Yields

[Thalassicus]

Goal

1 citizen costs 1 2, and can work a tile for 3 3 3 or 3.

Tile yields must be whole numbers. We cannot have 1.5 gold on a tile, so a ratio like 1.5 gold = 1 production would not be a practical goal. (The value of food drops when our empire is unhappy, so it's balanced with a slight bonus - lower numbers are more powerful.)


Problem

Consider a city with 12, a library, and 1 scientist. We get 1.8 sci per pop, which does not meet the goal. Scientists contribute 15% of the total science in this city. The numbers prove the intuition many people feel - scientists are too influential:

12 = 1. * 12 from pop
06 = .5 * 12 from library
03 = 3. * 01 from scientist
====================
21 / 12 = 1.8 per
03 / 21 = 15% from scientists

Now consider if 12 citizens are working 11 villages of 2 apiece, 1 merchant, and 1 market:

02 = city tile
22 = villages
03 = merchant
09 = market
====================
36 / 12 = 3.0 per
03 / 34 = 9% from merchants


Possible Solution

Now consider that if we double all science income and double all tech costs, nothing really changes, right? We just make all the numbers bigger. So let's consider what happens if we apply something like this to science. Consider a similar city to before:

24 = 2 * 12 pop
12 = 1 * 12 library
03 = 3 * 01 scientist
====================
39 / 12 = 3.2 per
03 / 21 = 8% from scientists

The final values of science are almost identical to those for gold, and very close to the goal of 3 yield per pop.

What are your thoughts on this analysis? Do you have any questions, or see anything I might have overlooked?

 

[wobuffet]

Sounds solid; even with VEM's nerfs to GS and buffs to other GP, Specialists aren't quite balanced, and Scientists are definitely the strongest of the bunch as is.

I almost never use Merchants, for instance. I've put some ideas for making Merchants better here.

I think a much bigger issue in the current game, from a yield equality perspective, is Gold becoming much much less useful over the course of a game, especially by turn 180 or so. Especially considering that the AI does not know how to spend its Gold well at all (hence the inevitable AI treasuries that perennially contain thousands if not tens of thousands of Gold each), I think either
(1) Gold yields have to be much harder to raise mid-game, and/or
(2) some maintenance costs have to go up.

Reducing the % benefit from Open Borders over time (once certain techs are researched, say) could mitigate, but not solve this issue, in my view.

 

[Thalassicus]

How about higher late-game research agreement prices? I could also increase the cost of maintaining armies, though it's already somewhat high... I'm concerned that if maintenance goes too high the AI won't be able to handle it well.

Open border benefits are rather low. I think reducing them further wouldn't have much psychological impact on the perceived value of gold. It's important that bonus be there to encourage friendly diplomacy.

 

[wobuffet]

I've just posted a humongous wall of text in the Specialists thread on this topic (sorry for being so pedantic there, incidentally).

You're probably right about the Open Borders. My impression is that all the AI players who are doing well in my games always seem to have more Gold than they know what to do with (Presumably because they get reduced maint and gold bonuses?). I'm okay with reducing Gold yields and/or raising prices to the point that the human player needs to scramble to catch up if it means the AIs won't be simply rolling around in useless hoarded piles of Gold: if anything, that'd be more challenging and fun!

I'll pay more attention to what's going on with both of these the next few games I play. (Thanks, InfoAddict!)

 

[Seek]

Possible Solution

Now consider that if we double all science income and double all tech costs, nothing really changes, right? We just make all the numbers bigger. So let's consider what happens if we apply something like this to science. Consider a similar city to before:

24 = 2 * 12 pop
12 = 1 * 12 library
03 = 3 * 01 scientist
====================
39 / 12 = 3.2 per
03 / 21 = 8% from specialists

The final values of science are almost identical to those for gold, and very close to the goal of 3 yield per pop.

What are your thoughts on this analysis? Do you have any questions, or see anything I might have overlooked?

I'm not sure what this would accomplish - wouldn't it be simpler to just make Scientist generate 2? It would remain higher than the Merchant in terms of percentage, but we can also adjust Merchants up to 4 as I suggested in the Specialists thread.

