A Discussion of Unit Maintenance Costs

[L. Vern]

Hello all,
the topic of discussion is GPT unit maintenance costs: as it currently stands it's a fairly cryptic system that calculates maintenance based on number of techs researched and number of units owned. In this post, I'm going to go over how the current system works, what I think the major problems with it are, and suggest a starting point for a simpler and easier to understand system.

Current System
The current unit maintenance costs are calculated as a function of the number of units (that have a nonzero maintenance cost) owned, and the number of technologies owned. When ignoring maintenance discounts and using the constants of the current version, the unit maintenance cost is equivalent to the following function

edit: I realized to better understand this equation it's worth pointing out that the reason I didn't simplify the fractions is that the number 82 is significant - this is the total number of techs in the game

It's a bit difficult to understand what's going on here due to having 2 independent variables in an exponential function, so I've created a visualization of how much the next unit GPT maintenance will cost across 8 values of techs known. I've also confirmed these match ingame for all the combinations I've tried with IGE, so should be a pretty reasonable representation of what we get in real games.

Problems with Current System

1. It is really not very clear. Often I can't tell how much my next unit will cost in maintenance, much less be able to plan or have an idea of what my army maintenance will cost in the future
2. Maintenance costs don't differentiate between the kind of unit - an archer and a stealth bomber both cost the same GPT to maintain, which is both silly as well as limiting to design space for unit identities - cheaper, mass produced, and weaker/lower utility units should cost less to maintain than specialized, powerful, rarer and impactful units
3. Marginal unit costs increase as number of units goes up. Not only does this go against the ideas of economies of scale, we already have a military supply system for limiting army sizes.

Simpler System
Imo, the previous equation should not be used and instead units should just have a flat maintenance cost, per unit. A simple, easy to understand, yet powerful and configurable alternative, such an abstraction is much more understandable, arguably a better representation of reality, and best of all - can be intuited by the average gamer.

Here's a starting point for flat unit costs that make sense to me, would love to get your thoughts and feedback on it and maybe eventually workshop it into a formal proposal?

*current values are calculated using an arbitrary number of total units and techs I felt were reasonable for that point in the game, though you can use the graph above to compare any combination of #units/#techs

Discussion Questions
The main reason for making this post was to gauge community feelings/interest and thoughts on the topic - here's some things in particular I'd like to hear more about

1. Is this actually something that warrants effort into changing and balancing it? I haven't seen a lot of conversation about unit maintenance, not sure if that's because everyone is content with the way it works now or perhaps because it's unclear how it works?
2. Thoughts on what I identified as the largest perceived problems? Do you agree or disagree with some, or are there any positives or negatives to the current system you feel strongly about?
3. How would you feel about a transition to flat, per-unit maintenance paradigm? Is it something you think would improve the game or can you envision problems arising from such a change?
4. In a flat cost model, what are your thoughts on the numbers presented? Are the base costs per era reasonable? Would you like to see more or less spread between different unit types within an era?
5. How would you prefer non-military maintenance be handled? I can make a decent case for it being free, or 1 * era, curious how you would like to see this done
6. Thoughts on unique units getting reduced upkeep or buildings/leader UAs with -1/-2 upkeep for some unit class or using flat gold reduction in some other capacity? Might be an interesting tool for modmodders

 

[Recursive]

How would you feel about a transition to flat, per-unit maintenance paradigm? Is it something you think would improve the game or can you envision problems arising from such a change?

How would we handle "X of your units are maintenance free" rewards?

I do think it would improve the game. Vague and abstract calculations < simple and transparent calculations.

 

[msw1]

This system is so much more intuitive than the current one that I thought it was already part of the game.

Could "X of your units are maintenance free" become "Your X most expensive units are maintenance free"?

 

[Rustonreddit]

I like this idea to change to a flat unit maintenance. If its done it should be visible in the production queue (along with hammer cost, etc) and in the unit description on civilopedia.

 

[L. Vern]

I like this idea to change to a flat unit maintenance. If its done it should be visible in the production queue (along with hammer cost, etc) and in the unit description on civilopedia.

yep I was also thinking that such a change would involve a UI component, definitely in line with the idea of making it clearer and easier to understand how much it will cost to play the unit

 

[Rustonreddit]

I was thinking that it may have some cheese as a side effect: the human player will have an incentive to not upgrade until right the turn before they use them, the AI will probably always upgrade right away.

