Tins

[Koyaanisqatsi]

quote: Originally posted by JosefGiven on 01-17-2001 02:28 PM To sum up, in order to survive with a place in our hearts, Civ III HAS to have a base in our human history, or at least feel like it does. With no historical basis, and with the complexity stripped from it...Sid and the Firaxis team will create a monster...

Oh, certainly. My initial point was simply that there seemed to be a lot of the attitude that historical accuracy automatically will improve the game, no matter what it takes to make things accurate. This is a game, not a simulator. Accuracy is good, but I buy games because they're fun, not because the musketeer has the right shape of hat.

[Ralf]

Good arguments, Bell! Interesting posts.

If you havent already, check out below mammoth-thread, that wined up to evolve around the exact same simplicity/complexity- and gameplay/reality-discussion as in this thread.

Click About health and protective border tolls

Then im writing this, its late. I try to comment your post better tomorrow or later. Comments (if any) about above link you can post here.

[This message has been edited by Ralf (edited January 17, 2001).]

[JosefGiven]

Ah! I think I'm starting to see your point! The fundamental difference between vertical and horizontal complexity...'good' and 'bad' complexity.

OK, to follow on from this, I would suggest that a few more horizontal branches at the base of the Civ III model would allow scope for more vertical level challenges... Can I think of an example? Hmmm...

Lets try the resource management concept.

If it's not too presumtuous to attempt to guess your thinking, it seems as if you see the idea of having to gather resources AS WELL as having to sort out production output as pure unnecessary horizontal complexity. Fair play, perfectly understandable.

But, how about this? If you make it very easy for players to gather resources, ie: Have special resource squares la Civ 2, but rather than yield trade/production/food bonuses, they produce these here contentious resources. Then, instead of having to source these resources about your civ via caravans, or whatever, they go to a 'national' pool, whereby all the cities in your civ can access them. Not only that, but any resources that go unused in your turn go to a national stockpile for use in future turns, and trade with foreign civs.

Right, that I think limits the horizonal complexities required to support the resource-gathering/management as much as possible. Now let us explore the possible vertical complexity benefits:

Vastly improved, and dare I say it, more realistic, inter-civ trading possibilities.

The possibility of comedy trade back-stabs!

Piracy in the CtP vein

Well, you know what I mean. There are loads of possibilities in there, as many other people on these forums have suggested in better ways than I could.



------------------
Josef Given
josefgiven@hotmail.com

[Koyaanisqatsi]

quote: Originally posted by Ralf on 01-17-2001 06:33 PM If you havent already, check out below mammoth-thread, that wined up to evolve around the exact same simplicity/complexity- and gameplay/reality-discussion as in this thread. Click About health and protective border tolls

I can't read the whole thing because it's too long, but I did skim it. A couple of points:

Scenarios--I've played some scenarios, but there's really only two that I enjoyed as much as the original game: East Wind, Rain, which was amazing, and Avalon, which I liked primarily because I made it based on a book I read.

Why don't I like scenarios? Well, for the same reason I don't like mayors, actually. Being the selfish little egomaniac that I am, when I play Civ-esque games, my empire has to be mine. I don't want the AI making decisions for me, and I don't want other people placing cities for me, historically accurate or not. A big part of the fun of Civ is growth and expansion, of which most scenarios don't really provide much.

AI--You know why a lot of suggestions would prove so difficult to program an effective AI for? Because there's so much horizontal complexity to them. AI's are very bad at horizontal complexity. Too many options that all do the same thing, and no real way to assign a priority to them...you end up having to program either very specific instructions, which makes the AI predictable, or program it to choose randomly, which makes it ineffective. Vertical complexity, however, is much easier to program for, since it involves lots of sequential actions that can be easily prioritized.

Um...I had another comment about something in that thread, but I forgot it now.....oops.

[Koyaanisqatsi]

quote: Originally posted by JosefGiven on 01-17-2001 06:40 PM OK, to follow on from this, I would suggest that a few more horizontal branches at the base of the Civ III model would allow scope for more vertical level challenges...

True, but there are limits. If you increase the base horizontal complexity too much, then the whole game gets unwieldy. Admittedly, there's room to add horizontal complexity to the CivII model without going too far...you just can't add huge amounts of it.

quote: But, how about this? If you make it very easy for players to gather resources, ie: Have special resource squares la Civ 2, but rather than yield trade/production/food bonuses, they produce these here contentious resources. Then, instead of having to source these resources about your civ via caravans, or whatever, they go to a 'national' pool, whereby all the cities in your civ can access them. Not only that, but any resources that go unused in your turn go to a national stockpile for use in future turns, and trade with foreign civs.