 

[Thalassicus]

It often takes me a few attempts to figure out a good way to explain things, sorry about that!

You suggested these numbers, right?

3 Engineers
4 Merchants
2 Scientists
2 Artists

If 3 is overpowered compared to 3... then 1 is overpowered to 1. It's the ratio that's imbalanced. Does this make sense?

 

[Seek]

That's one way of looking at it, but I think in game it's not so simple - it's not that Engineers feel overpowered but that Merchants feel underpowered, and these are not the same thing. A lot of it has to do with tile yields: One Engineer = one non-riverside mine, but one Merchant < one village because villages have two other yields depending on terrain.

 

[Zaldron]

One possible problem with your analysis is % multiplier buildings such as the university. Just adding that single building (which isn't THAT late in the game) to your example changes the equation to 28 science / 12 pop = 2.33 sci/pop. The weird interaction between pop based and multiplier based science buildings vs the strictly multiplier based market and bank will make it pretty tough to balance this overall, for a variety of population values.

Did you try doing your comparison without *any* buildings at all?

You'd also have to make scientists' bulb more powerful if each tech costs more.

 

[Zaldron]

That's one way of looking at it, but I think in game it's not so simple - it's not that Engineers feel overpowered but that Merchants feel underpowered, and these are not the same thing. A lot of it has to do with tile yields: One Engineer = one non-riverside mine, but one Merchant < one village because villages have two other yields depending on terrain.

And not only THAT but a GE can be the difference between getting and missing a key wonder. While a GM can be a flux of money and influence it won't buy you a wonder.

 

[Gamewizard]

If population will no longer equal science, then perhaps each yield should get its own tile improvement. Each building modifier and national wonder modifier would provide equal multipliers, so a 30%university = 30%opera house = 30%bank = 30%(production building). Specialist yields should be lower than the tile equivalent, but GP would make it equally as strong as a tile based economy. I think all tile improvements should adopt the Civ IV cottage method of getting better over time. I think that would balance out the specialist economy because GP get better later in the game. A GP could instant build a max level tile improvement.

I think this method promotes city specialization very well.

I also think social policy picking should be tweaked so that policies deep in a tree cost more than a policy early in a tree. I think this would make a balanced yield Civ = a focused yield Civ. I don't know if the AI would understand this or if it would always pick the cheapest policy available. It should allow a player to hold off on a victory condition until later in the game rather than going science victory from the start, for example.

Another addition I would like to see is a road/railroad equivalent for science and culture, perhaps making telegraph lines a buildable improvement. Or maybe just adding science/culture modifiers to trade routes at a certain tech or policy. This along with every yield having its own tile improvement should help the AI at higher levels since they already receive extra workers to start the game. The AI should start out strong because of this but would eventually get overtaken by a humans ability to specialize cities.


As for the tile improvements, I'm assuming the same thing that was done in Trading Posts Redesigned can be done to all improvements. Something as simple as changing the colors of an improvement as it levels up would be helpful in quickly distingushing a high level improvement from a low level one.

 

[orangecape]

I looked at the costs of RAs in VEM and although they scale with era they don't scale the same way the rewards do. For example, if you pay 600g in classical and 900g in industrial but the total beaker reward goes up from 400 science to 4000 science then the rewards are way out of line with the costs.

The most egregious example of this is vanilla where 250g got you 300 science early in the game and 6k science in the late game. *facepalm*

If you want to make gold as valuable as hammers you need to reduce income or massively increase outlay. For income reduction you could reduce the amount the AI will pay for things or reduce trade route gold, both of these just take gold out of the system without changing specialist/building balance directly) For increasing outlay you could implement more scaling in RA cost; if you want 4k science you should have to fork over 3k gold, say. The other possibility is mucking with the buy cost formula so it no longer has an exponent of .75 that makes buying endgame stuff cheap; you could go with a fixed ratio of 3:1 rushbuycost:hammercost to give people things to spend money on later on in the game.

In VEM I definitely found gold to be abundant to the point of being unimportant - I always found hammers to be vastly superior.

