A new resource system

[Shyrramar]

I think Judgement's idea was a reply to this: you could build more tanks than you got oil, but their maintenance cost would be greater. They should not be weaker, though. That would needlessly create two types of all resource-dependent units: a weak and a normal one. In any case, it MIGHT work with tanks, but hardly with, say, swordsmen. What would you need a weak swordsman for? And it is an odd idea, that swordsman with wooden swords...

 

[Shyrramar]

I made some calculations concerning the upkeep of excess units. By following judgement's simple logic I made a formula that counts the needed upkeep in relation to your oil-supplies and your units. This might be a bit hard to look at if you are not used to mathematical formulations, but the idea should be simple. You pay 1 gpt in upkeep for every allowed unit. So 70 tanks would cost 70gpt if you had >70 oil. For every excess unit you would pay more - and this sum would rise up rapidly. The first idea was with steps of 10, but I also tested with steps of 5. So unit numbers 71-80 would cost you 2gpt, #81-90 3gpt and so on.

First the used symbols:
r = the number of resource-units (oil-units, for example).
n = the number of your units dependent of the resource.
s = the used step (the cost would rise after 10, 5, or b excess units)
m = integer((n - r) / s), this is used to calculate easily the needed cost. It simply gives you the (integer) number of whole steps. If you had 88 units and 60 oil, the number of whole steps (if s = 10) would be 2. The first additional step would cost you s*2, the second s*3 and so on.
t = n - r - s*m. This seems ugly, but actually only gives you the excess units that do not comprise a whole step, the decimals of ((n - r) / s) if you will. So in the example above, t would be 8, marking the number of units of the third (but not complete) step.

Now for the formula. I used arithmetic sum to get this. I spare you of the details.
The cost would be:
r + 5m*(m + 3) + t*(m + 2)

Let us have an example. I have 37 oils and 72 units depending on them. What would they cost me? (let the step be 10)
r = 37
n = 72
s = 10
m = int((n - r) / s) = int((72-37) / 10) = int(3.5) = 3
t = n - r - s*m = 72 - 37 - 10*3 = 35 - 30 = 5
The cost would thereby be:
37 + 5*3*(3 + 3) + 10*(3 + 2) = 15*6 + 10*5 = 90 + 50 = 140 gpt

If the step was 5, the result would be:
37 + 15*9 + 40 = 77 + 135 = 212.

What does this mean in practice, then? First of all it means that it is relatively cheap to have some excess units, so you didn't need to worry about it (exactly my point). 10 excess units would in effect cost you 10 extra (as they would normally cost you 10, now they cost you 20), if using the 10-step. With 5-step the cost would be 5 + 10 = 15. But if would be disastrous to have a large amount of excess units.

Here I have tried to make a presentation of the costs with steps 10 and 5. I would prefer the 5-step system, as it would make the costs rise more quickly. It also seems that some of the variables should still be changed, as there seems to be too small an effect with having more oil. Or perhaps not. The difference between having 100 oil and having 10 oil is still very significant. With 10 oil you would be paying (depending on the steps) 550 or 955 gpt for 100 units, and 100gpt with 100 oil. The difference is huge, as it should be. I think it should be practically impossible to have 100 tanks with 10 oil, but quite common to have 20 tanks with 10 oil. Still, one's economy could handle for a short time 100 tanks with 10 oil, if necessary.

Well, I seem to be quite at loss how to insert the graphics. Perhaps someone could help me?

I will look more into this. And add the picture later. I try to attach it, but I think you must download it then..

In summation. This system of calculation may be hard to look at, but is quite simple to program. The idea of the calculation is easy to grasp even if it is hard to count the exact numbers. Depending on the step, you simply begin to pay more for every unit that exceeds the limit dictated by your resources. This way you can afford quite easily a small number of excess units (so if you lose one patch, you wouldn't be in very serious trouble, you could either get a new one of build other units to replace the excess ones and then disband the excess units - or sell them to other nations, which should certainly be possible in cIV! - or even decide to pay the extra gold), but a large number of excess units would be disastrous - not on the first turn, though. With steps of 5 the cost would rise up more rapidly - which I deem good - and would make it more important to have the resources. With steps of 10, I feel that there is too little difference in having 10 or 30 oil.

Anyway, here are the numbers, as the picture is still missing:

with steps of 10:
10 oil : 10 tanks = 10gpt, 30=60, 50=150, 100=550, 150=1200, 200=2100
30 oil : 10=10, 30=30, 50=80, 100=380, 150=930, 200=1730
50 : 10,30,50=10,30,50, 100=250, 150=700, 200=1400
100 : 10,30,50,100=10,30,50,100, 150=300, 200=750

With steps of 5:
10 : 10=10, 30=80, 50=230, 100=955, 150=1970, 200=3620
30 : 10,30=10,30, 50=70, 100=520, 150=1470, 200=2920
50 : 100=300, 150=1050, 200=2300
100 : 150=350, 200=1100

 

[EddyG17]

well the swordsman could be weaker in defense, or cost more shields. A non-oiled(?) tank could have just one movement point instead of two. But here come the question: Should this units be upgradeable to "normal units"? I say no.

 

[Shyrramar]

I made some more tests with this (after I figured out how Excel works...) and here are some graphs. These range from number of units being 20, 40, ... , 100 and resources ranging from 10 to 100 (marked with "r"). These all made with different steps (earlier marked with "s"), and as I forgot to mark them, they are:
steps of 3: upper left corner
steps of 5: upper right
steps of 7: lower left
steps of 10: lower right

What we can see here (pay attention to the scales, they differ some!) is that with lower steps the costs rise more quickly (obviously). With 10 resources and 100 units the costs are following:
s = 3, cost = 1500
s = 5, cost = almost 1000
s = 7, cost = about 730
s = 10, cost = 550

I would go with 5 or 7. perhaps with 7. I would then be paying seven times more from the units than with 100 resources. This would make it virtually impossible to have 100 tanks with 10 oil. But 50 tanks would "only" cost you about 250 gpt. That's five times more than normally.