Yep, that sounds much better. I still have problems with that specific idea, but you applied what I was talking about as far as complexity goes.

By the way, my issue with that idea is still a complexity problem, but a different kind. Unit production is a very, very basic function in Civ. The only thing more basic is resource harvesting, which is the base of all power in CivII. (As an aside, that's why ICS is such a sticky problem to solve...power comes from resources, and more cities mean more resources. Any change to that has to be very precisely balanced, or the game is shot. I'll believe they have ICS fixed when I see it.) Anyways, as I was saying before I was so rudely interrupted...I don't think that there should be artifical limits on unit production because it's such a basic function. It has a lot of potential to add strategy with fights over resources, but it would be too difficult to balance. Imagine if you started in an area with no oil, and were never able to break out of there. What chance do you think you would have for winning?

[Sarxis]

The more complex, the better. So long as it is not superfluous.

[Brent]

Okay, maybe Civ shouldn't be like Sim City, but do you think some other game might fill that role?

[Koyaanisqatsi]

quote: Originally posted by Brent on 01-21-2001 02:11 AM Okay, maybe Civ shouldn't be like Sim City, but do you think some other game might fill that role?

SimCity 4000?

[Ralf]

quote: Originally posted by Bell on 01-18-2001 07:08 PM I don't want the AI making decisions for me, and I don't want other people placing cities for me, historically accurate or not.

If above is a comment on the pre-designated AI-city placement idea, i must correct you, by emhasizing that this idea normally only applies to AI-civs alone - NOT to the human player. Once civers give themselves time to really read through and understand this idea, its not hard to see what really great potentials it has.

quote: A big part of the fun of Civ is growth and expansion, of which most scenarios don't really provide much.

Well, partly i tend to agree. But, scenarios shouldnt be just about rewamped rules/graphics in advanced mid- and late-game setups. I miss the ability to play standard early-game setups scenarios, with only 2-3 starting cities for each Civ. Thanks to the pre-designated AI-city placement idea, any scenario-designer can then tailor-cut exactly where and how the AI-civs should place their cities and expand their empires. This, together with template-guiding city-area tile-improvement and template CI/unit building priorities pretty much ensure stronger AI-civ competition. The human player can of course choose to place his cities, and expand his empire anyway he likes to. The player has complete freedom.

The whole issue above, is about taking out that stupid AI-settler out from the equation; both then it comes to placing AI-cities and developing AI-city areas. One only has to look at mayor-controled settlers/terraformers to understand how weak artificially intelligent unit-pathfinding really is - at least on huge totally random maps, like the ones in Civ-games.

quote: you end up having to program either very specific instructions, which makes the AI predictable, or program it to choose randomly, which makes it ineffective.

I keep my fingers crossed for a "spoonfeed-able" AI, there text-file tweaking players, can guide exactly what different AI-civs should emphasize, then it comes to overal principal strategic/logistic choices.

About the "predictable-problem":

Firstly, I dont think we should exaggerate how un-predictable human multiplayer often playes Civ-games. We all have our favourite (= predictable) strategies that we tend to stick to again and again.
Secondly; there is no reason why we couldnt have several parallell text-files with civ-specific instructions, from which the AI can choose randomly between with each new fresh game.

quote: Vertical complexity, however, is much easier to program for, since it involves lots of sequential actions that can be easily prioritized.

Yes, but before any Civ AI-programmer can prioritize, he must first gather ALL feedback-info each turn. This is very simple in games like tic-tac-toe, and somewhat harder in games like Chess.
But, the bigger the board/map/world is, and the more possible variables/combinations there are, and the more fuzzy-logic rules there are in the game, the harder it becomes to produce any half-decent AI-restistance. Check out http://apolyton.net/forums/Forum6/HTML/001531.html

This is why it is so very important to come up with ideas that can "spoonfeed" AI-civs in how to achieve better restistance - whether, this help deals with emphasize-guiding self-maturing AI-city areas, AI-city- and AI-unit improvement editable paths, editable AI-tech-tree paths, pre-designated invisible AI-city placements (which nevertheless starts and expands either unpredictably from any pre-designated point, in the main-game, or perhaps according to editable scripts, in scenarios), Strategic diplomacy thumb-rules, and much more.

[This message has been edited by Ralf (edited January 21, 2001).]

[Koyaanisqatsi]

quote: If above is a comment on the pre-designated AI-city placement idea

It wasn't really. All I meant was that I didn't like scenarios because they usually started with most of my cities already built.

