

The programmer's comment is not exactly translucent:


# note: scale values linearly between max-min when amount
# behind/ahead are between min-max


but I tend to aggree with Colorme that it means "Put the stars in where I just did". So Colorme is right in maintaining that in this example when the AI is less than 1.8 times as strong as the human it gets helped. Note though that when it is more than twice as strong as the human it gets penalized and bear in mind that this is Impossible level. On easier settings the above line is, roughly speaking, shifted upwards and to the left.

However, I believe that Daniel's original point concerned the situation in which the AI is treated as equal to the human: the horizontal line joining x1 and x2 above. It's true that in this example there's not a lot in it; but surely Daniel's point is valid: no civ-type AI can compete with a really determined human player. So consider his settings:


--------------------------------------------------------------------------------

max behind adj 1.5 I x
I *
I *
I *
min behind adj 1.1 I x
I *
I *
max ahead adj 1 I x - - - - - - - -
=min ahead adj I
Gold Adjustment I
Ancient Age I __________________________________________________ _
.8 1 .8 2 3
Daniel's max_behind min_behind min_ahead max_ahead

AI to Human ranking



--------------------------------------------------------------------------------

Here, unless the AI is more than twice as strong as the human it gets helped; and the further (relatively) "behind" it is the more help it gets. It's not treated as being equal to the human until it's at least twice as strong as the human. This is surely a better way of doing it and may go a little way to solving Harlan's problem:


--------------------------------------------------------------------------------

The problem with CTP1 was if you could survive long enough, the game eventually became a cakewalk. Whereas the game should be the other way: not too hard to survive initially, but the longer the game goes on, the greater the challenges.

--------------------------------------------------------------------------------


Of course you can argue that the existing settings do more or less the same thing but Daniel is right in maintaining that they don't do it enough. Clearly, there's a lot of room for experimentation here.

EDIT: BAH! Why can't I get this thing to re-produce graphs?

[colorme]

I played a game with the increased wages etc. and that's gotten rid of the huge cities problem completely.

This has been mentioned by other folks, but it does seem that the AI being at war with everyone in sight, seems to hurt its offensive capabilities. One sees huge spikes (mostly going south) in the AI power graph, probably because of all the fighting it does.
I've never had such spikes in my power graph, even when I think I've lost several big battles. So, clearly the AI must be losing huge armies because of infighting.

Does anyone know of a way to get the AI to be more friendly with each other, and less with the human?

[wheathin]

Strangely enough, this is exactly the opposite complaint that others have voiced about the AI not fighting amongst itself enough.

The real problem is that the AIs fight, but they don't crush.

Human players slowly gobble up their neighbors and expand. For the game to remain challenging through the mid-game, AI's have to do the same. It is the only way for an AI to keep up with the human player. One or two AIs have to wipe out the others and grow stronger - but it is this process that the AI is incapable of. The AI can fight all it wants; it just can't kill.

Quick survey: how often does one AI eliminate another in your games?

[OmniGod]

Well to weigh in on the does the AI destory other civs... in the umpteen million games... okay maybe only 25 or so I've had the Germans destroyed by either the Russians or Greeks about 5 times... that's only 20% but from what you guys are saying it never happens... it only happens if they start to run outta room or are pushed into a corner.. I find that they don't explore as much as they should... why fight your way out when they can expand to spots that are further away... example colonizing... they'll colonize only when the next island is within a turn or two travel from their home... why is that?

Omni

[WesW]

Diodorus, military support is taken out of your production, not your gold, so that is why you have noticed no difference in that category. Check your units tab, and notice what the support percentage is. It is usually around 20 to 30% in my games, and should not be more than 40 or so for the AIs.

Btw, it seems that my trip is going to be cut short, so I should be back home sometime late tomorrow. Then I can begin to experiment with some of my proposed changes.

[Chris B]

One Random Thought- Could A trigger be made to allow units to capture fortifications? This way yor natonal boundaries could look more solid, and U wouldn't have to build new ones after you conquered.