What I like about the ones with smaller steps is that the difference between having little or a lot of resources is bigger. This would increase the benefit if you invaded 20 oil more, increasing your 10 oils to 30, if you had, say, 60 units. The difference would be something like this:
s = 3, difference between 10 and 30 oils: 275gpt
s = 5, 170gpt
s = 7, 120gpt
s = 10, 90gpt

Some comments, anyone? Does this formula satisfy you? Which of the steps do you like? The difference is huge, so you should think something of them.

Here's the pic. I would put in midst of my text, but I don't know how. It's not in the internet, so I can't press the IMG (?), so this is attached.

 

[Shyrramar]

And here are the same graphs with unit numbers varying between 20 and 200, instead of 20 and 100.

 

[Suki]

this I's one of my favorite ideas.
good work guys.

that old game deadlock, is a good turm based strategy that does something like this with resources, not the upkeep thing but having a difference between work required and resources required to build something, it was a good game a lot of good ideas that could apply here.

total anihilation did something like that with resources required for upkeep every unit could take a number of energy and/or metal per turn, these could even be negative and produce resources.

Shyrramar:
good work but think a little more about what you're doing here that arithmetic summ is a parabola. wouldn't it be easier to give the formula as:

the upkeep cost linear, and if the you have more than the amount your resources would allow, add cost proportional to the difference squared

or in pseudocode

cost=C1*units;
if(units-resuorces>0)
{
cost=cost+C2*(units-resources)^2;
}

and your idea also would work in the version where you have to pay resources per turn, in fact it would pay to manage better your resources: it would cost the same amount of resource but a bunch less gold if you only use enough resource for 90 of 100 tanks for 10 turns than if you gave resource for the full 100 for 9 turns and no resource for the 10th.
which would make planning very important, and resource management a little too important

 

[Chieftess]

I had an idea like that back when Civ3 was being made. I was told that would be too much micromanagement. My idea was that for each turn, your cities could be allocated a certain number (tonage?) of resources.

i.e., iron would act like gpt, in that you get maybe 50 iron per turn. You have 2 cities that require 10 iron per turn for Swordsmen, using 20 ipt, but leave 30ipt elsewhere, which could be used to help rush something. (Units would have a resource per turn requirement to be built).

 

[judgement]

Yeah, I'm afraid that all the ideas we're suggesting here, cool as they are, might be considered too complex/too much micromanagement.

I had a thought last night on a simple modification of the current "one-tile-can-produce-as-much-as-you-want" scheme that would address the logical inconsistencies without introducing any new complications to the game at all. Simply put, take the idea to its logical conclusion: any tile can supply its resource to as many cities as its connected to, regardless of which civs those cities are part of. As we've discussed, logically, if one tile is enough to supply my entire empire when I have 50 cities, then if I have only 5 cities I should have a surplus to trade to other civs. With this very simple modification, you would always have a surplus to trade to other civs whenever you had access to a resource, no matter how big your empire, and no matter how many civs you wanted to trade it to.

The issues with this suggestion (as I see them) are that (1) the only incentives to get a second tile of a resource is in case your first runs out or to deprive a rival of it, and (2) having a single tile is now significantly more powerful since you can trade it as many times as you want, presumably generating significantly more income, etc. To address those issues, simply make the following additional simple modification: the chance of a resource becoming "exhausted" is related to how many cities are using it, simulating the "finite-ness" of each resource source.

The combined suggestion is thus this: One tile can supply as many cities, from as many different civs, as you want it to, but its more likely to run out the more cities it supplies. As far as I've thought it through so far, the results would be as follows:

1) You'd have an incentive to try to acquire more sources as your empire grows: getting a second source not only gives you a backup in case your first runs out, but it also decreases the chances that either source will run out, since each is now supplying only half your cities.

2) To balance the greater benefit from a single tile, you'd have a disincentive to trade a resource to every other civ since each civ you traded it to would increase the number of cities using your tile, increasing its odds of becoming exhausted. You could trade something to a small civ for a relatively small amount in return, but if you're going to trade something to a big civ with lots of cities, make sure you get a lot in exchange, because you're significantly increasing the resource's odds of disappearing. You're better off not trading to too many other civs at once unless you have multiple sources yourself.

3) You'd have a reason to want to acquire more of a resource by trade even if you already have it. If I already have oil but my empire is large, I might still be interested in buying some from a neighbor as well, because by doing so I decrease the chances that my own supply will run out: my own source now only supplies half my cities, while the trade source supplies the other half.

4) How much a civ wants in exchange for a resource would become dependant not only on how big you are but how big they are and how many sources they have. A small civ with many sources might happily sell you some relatively cheaply, since they're not worried about running out, but a big civ with few sources would be more reluctant and would ask for more in exchange, since they're already using the resources in a lot of cities back home. This would make shopping around to buy a resource more interesting and more realistic.

The beauty of this idea (as opposed to my other "stockpiling" idea, and many of the other cool ideas proposed here) is that it requires almost no change to the user interface, but resolves the logical inconsistency of the current system. The only interface change I see would be that on the screen where you're negotiating with a rival, next to the word "Cities" it should parenthetically add the number the civ has, allowing you to see at a glance how big their empire is without expanding the list and counting the entries. But that would be a very minor addition and useful anyway. And, of course, the same screen would not gray out resources if you only had one source, since they would still be tradeable in that case.

One last thing, as I've suggested elsewhere: resources should not always reappear somewhere else on the same turn they disappear, instead, these two events should be governed by separate random numbers (with potentially different odds, so that the number of tiles of a resource could tend to increase or decrease throughout the game). This would be even more important with the proposed change to resource usage, because otherwise, a really big empire would worry less about exhausting their resources since chances are good they'd immediately reappear somewhere else within the empire anyway.