But that is also the case with producing military units (gold sink while they are not at use), and the human player want to be ready for surprise wars and also appear strong in the eyes of the AIs, so maybe is not that bad.

edit: there is already an incentive to not upgrade the units until you use them: you want to use the gold in other things. so i think is fine

 

[CppMaster]

This system is so much more intuitive than the current one that I thought it was already part of the game.

Could "X of your units are maintenance free" become "Your X most expensive units are maintenance free"?

Haha, yeah, me too

@L. Vern Very cool idea! You have my vote if magi don't want to fast track it!

A simpler idea would be if unit cost maintenance = unit era, so 1 for ancient, 2 for classical etc.

 

[Agiwan]

I like the idea, we could even have UU with more or less gpt cost as an advantage or disadvantage over the unit it replaces.

 

[axatin]

It is really not very clear. Often I can't tell how much my next unit will cost in maintenance, much less be able to plan or have an idea of what my army maintenance will cost in the future

Yes, this is a problem. However, I think the rationale behind the current formula is sound, and the opaqueness can be mitigated by simply adding a text like "Producing one additional unit would increase unit maintenance costs by 6.53 gold." in the gold tooltip.

Maintenance costs don't differentiate between the kind of unit - an archer and a stealth bomber both cost the same GPT to maintain, which is both silly as well as limiting to design space for unit identities - cheaper, mass produced, and weaker/lower utility units should cost less to maintain than specialized, powerful, rarer and impactful units

It's already one of the more unrealistic parts of the game that you can have spearmen in Industrial and use them to garrison a city. There shouldn't be any kind of reward for not upgrading units. The strategy of using mass-produced expendable units in a war (like lancers in late industrial/early modern) is already implemented by having low hammer costs for those units.

Marginal unit costs increase as number of units goes up. Not only does this go against the ideas of economies of scale, we already have a military supply system for limiting army sizes.

It seems the ship has sailed here, but I think unit maintenance costs are a much better approach to limiting the number of units than unit supply. The supply system with its hard cap feels very artificial: I have a supply cap of, say, 52, then there's no difference between having 51 units and having 52 units, but once I have 53 units, I get a gold and production penalty. It is very much a win-more system (at least before the rework; maybe it's better now): Have a larger empire and you can have more units. The supply system gives no incentive of any kind to stay below the cap, so effectively it just tells you how many units you are supposed to have at a given point in time, which is boring and non-strategic. With unit maintenance costs, on the other hand, there's a trade-off and a strategic decision involved: Focus on a large army to conquer the world and you don't have money left for buildings. Or focus on infrastructure and maintain only a small army for defensive purposes. Or use the gold for diplomatic units, again leaving less money for other things...

For unit maintenance costs to serve as a limiting factor it's necessary that the costs per unit increase with the number of units: The larger your army is, the more difficult it is to make it even larger.

TL;DR I agree the opaqueness of the formula is a problem, but I think the two underlying design principles of a) having per-unit costs that don't depend on unit type and b) having a superlinear scaling in the total unit costs are reasonable and should not be changed.

 

[azum4roll]

Maintenance costs should be consistent on each tech tier, with exceptions on "premium" units (tanks and carriers, nothing else), and UUs that unlock earlier (maintenance costs should follow the original unlock tech unless otherwise specified).

There also still needs to be a growing maintenance modifier per supply, e.g. having 53 units that cost supply increases total unit maintenance by +26.5%.

This opens up some modding potential, like UUs with cheaper maintenance and buildings that reduce a certain unit combat's maintenance by a flat amount.

 

[CppMaster]

A combination of both solutions could be considered: Keep the superlinear scaller based on total supply and have units take supply based on their era. So ancient era units would take 1 supply, classical era units would take 2 supply etc. Obviously, numbers for the scaler and cap limit would need to be changed.

 

[CppMaster]

For unit maintenance costs to serve as a limiting factor it's necessary that the costs per unit increase with the number of units: The larger your army is, the more difficult it is to make it even larger.

In theory yes, but in practice I usually have plenty of gold for both buildings and units.