[Martock]

had another random thought bubble....how about making barbarians change units based upon the age of the game. i remember this happened in CIV II...seems a bit daft to me that warriors and hoplites are still hopping about in the diamond age...and still further in that thought bubble...how about creating barbarian STEALTH units...i had great fun creating a barbarian infector and watched as it went from city to city infecting it's merry little way, while the ever alert and always prepared AI (snicker snicker) did nothing to stop this little terror. i actually plopped down a fort and then a listening post next to each other, manned the for with 3 spies and 1 infantry man just to see what happened. yes the listening post has been tweaked to see out to 8 tiles and the whole shebang was done with the help of the cheat menu but that's besides the point! it would seem to me that as the game progressed, barbarians would change from "barbarians" to "anarchists" to "terrorists" and use appropiate units. i don't know if this could be done, but it'd make the game a lot cooler if in the diamond age i had a loose barbarian INFECTOR running about in my back yard the a barbarian hoplite.

is this something that can be done?!

[Ioannes]

This may be of the recent subject line but a simple and nice tweak to add to the difficulty of the game. Wes, would you consider changing the starting contentment level to 72 on the impossible level and 73 on the very hard level. This change does seem to add to the difficulty of the game when desired.

[WesW]

Below is a copy of a post I made in Harlan's happiness thread. Hmm..."Harlan's happiness thread." Does that sound like a Beatles album to anyone?

"I have been thinking about the new happiness system myself. Remember that happiness is directly related to crime, and that a few percentage points really add up over time. Not as much as the benefits of the lowered sliders though, I assume.
Gedrin sent me a copy of the his mod today (who knew he had his own mod?). Anyway, in looking through his govt stats, I noticed the flags for positive and negative coefs in the three slider areas (food, prod, and wages). They are all set to 1 right now. I am going to set them to 4 for positive and 6 for negative, and see how that works. This should mean that moving the slider one notch causes these changes to happiness, and that the resulting changes to crime should about equal the benefits or penalties incurred from the slider areas."

Martock, your Barbarian ideas seem good, and very funny as well. Look in risks.txt and strategies.txt and see what you can piece together. I think everything you mentioned can be done, if you want to take the time to learn how strategies.txt works.

On other topics, I have been working on the text files, and have learned a lot about how the various costs and benefits work together when figuring out gold and science.
I think that the new improvement cost system is going to work out very well, perhaps even better than I expected. As you know, cities get their wealth in the real world either from having a valuable commodity they can trade, or from being on a trade route, like an ocean port.
With the previous cost system, this wasn't shown as dramatically in the game as in the real world. If you had a large pop, you were going to have a pretty good income regardless, and at least decent gold and science production, since many improvements' benefits increased directly with size, but costs did not.
With the new system, you really have to think about where your city's strengths lie before deciding to build something. I think it will be possible to see large but poorly producing cities, ala the third world, with the new system.
Unlike wages and upkeep, which are taken out of commerce, the pop-based costs come out of gold only. I don't know if this is better or worse, but it does make things simpler, since you don't have to worry about balancing science improvements as well.

I have doubled the effect of gold improvements, to 40%, for the initial trials. This seemed to give a pretty good balance in my very limited testing so far.

Right now, I have the food/prod/wage sliders set to 12kg/8hr/5commerce at the neutral position. Each notch represents 3kg/2hr/2comm right now.

Finally, I have been getting 1 or 2 emails a day asking when the mod will be ready, or where people can get it.
All I can say is to keep up with the forum here. I don't know when I will get my new website, but once it's up, I should be ready to post the alpha version of the mod.
As to when a public verison will be ready, that is anybody's guess at this time. There is just no way to know what we will be able to do as far as slic is concerned, much less when things will be balanced and stable. There are many, many things I want to do with the standard text files as well.
We are still figuring out what things need to be balanced in the existing features, and uncovering all the new flags that will allow us to balance them, as well as do new things with the game.
The more I look into the texts, the more new things I see. These new factors make the game harder to balance, but at the same time give us more opportunities than ever to customize and improve it.
[This message has been edited by WesW (edited December 21, 2000).]