 

[Shyrramar]

Originally posted by Suki
Shyrramar:
good work but think a little more about what you're doing here that arithmetic summ is a parabola. wouldn't it be easier to give the formula as:

the upkeep cost linear, and if the you have more than the amount your resources would allow, add cost proportional to the difference squared

Yes the formula is in effect a parabola. I have of course thought of parabolas and exponential functions, but there are two reasons I had in making the formula as it is.

(1) The formula is based on a simple idea of how the cost would rise. The idea is easy to grasp even for those who aren't math-experts and the effects would be easy to calculate and to control. The idea is that once you unit number exceeded your resources, each unit would cost two gpt instead of one (or it works just as fine with the allowed units costing nothing). If the step was ten, then you would know that each exceeding unit would cost first two, then three and so on. This would be very easy to understand. If the function was made parabolic, it would become abstract.

(2) The actual code is unimportant. The maths involved is relatively simple to coders, so it should not present a problem. There is no need to make the code any simpler. In effect it is a parabola, but with important differences caused by the integer-part of the formula. It could be turned into a parabola that followed along the lines of this function and then turned into integers, but I am not sure of this and will not pursue the matter for now. The integers might have an effect on the parabola so that it wouldn't actually work.

So the point here is simply this: It is not important if the actual formula is easy or not - both suggested are easy enough. What IS important is that players can easily calculate their losses. This is what my system allows.

And another point: the parabola cannot be made as simple as you have suggested. My formula includes first-degree terms whereas the suggested one does not...

and your idea also would work in the version where you have to pay resources per turn, in fact it would pay to manage better your resources: it would cost the same amount of resource but a bunch less gold if you only use enough resource for 90 of 100 tanks for 10 turns than if you gave resource for the full 100 for 9 turns and no resource for the 10th.
which would make planning very important, and resource management a little too important

It is a good idea, but something I would rather avoid. I do not like the idea of being forced to "manage" your resources. I like the same idea applying to resources as not applies to money: I know in theory where it is coming from, I know how to get more, but I don't have to pay attention to the cash-flow if it is okay. Same idea here. You would know where the resources come from and how to acquire more. If you have less units than resources, you needn't worry. When your units exceed the number (this would of course be shown in detail in trade-advisor or some other advisor) allowed, you would begin to lose money. This I know is easily fixed (either buy more resources or disband your units). I can also easily calculate that I can afford having 10 extra units. The advisor screen would also show the exact amount of extra money you are paying.

I would also dread the idea of having to somehow remember which turn some units would cost something and which not.

Those are good comments, though. I would like to hear some more comments on the matter

Originally posted by Chieftess
I had an idea like that back when Civ3 was being made. I was told that would be too much micromanagement. My idea was that for each turn, your cities could be allocated a certain number (tonage?) of resources.

i.e., iron would act like gpt, in that you get maybe 50 iron per turn. You have 2 cities that require 10 iron per turn for Swordsmen, using 20 ipt, but leave 30ipt elsewhere, which could be used to help rush something. (Units would have a resource per turn requirement to be built).

I would like to point out why this wouldn't cause micromanagement - not so much to answer you as to answer those who complain about the micromanagement. I agree that this sort of resource-handling could lead to micromanagement. Your idea seems good, but could indeed be a bit complex to actually use in the game. It is hard to say until playtesting if it would work.

In the system now things are as easy as they can be. You have a resource, you can build how many units you want to. You don't, you can't build any units. You can buy an infinite stack of resources from others and you can sell them if you have more than one. Simple, but has many flaws.

My system would works as this. You have a resource, but that resource is not infinite. It can support a certain number of units. This would be a mere number, so it would not cause any trouble. You can also decide to sell SOME of the resource to others, or buy SOME more. The difference in trade would be that you didn't simply buy "oil", but perhaps "30 oil" - no different from gold. Okay then, this is easy. It has actually produced NO micromanagement and no complexity whatsoever. You only need to keep an eye of the number of resources being used. An advisor could pop up and say: "Sire, Washington is building a tank, but our resources is low. Shall it build it anyway?" This way you needn't worry about that either - not more than of your money currently.

What would this fix?
It would fix the problem of No Oil. A real game breaker. If you have no oil, you should pay ridiculous amounts of money to get it - and you many not even be able to get it anyway. My system assumes that resources are more common, but they would not contain an infinite amount of the resource. This way it should be almost impossible to be left COMPLETELY without oil, but more common to have too little of it. This would allow a small army of, say, tanks created to invade more oil. It's pretty hard to invade rubber and oil with riflemen and knights!

The only part of this what could be called complex is the situation where you have too many units compared to your resources - either by simply building too many units or by loss of resources. The exact effects of this have been debated. Here are the thoughts presented about this:
(1) Your excess units would be randomly (?) disbanded - exactly as when exceeding your treasury.
(2) Your excess units (which of them?) would get penalties to their ADM.
(3) You could only move the allowed number of units each turn.
(4) You would pay extra for the units - representing smuggling and black market sales and other related things.

Now the first is too harsh. Losing your city for one turn and then recapturing it would mean that you excess units would have already been gone? The second is possible, but I wouldn't like it in my game: which units, how much? It is not out of question, though. Now the third would be too complex and would result in micromanagement, although perhaps being the most realistic of them all. I would go for the fourth. This would not cause any short-term catastrophies, so you could easily pay the extra during war, but would in the long run severely slow your economy. It would be easy to handle and would not bring any new concepts into the game.

Now I have argued (whether my arguments are sound or not) that everyone could grasp the simple idea of paying two gold instead of one in upkeep, and then three, and then four. It is countable, and understandable. The exact formula may look a bit hard, but that is not something a common player should worry about. The developers would think that as simple (Compare to the corruption calculations!) and modders should be able to understand the mechanics.