 

[Pipiskus]

I feel like this change would collide with UUs in modded civs, so all of them would need to be updated to match this new philosophy, so iam not very supportive of this proposal. if you want to change maintenance cost that much, work with existing equation, why inventing the whole new system?

You can just do something like this as one man suggested:

Yes, this is a problem. However, I think the rationale behind the current formula is sound, and the opaqueness can be mitigated by simply adding a text like "Producing one additional unit would increase unit maintenance costs by 6.53 gold." in the gold tooltip.

P.S Or you see the way for it to work with modded civs? I may not know something but as ive heard new UUs and UBs are all new units and building and they don't truly replace anything so you would need to attach this new maintenance cost system to every unit, it would not apply automatically, right? This is my main objection if its true.

 

[ma_kuh]

I like that flat maintenance costs are a better parallel to building maintenance costs, which do not have formulaic calculations.

I was going to say I also prefer a system where strategic resource = more maintenance, but on second read that's basically what you've already laid out, with the exception of archer-line units which are not as cheap as melee-line units. I would maybe put archers and spears/tercio+ in the same maintenance category, but otherwise I like the proposal.

I'm personally a fan of the emergent behavior around Warcraft III's upkeep system, where going over 50% supply results in upkeep costs, and going over 80% supply results in heavy upkeep costs. If you keep your army light, you get a heavy economy advantage, at the risk of being taken by a surprise fight.

 

[hokath]

To Axatin:

a) having per-unit costs that don't depend on unit type

So the hybrid suggestion would be when you hit a new Era the base cost of all units goes up
Then to get differing costs for better units they would have specific exceptions. I.e. a cannon would say on its tooltip "Costs 1 extra Maintenance"

b) having a superlinear scaling in the total unit costs are reasonable and should not be changed.

As you discussed, it is somewhat redundant now we have supply. To me, having both systems is overkill.

and on that point to Azum:

There also still needs to be a growing maintenance modifier per supply, e.g. having 53 units that cost supply increases total unit maintenance by +26.5%.

Could you speak more about why there "needs to be" this mechanic?

I'm personally a fan of the emergent behavior around Warcraft III's upkeep system, where going over 50% supply results in upkeep costs, and going over 80% supply results in heavy upkeep costs. If you keep your army light, you get a heavy economy advantage, at the risk of being taken by a surprise fight.

Zug zug.
I suppose we could consider this. It requires AI work though. The strategy is to stay under 50% pop and then ramp up for a war -- AI currently just maxes out supply.

 

[ma_kuh]

Zug zug.
I suppose we could consider this. It requires AI work though. The strategy is to stay under 50% pop and then ramp up for a war -- AI currently just maxes out supply.

Okie dokie.

The Civ model is already sort of like the inverse of this, where you have no penalties up to current supply, and once you go above you start incurring eco penalties. The difference I guess would be that being 101%-150% supply should have the same %increase, instead of continuing to get worse. And the AI problem would be flipped, aggro civs would need to be taught it's okay to go over-supply if they're in a war or preparing for a war. Everyone else would probably rightly be on the eco gameplan and stick to up-to-100% supply.

 

[CppMaster]

Nah, those artificial steps seems gamey. I prefer the steady increase

 

[ma_kuh]

You can't play around smooth scaling though, the next one is just "a little more" no matter where you are on the curve.

Steps offer trade-offs, they give you something to play around and optimize.

Just a player preference, though, I wouldn't say either is better design, strictly speaking. They just make you play the game (optimally) differently.

 

[CppMaster]

You can't play around smooth scaling though, the next one is just "a little more" no matter where you are on the curve.

Steps offer trade-offs, they give you something to play around and optimize.

Precisely why it's gamey. Smooth scaling is more natural.

 

[L. Vern]

The supply system with its hard cap feels very artificial: I have a supply cap of, say, 52, then there's no difference between having 51 units and having 52 units, but once I have 53 units, I get a gold and production penalty.

Interestingly, I had the exact opposite reaction to seeing this mechanic - I think it's a great abstraction of war economies:

If your economy has a productive capacity that requires one million workers, your country has two million able-bodied men, you basically have a military supply cap of a million people. At 900k and 1m military, there's no difference to your economy, but once you go up to 1.1 mil your country will indeed be less productive and generate less economic value due to less populace available to work in your countries economy, which is totally in line with -% /