[Pintello]

Hi Wes,

Good to have you back. Just thought I would alert you to another tread that is going on in the General Forum. It is titled "My AI Analysis." I don't remember who started it, but is has some good ideas about improving the AI's attack ability. I thought you might want to look at it to see if there is anything there you haven't thought of or could use.

Regards,

Timothy Pintello

[Martock]

i'll take a look into those 2 files tonight but i don't expect to have any reports on them till at the earliest, dec 28. i'll be going home for the holidays and my folks don't have a pc worth it's weight in peanuts. once i'm back i'll more time to tinker around with this.

[Taliseian]

Wes,

I went to take a look at your site...and all it did was force me to www.zip2.com

How do I get to your page??


Taliseian


=====
The World of Lyrlusa
A Dungeons and Dragons Campaign World http://www.crosswinds.net/~lyrlusa
=====

[Martock]

i've been doing some thinking here about the barbarian stealth units. i've decided on the following:

1) i always loved the 'love bug' of doom from CTP1 but was flabbergasted at the sight of the guy on the blue bird of happiness that replaced my doom bug. so i've decided to switch the sprites around and use the love bug as it was and use the new bugette as a barbarian stealth unit. unlike it's bigger and badder brother, the barbarian unit will explode as a NUKE. i will have to see about the nanite defuser though and see if that screws it up. i don't want ABS to shot my little fella down so if i have to make the nuke attack like a plant nuke attack, i'll have to give it an 150% success rate.

2) i don't know if there were plans to revive the old spy sprite but if not, i'm thinking of using that as the middle age bad ass for the barbarians. only problem, what does it do?!! can't give it nukes or parks...hmmm...i wonder if it can be set up so that the anarchist can be used to target specific buildings or at least random ones. i'd definitely give him the ability to use a form of the plague though if we could tone it down a bit that might be more useful.

3) as for the barbarians using forces of the appropiate ages, i've decided that it would be best to limit it to just infantry and stealth units. i think it'd be odd to see a barbarian stealth bomber flying about. i'm gonna use some of the naval units however. i know that the barbarians tend to use them like pirates as i've tested that. i basically gave the barbarians 2 of each kind of ship upto battleships and it used them to pillage, plunder, and attack everything in site.

i've not done any sort of SLIC modding before save for changing some of the sounds in my game so this will be quite a learning experience for me. i do wish someone would put together a more realistic nuke explosion though...seems so puny and crappy the way it is now... if i had any talent for that sort of thing i'd have done that as well.

anyone want a realistic sounding nuke complete with air-raid siren in the background??

[WesW]

I have no website at the moment. Hopefully, that will change soon.

Thanks for the heads-up, Tim. I have made some changes like the ones Matte mentioned there. Some of the other things I have different ideas about.

Is the Spy sprite different from Ctp1? I thought it was the same. I will try and remember to insert the old sound files for it, if people like them better. I usually play with the sound off, so I can listen to the TV when I play, so the sounds are not a priority for me. Lev sent me the Ctp1 sound files and a modified text a while back. I will get to those once I have the alpha ready.

Mark, you need to give Martock the special designation "Klingon Barbarian from Hell". (Kind of like Caveman Lawyer.)
If you get the Barbarians to doing the things you talked about, send me the files, and I will try and work them into the mod. I have already changed some of their force-matching settings, and have them building garrison troops and naval units.