The more I think of it, two things become clearer and clearer: SOME change is needed and this suggestion is fully possible, if not the best.

@Judgement: Once again you astound me with your innovative thinking. These sort of ideas are sorely needed - we don't want to bang our heads into the same wall all the time Let me try to ask some questions about it...

Originally posted by judgement
Yeah, I'm afraid that all the ideas we're suggesting here, cool as they are, might be considered too complex/too much micromanagement.

Do you still think so, after all my ranting?

The issues with this suggestion (as I see them) are that (1) the only incentives to get a second tile of a resource is in case your first runs out or to deprive a rival of it, and (2) having a single tile is now significantly more powerful since you can trade it as many times as you want, presumably generating significantly more income, etc. To address those issues, simply make the following additional simple modification: the chance of a resource becoming "exhausted" is related to how many cities are using it, simulating the "finite-ness" of each resource source.

This idea has potential, I must admit. I am bit sceptic, though. This would of course assume that there aren't more resources than there are currently in the game, right? With more, this idea would have no significance as virtually all civs would have their resources. If less...well, perhaps less would even be tolerated if the probabilities are handled well and AI would indeed sell it (and NOT from outrageous sums).

1) You'd have an incentive to try to acquire more sources as your empire grows: getting a second source not only gives you a backup in case your first runs out, but it also decreases the chances that either source will run out, since each is now supplying only half your cities.

I'm with you mate!

2) To balance the greater benefit from a single tile, you'd have a disincentive to trade a resource to every other civ since each civ you traded it to would increase the number of cities using your tile, increasing its odds of becoming exhausted. You could trade something to a small civ for a relatively small amount in return, but if you're going to trade something to a big civ with lots of cities, make sure you get a lot in exchange, because you're significantly increasing the resource's odds of disappearing. You're better off not trading to too many other civs at once unless you have multiple sources yourself.

Here is your weak spot, in my opinion.
The probabilities would be impossible for mere mortals to handle. It would remain abstract. You would only be vaguely aware that selling it abroad would increase your risk of losing your resources, but you would certainly not understand the risks involved. Even mathematicians would have hard time trying to figure out if they should sell their only resource to a third civ or not, even if they could calculate the exact risks. This would of course increase tactics, but I think that you should always understand the rules and parameters that you play within to actually REALLY use tactics, and to really enjoy the game. That's the thing behind the popularity of games like chess: few rules, endless possibilities. Each move's effects are concrete and understandable. That's where MOO3 went awry, I think: atleast I had NO IDEA of the mechanics behind my economy - why it was soaring now when just a moments ago I was losing tons of gold per turn?

This change in probability should either be simply displayed somewhere or made simple enough for everybody to grasp - as I have tried to do with my resource-cost-model.

The beauty of this idea (as opposed to my other "stockpiling" idea, and many of the other cool ideas proposed here) is that it requires almost no change to the user interface, but resolves the logical inconsistency of the current system. The only interface change I see would be that on the screen where you're negotiating with a rival, next to the word "Cities" it should parenthetically add the number the civ has, allowing you to see at a glance how big their empire is without expanding the list and counting the entries. But that would be a very minor addition and useful anyway. And, of course, the same screen would not gray out resources if you only had one source, since they would still be tradeable in that case.

I agree with you on this. This would indeed not change the interface (not by much, anyway). But I think the same applies to my system, does it not?

One last thing, as I've suggested elsewhere: resources should not always reappear somewhere else on the same turn they disappear, instead, these two events should be governed by separate random numbers (with potentially different odds, so that the number of tiles of a resource could tend to increase or decrease throughout the game). This would be even more important with the proposed change to resource usage, because otherwise, a really big empire would worry less about exhausting their resources since chances are good they'd immediately reappear somewhere else within the empire anyway.

This is of course irrelevant to your system, and it is good that it is so, as I think it is a bad idea. Sorry.
For starters let me defend the current system: it makes it certain that at any given time there are always enough resources to go by. So the game wouldn't break because there was only one civ having oil.
Then why is your suggestion not good? Because it could cause a snowball effect. Let's say that all 5 resources of oil are being used by 7 civs globally. Now one disappears. This would mean that the remaining oils are subject to a bigger stress of running out too. This would either cause the remaining civs not to sell their oil to the civ left without the oil, or not. If not, then the probability of the remaining oils to disappear would become greater. Then if because of that the fourth oil would disappear, it would of course stress the remaining oils even more. This could cause an effect that would soon have all the oils gone, or simply that many civs would be left without - and in your system it would mean COMPLETELY without.

This would of course be rare, but possible anyhow. And because probabilities work in such a fashion, that it is very probable that something very improbable will happen quite often (like people winning in lottery, which is near impossible). So there would most certainly be many games ruined by missing resources. The current system totally eliminates this risk - even if it is less realistic.

Thank you for your great comments! As I mentioned earlier, I see this as a very important change in the game and it should be pondered most thoroughly. Please do keep the conversation going!

 

[judgement]

Originally posted by Shyrramar

Do you still think so, after all my ranting?

No, but I don't work for Firaxis, and that's whose opinion matters here. I like your idea (I like my stockpiling idea, too) but this new suggestion is an alternative in case the Civ 4 developers respond to our other ideas the same way the Civ 3 developers responded to Chieftess.

This would of course assume that there aren't more resources than there are currently in the game, right?

Yes, same as the current game, or possibly slightly less to compensate for trading being easier.

Here is your weak spot, in my opinion.
The probabilities would be impossible for mere mortals to handle. It would remain abstract. You would only be vaguely aware that selling it abroad would increase your risk of losing your resources, but you would certainly not understand the risks involved. Even mathematicians would have hard time trying to figure out if they should sell their only resource to a third civ or not, even if they could calculate the exact risks. This would of course increase tactics, but I think that you should always understand the rules and parameters that you play within to actually REALLY use tactics, and to really enjoy the game. That's the thing behind the popularity of games like chess: few rules, endless possibilities. Each move's effects are concrete and understandable. That's where MOO3 went awry, I think: atleast I had NO IDEA of the mechanics behind my economy - why it was soaring now when just a moments ago I was losing tons of gold per turn?