 

[wobuffet]

@orangecape:
I'm not sure if you know this, but Research Agreements have been completely redone in VEM, and the costs are being tweaked as we speak. Likewise, in VEM Thal has completely gotten rid of the rushbuy cost formula and given each buyable unit/building a Cost % (e.g., 310% on a 100 item means rushbuying it would cost 310), and as far as I know, these Cost %s do not systematically go down over time.

 

[Jaybe]

..., and as far as I know, these Cost %s do not systematically go down over time.

As you advance in the tech tree, the cost %s of these later buildings DO go down.

 

[wobuffet]

As you advance in the tech tree, the cost %s of these later buildings DO go down.

Ah. I might be missing something, but why would we ever want that, especially when Gold gets more and more plentiful as the game goes on?

 

[Jaybe]

Next time you have the tech tree open, do a comparison. With the 'over-abundance' of gold people have been referring to (and typically Persia along with my current v.126 game as Rome), I agree that I would NOT want it. A Thal design decision which I wouldn't mind being reversed.

It also applies to units, IIRC.

 

[wobuffet]

My guess is that it wasn't really a Thal design decision, and that he pretty much kept costs in line with what Firaxis had in the vanilla game (whose inelegant and hidden hurry cost multipliers and such VEM makes transparent or eliminates).

My guess would be that the scaling-down of / costs better keeps -to-actual building yields in line, since costs increase more than linearly to account for % production bonuses and such. In any case, I'd strongly support harmonizing the Cost %s over time.

If this ends up having unintended side effects (I can see it encouraging wide empires with very few, -cheap buildings each, for example), these can be dealt with afterwards.

 

[Thalassicus]

[...] reduce trade route gold, both of these just take gold out of the system without changing specialist/building balance directly

I hadn't thought of this! This is a great idea. Trade routes depend on the number of cities and population we have, so it's probably a major factor in late game gold excess.

@orangecape, wobuffet
Purchase cost has three parts in vanilla:

1. Base cost.
2. Modifier depending on #1.
3. Modifier depending on place in tech tree.

Vem eliminated 3, leaving only 1 and 2.

The tricky part is #2 determines which cities we purchase things in. Undeveloped cities have low production, so they're a high priority for purchasing. The purchase discount on later buildings raises the priority of developed cities. Without that modifier, we would always purchase in small cities, and always build in large ones. So removing #2 would shifts the fun scale in two ways:

• Right shift - yield to focus on (production overpowered &#8594; more equal).
• Left shift - city to buy things in (small &#8592; large).

Two shifts in opposite directions might not have a net effect on the game. This is an option, but I'd like to first pursue ideas like the trade route gold idea, with less of a counteracting effect.

 

[orangecape]

@wobuffet

Yes, I know RAs in VEM are quite different, I think I just didn't articulate clearly enough. What I meant to communicate is that we can estimate the amount of science we get from an RA at any given point. We can compare that to the cost of the RA and figure out how many beakers we are getting per gold. Obviously in vanilla you are getting half a tech which ranges from 150 beakers to 5000 beakers at the fixed price of 250g, which is crazy.

I don't know the math behind RAs in VEM but I understand the principle - it is a % increase to science based on both civ's current science rate. You can still look at average science creation in a given era and figure out how much science you get. *Very* rough estimate:

100 beakers/turn in medieval.
1000 beakers/turn in industrial.

Therefore the RA cost in industrial should be 10 times that in medieval.

I don't know how much science you actually get, mind, but I do think that the proportions should be equivalent. I assume Thal has the math for how much you get from RAs exactly and that would allow Thal to set costing appropriately. This would mean that you can cost RAs appropriately for the benefit you get from them. I would think that you would want to get more than 1 beaker/gold because investment should have a positive return but honestly I don't know what the return should be.

 

[Zaldron]

I hadn't thought of this! This is a great idea. Trade routes depend on the number of cities and population we have, so it's probably a major factor in late game gold excess.

This could work well, but be warry of unintended consequences to tall empire gold generation (which is already tight).

 

[wobuffet]

Reducing trade route income would be great. Offhand, maybe halve both road/raliroad maintenance cost and trade route income?

This could work well, but be warry of unintended consequences to tall empire gold generation (which is already tight).

Reducing trade route income would hurt wide empires more than tall ones, no?