A simple, elegant way of making extra resources domestically useful

[Roman]

Fosse, I suppose you are right that we should not be 'defeatist' at this time. Perhaps we should strive to provide at least two ideas for each category of gameplay - one radical idea that favours realism and requires substantial modification of the current system and one more conservative idea that builds on the current system yet does add something to both realism and gameplay. It is much more likely that Firaxis will choose the conservative idea in most cases (or an idea of their own, or not modify the system at all...), but in some cases they might be inspired to choose the more complex system.

[Fosse]

Regarding AI not getting "deep." You're absolutely right.

However, I am a firm believer in having the program simply telling the AI what to do with good solid heuristics.

For example: People will say that the AI will be bad at organizing units into armies, or keeping stockpiles of resources, etc. I say, create a database for the AI of 50 to 100 unit-type ratios for armies, and a set of instructions about when it is good to stockpile and when it isn't.

The AI will never be as good as we are, but we can at least teach it to use some tried and true approaches.

For instance... nobody expects an AI to be able to "think up" clever flanking attacks. But we all get upset when the programmers don't provide code that simply says, "Attack on two fronts, with considerable force!" (in computer terms, of course ) Such simple code (which seems to have been added to C3C, though I have yet to play it) gets rid of the endless trickle of units that the AI "figures out" on its own.

I agree that it's a poor idea to have crippled AI because a game is too tough for the programmed AI. But I think that the mistake lies in the poor AI, and not the good game model.

Roman... Totally. If we can get one or two great new ideas into the game then we'll have a great victory. But if we don't provide dozens and dozens of such ideas, then we might not hit on ones Firaxis likes...

[Frozzy]

How about you need, say, 1 oil square per 10 tanks, otherwise their HP goes down.

[Fosse]

poll

A new poll takes the question of how to handle resources to task.

Hopefully it will help us guide our discussion.

[MrBaggins]

The Civ3 designers saw the possibility for castastrophy with strategic resources, and fudged that by giving the AI knowledge of these resources beforehand... if they didn't then the AI would most likely not see the later ones, the human being predisposed in planning to deny them.

The situation we are talking about... "Extra strategic resources" giving significant production bonuses isn't a heuristic problem, except at the grossest level. The solutions aren't hard and fast... since there are competing solutions. You might go with a genetic algorithm over the top of heuristics, and then develop that to a point where it was smart... but thats not been done in an empire game... (except maybe Evo, but thats a bit of a backwater,) or use a fuzzy system, but again, implementation is on the tough side, and you still have to define the operators appropriately, which means you need to understand how the system works beforehand, then defuzzify.

The Civ3 team have done very well, in general on the AI, but its proficiency is largely its due to a brute force method of resources and strategems... like tons of bonus resources, rabid expansion, forethought and lack of ZOC's.

One of the biggest problems is that "point defence" isn't easy to implement, or actually, just that humans are so much better at "point attack".

[Kuciwalker]

Originally posted by Fosse
Jon and skywalker, you'd both rather stick to the ON/OFF resource model of Civ 3, where one source of incense can make your 1000 people in 1000 cities happy, but not two people in one city?

Yes.

[Kuciwalker]

Originally posted by Roman
Thanks for the voice of confidence.

We can do something similar for luxuries. Again, having more than one luxury square of the same type currently yeilds no benefit beyond the ability to trade. I suppose we could have a single luxury square only be sufficient for a certain limited number of people, but this creates the same complexity problems as having a single resource only be sufficient for a given number of cities. A simpler approach would be to have each extra luxury square plus one, provide the same benefit in happinness as the first square.

e.g.

1) 1 luxury would increase happinness by 1.
2) 2 additional luxuries (of the same type) would increase happinnes by an additional 1.
3) 3 additional luxuries (of the same type) would increase happinnes by an additional 1.
4) Etc. (There could but would not have to be a limit.)

This is a simple system that ensures the benefits of additional luxuries of the same type apart from trade yet does provide an incentive to try to gain different types of luxuries and indeed to trade your luxuries for different ones.

This system provides no incentive to trade. The strength of the C3 luxury resource system is that it REQUIRES that you trade for different types.

[Roman]

Skywalker, it gives a very big incentive to trade. You will get a much higher bonus to hapiness by having 6 different types of luxuries (+6 bonus) than 6 tiles of the same luxury (+3 bonus). It therefore makes sense to trade your identical luxuries for different ones - the system merely ensures that the extra identical luxuries are not completely useless domestically.

[Kuciwalker]

Oh, I see what you meant. Sorry, I though it meant 1 lux = 1 happy, 2 lux = 2 happy, etc.