An interesting point, but I disagree, both about the math being complicated, and about that being a big problem. Let me see if I can articulate the latter point. If I attack your warrior with my swordsman, I know, based on the simple (easy to grasp) mechanics of the game, that my 3 attack vs. your 1 defense means I have a high probability of success. But there's still no way for me to know ahead of time who will win, or even to know roughly how many times I'll win if I try the same thing 20 times (I could always have a run of good or bad luck). Understanding the mechanics makes it more likely that I can guess what might happen (I'm likely to win most of the time) but its still just a guess. Because random chance is involved, understanding the mechanics can never give me complete knowledge.

I play Civ 3 just fine without knowing the exact probability that my oil resource will disappear on any given turn (BTW, does anyone know this number? As I recall, the number in the editor isn't a simple percent chance per turn - or is it?). Its enough for me to understand that there is some chance, that it is completely random (no way I can affect it), and that when it does happen oil will reappear somewhere else on the map. Being able to calculate the exact probabilty wouldn't really add anything to my game. For that matter, when I attack your warrior with my swordsman, I don't bother to calculate my exact odds of winning, either, I simply rely on my understanding of the mechanics to give me a feeling of whether the odds are reasonably good or not. (I'll admit, though, that some people do make use of those combat calculators to know exactly the odds, but those aren't part of the game as its sold).

Since the resources disappearing would be (and is currently, for that matter) a random event, there's not much point in being able to calculate exactly the odds of it happening. Knowing that there was only a 1% chance of something happening is a small consolation when you get unlucky and it happens. That's why the tank-killed-by-a-spearman is so irritating to people: they think that because they've calculated the odds and the odds are very small, it should never happen to them, but then it does.

I'm rambling on, probably not making my point very well, sorry. You'd know, quite simply, that the more civs that you traded something to, and the bigger they were, the more likely resource disappearance would be. If you had only one source and wanted to safeguard it, you wouldn't trade - you might not know the exact odds of disappearance, but you'd know what to do to minimize those odds (and you wouldn't do anything different even if you knew the exact probability). In the current game, you can't trade when you only have one tile of something anyway, so making it risky to do so isn't so terrible. If you had more than one source, trading wouldn't be too risky, since if you lost one you'd still have access. Resource disappearance is unlikely enough that the chances of more than one source disappearing in rapid sequence would be very small indeed (hopefully even less likely than a tank-losing-to-spearman ), so if you had two and you lost one, you simply could avoid renewing trade deals. Essentially, you wouldn't really need to bother calculating odds when deciding whether to trade your resource to a third (or fourth, or whatever) civ. Simply don't trade when you have only one source, unless you're desperate and willing to take the risk, and if you have multiple sources, trade away, but remember to demand more in exchange from civs that have a larger number of cities.

Besides, it wouldn't take a graduate degree in math to add up the total number of cities used by a resource and see how that compares with and without a given trade. If its 50 cities without the trade in question and 55 with it, no big deal, but if its 20 without and 60 with, its pretty obvious you'd triple the disappearance chance. To make this easier, one further interface change: Next to each of your resources on the negotiation screen, it would tell you how many cities currently use it, i.e., replace "Oil (2 extra)" with "3 Oil (24 cities)". (The number of cities could be different for different resources if you have traded some to other civs as well). If your oil is currently being used by 24 cities and your potential trading partner has 8 cities, then the only math you need to do is to compare 24 with 24+8. All the info you need is right there on one screen, no fancy math needed! The only relevence of the fact that you have 3 sources is that you know how much cushion you have in case a source disappears: switching from supplying 24 to 32 cities alters the disappearance odds by the same amount regardless (to be extra clear here, the disappearance odds aren't the same if you have 2 sources, or 4, but the relative change in odds due to the proposed trade is the same).

This change in probability should either be simply displayed somewhere or made simple enough for everybody to grasp - as I have tried to do with my resource-cost-model.

Nowhere does the game display the odds of winning a combat, and yet, even though my attack strength versus your defense strength seems simple enough, the actual odds are quite complex, taking into account terrain and how many hitpoints we each have and whether or not you have defensive bombard and so on. Unless one uses a combat calculator, calculating the exact probability each time one decides whether or not to attack is simply not possible, yet that doesn't stop one from understanding the mechanics and using that knowledge to make an informed decision. Resource disappearance would work the same way: you might need an exterior calculator if you wanted to know the exact chances, but without that you would still know that trading something to a civ with 50 cities would be a lot riskier than to a civ with 5.

This would indeed not change the interface (not by much, anyway). But I think the same applies to my system, does it not?

(from elsewhere in your post...)
An advisor could pop up and say: "Sire, Washington is building a tank, but our resources is low. Shall it build it anyway?" This way you needn't worry about that either - not more than of your money currently.

Your system certainly would require less interface overhaul than my "stockpile" system, but more than the (virtually zero) changes in my more recent idea. You admit your idea might require extra advisor popups, and there would need to be some screen somewhere that told you how many you were using and trading out of how many you had total. The latter would be a very minor thing to add, but the former is quite unacceptable: the best improvement of Civ 3 over Civ 2 is how much less often you have to click "Okay".

(In response to my idea about separating the appearance and disappearance of resources)
This is of course irrelevant to your system, and it is good that it is so, as I think it is a bad idea...
... it could cause a snowball effect.