As far the link you mentioned, I can't say I like that idea of scrapping the map generator. Yes, the net allows you to download all kinds of handmade maps, but it's better to have a system built in to the game that gives you infinite maps. It would be nice for scenarios and custom maps though if the hooks were in place to give the AI sites to build locations though. On random maps, the AI's would just be on their own.

quote: Firstly, I dont think we should exaggerate how un-predictable human multiplayer often playes Civ-games. We all have our favourite (= predictable) strategies that we tend to stick to again and again.

Oh, I think human players are more predictable than the AI's. Doesn't mean the AI's can't get even less prediactable.

quote: Secondly; there is no reason why we couldnt have several parallell text-files with civ-specific instructions, from which the AI can choose randomly between with each new fresh game.

Works for me.

quote: But, the bigger the board/map/world is, and the more possible variables/combinations there are, and the more fuzzy-logic rules there are in the game, the harder it becomes to produce any half-decent AI-restistance

This is exactly what I mean. Vertical complexity reduces variables and provides more structure for an AI to use. Horizontal complexity adds lots of variables, with no real way for an AI to choose between them.

[Ralf]

quote: Originally posted by Bell on 01-21-2001 11:37 AM As far the link you mentioned, I can't say I like that idea of scrapping the map generator. Yes, the net allows you to download all kinds of handmade maps, but it's better to have a system built in to the game that gives you infinite maps.

A short note: I quickly abandoned the idea of scrapping the map-generator further down the thread, because it really didnt collided with my main "invisible pre-designated AI-city placement" idea. My very last post in that thread is a short summarization of the whole idea, by the way.

quote: On random maps, the AI's would just be on their own.

IF Firaxis can implement the idea together with the automatic map-generator, and IF the AI-cities, as a result of this, is placed noticeably more efficient; then they should of course implement it.

After all; its the practical end-result that should have the final verdict!

[This message has been edited by Ralf (edited January 21, 2001).]

[Brent]

Would SC4000 be a history game?

[Koyaanisqatsi]

quote: Originally posted by Ralf on 01-21-2001 02:56 PM IF Firaxis can implement the idea together with the automatic map-generator, and IF the AI-cities, as a result of this, is placed noticeably more efficient; then they should of course implement it.

If you can program the map generator to find ideal city sites, why can't you give the same algorithm to the AI? It would be the same effect, but it would give the AI the chance to react to game conditions.

[Koyaanisqatsi]

quote: Originally posted by Brent on 01-21-2001 02:56 PM Would SC4000 be a history game?

*shrug* I'll tell you when it comes out.

[Ralf]

quote: Originally posted by Bell on 01-21-2001 03:43 PM If you can program the map generator to find ideal city sites, why can't you give the same algorithm to the AI? It would be the same effect, but it would give the AI the chance to react to game conditions.

Its much easier to calculate AI-city placements from top-left to bottom-right once-and-for-all, directly after the map is fresly generated, then it is to - in the middle of a game; navigate hordes of AI-settler, and then - in the midst of that short game-turn, with so many other duties thats calling for AI-attention - try to calculate best AI-city placements.

Heres an example how map-generated AI-city placements could be done - at least in theory:

"Infact, once the continents and the terrains of the map is generated, the map-generator then could continue in calculating potential 21-square AI "city-places" (or rather "city-areas") locations from up-left to down-right, shoulder-to-shoulder over the whole map. Only those areas who had its central square on terra-firma would be account for.
The map-generator should be able to calculate average values from from all 21 tiles in each AI city-area. The lower food/resource sustain threshold-values for such AI-city generated places should of course be manipulatable by the human player - giving him some control over what the map-generator AI should accept, or not.

Above by itself however, would give rather bad results because small 6-10 square island/capes would often be left out, and also not less important: the general AI-city distribution would be relatively squarish and unimaginative (= plain ugly). The shoulder-to-shoulder generation of AI city-areas should therefore be alloved to overlap, disjoin, disfigure each others areas max 1 square-row (the outer 12 squares - the 9 inner ones are untouchables). in random ways.

Also, (important!) this first map AI city-place sweep is followed by a several additional sweeps that distributes all the potential AI city-places in an alltogether different fasions. If these new sweeps finds any potential city-placements that has a better food/resource sustaine-value then the previously designated ones, 1-3 squares away - the new one automatically overrides any old weaker AI city-placements.
Finally, any second and above sweep coustal city-places should most often override any previous sweep 1-3 squares-away inland city-places. The reason for the latter is that the terrain on small/medium-sized island can be utilized much better, if coustal city-places have high priority."

----------------------------- added comment:

You can compare above with a map punctured with unlit red diodes all over it (= invisible for the human player). Under the actual game these AI-city "red diodes" is turned on (= habited; founded AI-city), one after the other.

[This message has been edited by Ralf (edited January 22, 2001).]