Also, before I forget, I need someone to take a look at the diplomacy files and see what they can do with them. This will be a big and complicated job, so be forewarned. From all the griping I am reading, this is an area that is full of promise, but sadly underdeveloped. This theme of unfulfilled promise is something that I keep finding as I go through the texts and read the forums.
The more I explore the files, the more I believe that the tools are there to make a game that will totally rock. The trouble with the original settings is that the play-testers simply didn't have enough time to play with all of them, or even half of them in certain sections. I wonder also who the play-testers were at Activision. From what the Activsioners told those of us on the A-team during beta-testing, I think that most of them were regular employees who do this for all their games, mostly for de-bugging purposes. If that is the case, then they probably had little experience with civ-games, since Activision doesn't make any other civ-type games.
Oh well, WE have the game now, and we know what to do with it.

I have spent today on the strategies.txt and the files which relate with it.
I have made many changes to the various force-matching lists, raising them so that the AI will not attack unless it is strong enough to win. This goes for field battles as well as cities. Basically, the AI's offensive strength needs to be 1.2 times as much as the defender's defense strength when attacking in the field, and needs to be 1.6 times as much when attacking cities. These settings may need to be raised.
I changed the Harass force-matching files to be those used in taking cities. The Offense settings deal with field battles.

The most important thing I did today dealt with the way the AI chooses which city improvements to build. I had asked in months past for the ability to have the AI pick improvements to either play to the city's strength, or shore up a weakness, depending on the type of improvement. This was something we were unable to do in Ctp1, and something which is sorely lacking in all civ-type games. Well, I don't know if the programmers were listening to me or not, but this ability is present in Ctp2, and I think it is going to result in dramatic improvements to AI performance once the settings are set properly.
This is one of the instances I alluded to earlier, where you have all this potential that isn't being used properly. I say that because the settings initially present had the AIs doing the same things that they were in Ctp1. Specifically, it had them building science and gold improvements based upon factors other than the city's commerce income, building production improvements in its worst-producing cities, which doesn't make sense when you think about it, and a few other things that didn't make sense as well. I think that people are going to see major improvements from simply having these flags used properly. I also went through the building element lists to make sure they were appropriate.

All in all, it's been a good day. This game has so much promise, it just makes me drool as I continue to explore it. It is just too bad the programmers didn't have more time to develop all the game's tools.

Btw, Gedrin has created a monster Excel spreadsheet for his mod in which he has set up most of the major default/gamedata files so that you can make changes to the sreadsheet DB, and it automatically changes the text files sheet accordingly. Then you can save the text sheet as the game's text file. This is going to have huge improvements to altering such things as units, advances, terrains, governements, etc.
As some of you may know, advances are the foundation upon which everything else is built upon in game. Well I *think* that Gedrin's spreadsheet will allow you to decide when you want a new advance to be discovered, and what things are linked to it and enabled by it, and the spreadsheet will calculate how much the advance should cost and make the appropriate changes to all the things that are linked to it. Amazing.
Gedrin is still working on the spreadsheet, and it is about 1meg zipped, so please wait until I post it on my website before asking for a copy from me. I only have a 26k connection for the next couple of weeks.

Diodorus, let me know sometime how your advances work is coming, either in this thread or in one of your own if you think that is better. Since the new improvement cost system comes out of gold rather than commerce, you can use the standard cost system as a basis for your new advances.

Wouter (Locutus) emailed me today, and said that he would have a militia trigger sent to me by tomorrow, when he goes on vacation. It may not work perfectly, but should be good enough for the alpha version. I just need to get an idea of how it affects general game flow. For those of you who didn't use the Ctp1 Med mod, I think you are really going to like this feature.

I have done about all that I think I can to the growth and development side of the game prior to sending it out for play-testing. How do you guys feel about the current government settings, aside from Fascism being in the wrong place? I plan to add Celestial Dawn's Fundamentalism gov to the game, and my Constitutional Monarchy. I noticed that many of the govs have different settings than they had in Ctp1. What seems off, and what do you think should be done?

[Alpha Wolf]

arrrrrrrrrrrrrrrrggggggggggggggg. I spent most of last night and tonight putting together a spreadsheet with all my changes. Even setup an access database with all the unit flags, and had just started loading it. Well, that'll save me a weekend of typing if someone is already doing it.

------------------
History is written by the victor.