There's no reason the chance of appearance has to be different from the chance of disappearance, and even if it was different, no reason it has to be unrelated. As you say, the more a resource disappears, the more stress on the remaining tiles, increasing the likelihood of disappearance. But the snowball effect is only there if the chance of reappearance remains unchanged. If it also increases as more resources disappear, then scarcity of resources would at worst be temporary: the fewer there were, the more likely that new ones would appear. Likewise, if a few extra oil (or whatever) tiles happened to appear before any of the originals had disappeared, the chance of further appearances could decrease, preventing a snowball in the other direction. The number of oil tiles on the map would vary throughout the game, instead of remaining constant as in Civ 3, but that variation would always tend to return towards the default amount. It would add periods of shortage and periods of surplus to the game, which could be a nice, "realistic" touch.

If would be quite simple to have the computer, on each turn, sum up the total chance than any oil tile will disappear, and assign the same probability to the chance that a new source will be discovered. As resources were used heavily and the disappearance odds increased, the reappearance odds would increase to match. As a variation, the chance of reappearance for some resources could set each turn to 90% or 110% (or whatever, and these could be set in the editor) of the disappearance chance, to allow you to have resources that tended to increase or decrease in number throughout the game, if you so wanted. The reappearance chance would still be tied to the disappearance chance, so the snowball effect would still be eliminated, and the result would be oscillation about and upward or downward trend rather than a fixed value.

Thank you for your great comments! As I mentioned earlier, I see this as a very important change in the game and it should be pondered most thoroughly. Please do keep the conversation going!

I'll try, but judging from those plots you posted, I think I'm up against someone with way more free time than myself.

 

[Shyrramar]

Originally posted by judgement
An interesting point, but I disagree, both about the math being complicated, and about that being a big problem. Let me see if I can articulate the latter point. If I attack your warrior with my swordsman, I know, based on the simple (easy to grasp) mechanics of the game, that my 3 attack vs. your 1 defense means I have a high probability of success. But there's still no way for me to know ahead of time who will win, or even to know roughly how many times I'll win if I try the same thing 20 times (I could always have a run of good or bad luck). Understanding the mechanics makes it more likely that I can guess what might happen (I'm likely to win most of the time) but its still just a guess. Because random chance is involved, understanding the mechanics can never give me complete knowledge.

I must in turn disagree with you on this point (who's disagreeing who in what point...and WHY? ). The difference between understanding probabilities that are between 10 and 90 percent and the probabilities being very small (or big, it does not matter much) if big. I understand that I have relatively good chances of beating up that warrior, as you said. I do it often and get a grip onto the probabilities. Enough to rip my hairs off if my swordsman lost - which is not the case if it lost to a spearman. The counter-argument I will use is actually from your text: people get annoyed when their modern armor lost to a spearman (which is bound to happen time to time). Now the disappearance ratio of the resource should be very small, agreed? It is easy to grasp 1:3 probabilities, as they are often on display. But if something happens very rarely, you don't get a similar touch to it. Like how probable it actually is to win in lottery? There is actually a certain probability of iron disappearing (you asked this somewhere). I don't remember it, but it is something like 1:400. The disappearance ratio must be the divider and 1 the divided. This is the possibility of it disappearing each turn. It happens so rarely now that I almost always express curses when it happens - and I wouldn't understand the difference between 1:200 as it happened only once in 200 turns! (in opposition to my swordmen winning the warrior something like 19 times out of 20 - this is just a throw, nothing calculated or even thought).

Now if I had one oil, I would trade it. If the disappearance ratio was 1:400, 1:200 wouldn't make a difference: I would deem it better to trade than not trade, as I would (in my opinion) be taking a very little risk. The problem here is that you should play the game VERY much in order to grasp to real risk you are taking - due to the very small probabilities.

Do you understand my point? I am not sure if I am getting it across properly. The point is simply that having something happening many, many times during a game gives me an idea of what I should attack with what. But if it happened just a few times in a game, it would be really hard.

I play Civ 3 just fine without knowing the exact probability that my oil resource will disappear on any given turn

This is because you have no choice: you don't affect the outcome. It simply happens if it happens. You can't affect it, so you don't bother about the probabilities, just with a curse when it happens.

I am making this point: if the probability is small, people will not take it into account. If the probability is big, it would be better - but the constant shifting of resources would get on my nerves, anyway. So how to get the probability right: so that it would be big enough to handle and small enough to not cause irritation? Could you actually post some numbers, so we could somehow try to understand how it works?

Don't get me wrong: the idea is good, I have some reservations that I am expressing.

the best improvement of Civ 3 over Civ 2 is how much less often you have to click "Okay".

Good point! Well, it could be something that you didn't have to press "ok" to. Just a notification that shows up in the right upper corner of your screen or something. All the data would be in the advisor screen, the notification is just a wake up call, so you wouldn't have to check it manually all the time. It could also say someting like "Athens has produced a tank. 5 oil, 7 rubber left". It does indeed get irritating to press OK whenever the advisor comments about my way of handling the treasury!

There's no reason the chance of appearance has to be different from the chance of disappearance, and even if it was different, no reason it has to be unrelated. As you say, the more a resource disappears, the more stress on the remaining tiles, increasing the likelihood of disappearance.

You totally convinced me with this one! I withdraw my objection: if the probabilities are handled well, I am with you in this. This would, as you said, indeed bring welcome variation to resources. This way a certain time without any oil would actually be interesting - provided that it would take probably about 10 turns until they started popping back! Nice thinking there, pal!

I'll try, but judging from those plots you posted, I think I'm up against someone with way more free time than myself.

Yes, beware. There is nothing more dangerous than a man with opinions and way too much time on his hands!

 

[EddyG17]

Different classes of resources could be added too. You could have normal oil but if you invade your neighbor you'll have high quality oil which is better that your normal oil. Or since we are talking about how much a oil resource will last or produce oil units, this high quality oil will be a larger deposit of oil than normal.

 

[Shyrramar]

The size of deposits is what I would like to see (it's implemented in my system), but different qualities of resources I think would only make the game more complex. I wouldn't like to keep track of what quality my oil is - I should have better things to do (such as invading Russia

 

[judgement]

Originally posted by Shyrramar
I must in turn disagree with you on this point (who's disagreeing who in what point...and WHY? ). The difference between understanding probabilities that are between 10 and 90 percent and the probabilities being very small (or big, it does not matter much) if big. I understand that I have relatively good chances of beating up that warrior, as you said. I do it often and get a grip onto the probabilities. Enough to rip my hairs off if my swordsman lost - which is not the case if it lost to a spearman. The counter-argument I will use is actually from your text: people get annoyed when their modern armor lost to a spearman (which is bound to happen time to time). Now the disappearance ratio of the resource should be very small, agreed? It is easy to grasp 1:3 probabilities, as they are often on display. But if something happens very rarely, you don't get a similar touch to it. Like how probable it actually is to win in lottery? There is actually a certain probability of iron disappearing (you asked this somewhere). I don't remember it, but it is something like 1:400. The disappearance ratio must be the divider and 1 the divided. This is the possibility of it disappearing each turn. It happens so rarely now that I almost always express curses when it happens - and I wouldn't understand the difference between 1:200 as it happened only once in 200 turns! (in opposition to my swordmen winning the warrior something like 19 times out of 20 - this is just a throw, nothing calculated or even thought).

Now if I had one oil, I would trade it. If the disappearance ratio was 1:400, 1:200 wouldn't make a difference: I would deem it better to trade than not trade, as I would (in my opinion) be taking a very little risk. The problem here is that you should play the game VERY much in order to grasp to real risk you are taking - due to the very small probabilities.

This is funny, cause we're using the exact same examples to argue opposite sides
I would agree, its very hard for the human mind to understand the different between 1:200 and 1:400, unlike the difference between 1:2 and 1:4, which is quite obvious. But to me, that argues in favor of my idea: the odds of a resource disappearing are so low that you wouldn't really understand them even if you did calculate them precisely, so why worry about trying?

I guess I do see your point, somewhat, though: its better if the decisions you make are based on understanding. If your brain can't wrap itself around the difference between 1:200 and 1:400, then why would you choose to trade with a civ with 10 cities but not with one with 50 cities - is that what you're asking? Both situations seem the same when the odds of disappearance are small enough?

I am making this point: if the probability is small, people will not take it into account. If the probability is big, it would be better - but the constant shifting of resources would get on my nerves, anyway. So how to get the probability right: so that it would be big enough to handle and small enough to not cause irritation? Could you actually post some numbers, so we could somehow try to understand how it works?

Lets try to come up with some numbers, then. Suppose at first that the world is evenly divided among all the civs in it, so that, if there are 10 civs playing, you have 10% of the territory. Set the per-turn chance of disappearance such that the overall chance of a disappearance occuring to you during the entire duration a resource is relevant (for oil, that's about half of the game, turn-wise) is maybe 1/3 or 1/4. That way, it would happen to you in a reasonable number of your games, but not all. Now, if you either trade your oil to one other civ, or conquer an additional 10% of the map, you'd double the number of cities using the oil, so instead of happening less than half your games, disappearance would happen in the majority of them, although usually only once per game still, more than that would involve a little bad luck. But if you trade the oil to 4 other civs at once, or your empire expands to cover half the map, then its very likely that resource disappearance will happen more than once to you during a typical game (in the cse of covereing half the map, this would be balanced by the fact that you'd probably control more than one source). If you played one game where you kept your resources to yourself, and then played another where you traded them as much as possible, you'd really notice a difference.

Let's assume there are 200 turns in which oil is relevant, and maybe the per-turn chances are such that if 10% of the cities are using a source, it has a 1:400 chance of running out each turn. By my rudimentary calculations, that gives a roughly 39% chance of that source running out before the end of those 200 turns. But if you supply 20% of the cities with that source, the odds are 63% that it will run out before 200 turns, with the average time until depletion being 138 turns. And if you supply half the map, there's a 92% chance of at least one disappearance, with an average of only 55 turns between disappearances.

The point is, although the chance per-turn is quite small (too small for our brains to really digest), over the course of many turns, those small chances add up to larger chances that are big enough to notice and understand. There's a noticeable difference between 39%, 63%, and 92%, and a very noticeable difference between something usually not occuring and something occurring multiple times per game.

I admit these calculations would not be readily available to you during the game, but it wouldn't be hard (or take very many games played) to notice that when you didn't trade many resources and acquired a number of sources roughly proportional to you empire size, you usually had enough, but if you traded a single source to everyone and/or never acquired more as your empire grew huge, you'd frequently have shortages.

I don't recall the exact numbers of resources in Civ 3, but its seems that there's usually roughly as many sources as civs, give or take a few. Maybe a few less of the later game ones like oil, a few more of early ones like iron and horses. Assuming that didn't change, and the per-turn chances were set right, you could look at the map and think things to yourself such as "This is a 10-civ game and I control about 25% of the map, so I probably need 2 or 3 sources to have good odds of not running out. I only have one, I'd better look to get more, and not try trading my single source with others" Or, something like "There are 8 civs in this game and I control roughly 1/8th the map, but somehow I find myself with 3 sources of oil. That's plenty for a nation my size, I should trade some to others."

As an alternative to this system, here's another idea: instead of a random chance each turn for a resource to run out, it could run out after a random number of uses, which is determined when the resource first appears (at the beginning of the game, or later for reappearances). That way, you would know for sure that by trading a resource to multiple partners, you would be shortening its lifespan. I don't like this version as much, but it would remove your complaint about small percentage chances being hard to deal with intelligently.

 

[EddyG17]

you wouldn't need to worry much about your oil type because the number of units you could build with it would be mixed together with all the oil types you have in a grand total.

But intead of conquering a Ruissian city with 6 low quallity/small deposits of oil you might be interested in the city next to it which has 2 high quality/large deposits of oil, which are like 8 low quality/small deposits of oil.

 

[Shyrramar]

@judgement: I just replied your post and in the end decided to have a nice little smiley in it, and guess what? It opened the smileys-window ON TOP OF MY POST, so the reply was gone in a blink of an eye. It's past midnight already, so I would reply to your post tomorrow in greater detail. Let me just ask you one more thing: I got the same probabilities, but I couldn't figure out how you got the average turns between disappearances. It might be the time or the fact that it's been long since my last probabilities, or both, or the fault might be in you. So could you apply the calculations also? Thanks in advance. As an overall comment: you once again managed to convince me! (and that's a hard thing to do, mind you )

Nice job!

EDIT: Oh boy! It seems I am on warlord now. I guess the barbarians will write tougher posts now, and I should be wary of my fingers, as they will now strike the wrong buttons more often (I must get them some more luxuries), or something. But one thing is certain: this is going to get much harder from now on, as you guys become more intelligent - and get to cheat! You now need only press seven buttons to get eight letters, so you can write faster now and overwhelm me with your posts and quick replies....

 

[judgement]

Originally posted by Shyrramar
@judgement... I got the same probabilities, but I couldn't figure out how you got the average turns between disappearances.

Well, if you must know, I made plots of cumulative probability versus number of turns. the 39%, 63%, and 92% were the values at 200 turns, while the "average time until a disappearance" was simply the turn at which the curve crossed 50%. Half the time, a disappearance will happen sooner, half the time, it will take longer. Not the most rigorous, I grant you, and I guess its technically the median turns until a disappearance, which is not exactly the same thing as the average (mean) number of turns between disappearances.

It occurs to me that your trade advisor could also be useful here. Rather than always giving you useless messages like "We should acquire more sources of oil!", the advisor could take into account the size of your empire and the number of sources you have, and say "We should acquire more..." only when the chance of disappearance was abnormally high. In fact, something like "our empire is growing, sire, we should seek more sources of oil" would be even more to the point. If you had plenty, and the chance of disappearance was low, the advice would switch to "We have plenty of oil, sire, lets see if anyone would like to buy some!"

And one final thing on this topic: if you had one oil and so did your friendly neighbor, you could contact them and offer to trade oil for oil. Then you'd each have two sources. The total chance of disappearance would remain the same, but now if your source was the one that vanished, you'd still have access to their source, at least until the trade expired. Vice versa if theirs vanished - they'd have access to yours. You could cover each other's back that way. Maybe, in the same way that the AI civs sometimes contact you and ask to trade world maps, the would occasionally contact you and say something like "Let us pool our oil resources in case either of our oil fields run dry!" (Cleopatra offers Oil and 5 GPT, asks for Oil). The gold might be in there if Egypt is bigger than you, to compensate your for the increased likelihood of losing your source if you have to supply more of her cities than she has to supply of yours.

 

[rcoutme]

One of the things I find missing in the resources is the ability to trade from one civ to another via an intermediary. The entire reason that N&S America were discovered was because the Europeans were tired of paying the exorbitant prices for spices and silks. My civ (or the AI) should be able to purchase extra resources (beyond what I need) and sell them to other civs who do not have contact with the supplier (or who the supplier refuses to sell to).

 

[Shyrramar]

Well, if you must know, I made plots of cumulative probability versus number of turns. the 39%, 63%, and 92% were the values at 200 turns, while the "average time until a disappearance" was simply the turn at which the curve crossed 50%. Half the time, a disappearance will happen sooner, half the time, it will take longer. Not the most rigorous, I grant you, and I guess its technically the median turns until a disappearance, which is not exactly the same thing as the average (mean) number of turns between disappearances.

Of course I must know! I'm nosy!
I know very little of mathematical terms in English, but I try to sound understandable. I think you needn't use plots. The first percents are simple enough to count. If the disappearance probability is 1:200, then the odds for having atleast one disappearance would be the complement of having no disappearances whatsoever:
1 - (199/200)^200 = 39%. With the same calculations you would have 63% and 92 %.

The median it is, and that explains. But I still think there is something wrong there. With binomical probability/binome probability/I-have-no-idea-what-probability (I don't know if it's the proper term in english) the probs should be easy to calculate. With 1:80 prob of disappearance, the number of disappearances in 200 turns would be:
0: (79/80)^200 = 0,08
1: (200 nCr 1)(1/80)^1*(79/80)^199 = 0,20
2: (200 nCr 2)(1/80)^2*(79/80)^198 = 0,26
It is actually here where the 50% line is crossed. This would make the median disappearances 2, which would mean something like once in 100 turns. Actually, come to think of it, if the probability is 1:80, shouldn't it simply occur once in 80 turns? That's what probabilities mean, isn't it?

This does give a slightly rarer picture of the disappearances. This would of course be corrected by lowering the probability to start with, or make the selling affect the disappearance ratio more.

Whatever the case, I think your calculations have shown that this could indeed work in reality with quite simple formulas.

And what's with the plots, judgement!? So it seems that you are not so busy yourself! Tricky, tricky



Thanks for the calculations, by the way. I think they are an important part of this kind of suggestions: you can't really know it is something you can actually implement until you have a rough idea how it would work actually. I am still on the process of actually showing that my new AI-system (or the system is old, but anyway) would actually work. That can't be done with simple numbers, but the idea can be shown and made concrete.

You're a real treat, judgement!

 

[Shyrramar]

@rcoutme: A great idea! This would become as a given in my system (I do advertise a lot, it seems!). You could buy resources even though you had enough - this excess could of course be re-directed. I didn't thought of this, but now that you mention it, it is an important feature - and luckily part of my system

EDIT: Welcome to CFC, by the way! (silly me didn't notice it in the first place. That's what you get for having the same head summers and winters through! )