When should I stop building Civs? A: Don't build them.

[personwithhat]

Heyoo,

Attached is a quick and (I hope) logical breakdown of Civ's in Hoi4.

Lemme know if anyone see's logical issues or has other tips to keep in mind.

I welcome ya'll to do test runs and compare.

The basic result, is that building Civs then Mils (Case 1) is worse than building only Mils (Case 2).
Well, until 1944 or thereabouts, but most wars end before then.

==========================
Addendum, quick demonstration
According to graph, should break even in about 5.5 years if Mil construction bonus is 1.2

Scenario:

• Japan
• 90 Civs
• All infra at 100%, own China/Machuko.
• Basic tools and up to Concentrated Industry IV (For the factory slots)
• Partial Mobilization as economy law
• Limited Exports as trade law
• Total of 10% construction bonus for Civs
  Total of 20% construction bonus for Mils
• All Military factories on Gun_I

======== CASE ONE ========
Build 3 Civs, 15 builders on each. Done by March 11, 1936
Everything else (and after) is building Mils.
January 1938 - 478K guns
January 1939 - 799K guns
January 1940 - 1177K guns
January 1941 - 1607K guns
January 1942 - 2082K guns

======== CASE TWO ========
Build only Mils
January 1938 - 490K guns
January 1939 - 815K guns
January 1940 - 1193K guns
January 1941 - 1620K guns
January 1942 - 2091K guns <--- Still 9K guns behind.

(FYI 2147K is largest stockpile, it caps out)

RESULT: Case Two wins, as expected.

 

[KDEstroy]

Hardcore players: Hoi4 is dumbed down!
Hoi4: literally requires players to solve differential equations in order to calculate the optimal build order

I think most of your simplifications are reasonable, except for this one point: " Resource limitations can be safely ignored as they have no direct effect, andare typically resolved through alternative means such as infrastructure/focuses" If you actually play as any country except US or France and build nothing but military factories, you will run out of resources very quickly and be forced to rely on trading, which reduces construction time further. For Germany you will start to run out of steel after about 100 Mil factories. For UK, even less. Building more infrastructure only helps to a certain extent. Sooner or later you will be forced to trade.

 

[DGuller]

I'm not sure I 100% understand the math, but it seems like you're ignoring all the maintenance costs of MILs, such as resources and consumer goods. Just like building CIVs forgoes immediate military production in exchange for having greater production from more MILs down the road, building MILs reduces your capacity to build MILs later on. Having to buy resources and cover the consumer goods requirement can be a real drag on your economy.

I also find it hard to see how there can be no break-even point, as it seems like a certainty that starting off with building a CIV will eventually leave you with more MILs at some point. That point may be 2100, but there has to be that point.

 

[personwithhat]

Hardcore players: Hoi4 is dumbed down!
Hoi4: literally requires players to solve differential equations in order to calculate the optimal build order

I think most of your simplifications are reasonable, except for this one point: " Resource limitations can be safely ignored as they have no direct effect, andare typically resolved through alternative means such as infrastructure/focuses" If you actually play as any country except US or France and build nothing but military factories, you will run out of resources very quickly and be forced to rely on trading, which reduces construction time further. For Germany you will start to run out of steel after about 100 Mil factories. For UK, even less. Building more infrastructure only helps to a certain extent. Sooner or later you will be forced to trade.

Infrastructure is a better payoff
Through numerous runs on e.g. Japan, you can generally get more than enough steel/aluminum/tungsten by appropriately timing your infra/focuses

There are some ofc that are quite difficult .e.g. Australia

Regardless, having fewer factories for longer actually makes building straight mills a better option!
Less factories == Less resources necessary
You maintain similar (if not better) output by just outputting earlier.

Mayhap you are running out of resources, because you are spamming a bunch of Mils later?

I'm not sure I 100% understand the math, but it seems like you're ignoring all the maintenance costs of MILs, such as resources and consumer goods. Just like building CIVs forgoes immediate military production in exchange for having greater production from more MILs down the road, building MILs reduces your capacity to build MILs later on. Having to buy resources and cover the consumer goods requirement can be a real drag on your economy.

I also find it hard to see how there can be no break-even point, as it seems like a certainty that starting off with building a CIV will eventually leave you with more MILs at some point. That point may be 2100, but there has to be that point.

The break even point tends to come 5+ years after you finish building the Civ.
Considering most major wars start in 1941 (a mere 5 years after Day 0), most of your Civs would not pay off in a reasonable amount of time.

 

[blahmaster6k]

I beg to differ. With each assumption you make, your calculation necessarily becomes more inaccurate, particularly (2) and (3), and to a lesser extent (4). Consumer goods are especially relevant because the number of factories required for consumer goods are a function of your total factories, not just your civilian factories. Likewise with resources available. If you build only military factories, your number of civilian factories available for construction and trade will shrink dramatically due to consumer goods, even as the amount of civilian factories required to trade for resources to supply your new military factories goes through the roof. To me, the combination of these assumptions makes the entire calculation irrelevant, as it's basically a vacuum so wildly different from what will happen in an actual game as to be completely useless. I'm sure you're good at math, but you can do all the calculus you want and the classic rule of garbage in, garbage out makes all that time and effort wasted.

Honestly, this whole post seems like an ill-conceived I-am-very-smart flex, because I'm pretty sure 99% of the people on this forum have no idea what a differential equation is, let alone a basic integral. If you don't have a STEM degree you likely have never seen these symbols in your entire life. What good is it to put all this math out there backing up calculations if people have no idea what any of it means? They just have to trust that your conclusions are correct, which is bad because as I noted above, they aren't.

 

[personwithhat]

I beg to differ. With each assumption you make, your calculation necessarily becomes more inaccurate, particularly (2) and (3), and to a lesser extent (4). Consumer goods are especially relevant because the number of factories required for consumer goods are a function of your total factories, not just your civilian factories. Likewise with resources available. If you build only military factories, your number of civilian factories available for construction and trade will shrink dramatically due to consumer goods, even as the amount of civilian factories required to trade for resources to supply your new military factories goes through the roof.

Ok, consider case #1 then.

The main concept behind a Civ is that it pays for itself in the future by speeding up future military builds.
Ofc, you can build 1.5 military factories quite a bit earlier than the Civ can build 1.5 factories.
To make up for the lost time, the civ would have to build more than 1.5 Mils right?

These require more resources, and each successive Mil will also penalize your consumer goods.
Since the whole idea behind Case #1 paying off hinges on the fact that you need to build more factories than in Case#2, consumer goods are actually detrimental.

=========================
Addendum:
Regardless, if you are building Civs for resources..... you are building them for resources.
You are not building them to kickstart your Economy to have a economic boom later on.
That would be working within a resource limitation which is country specific.
You would predict how many resources you would need, that you cant get from infra, and build that many Civs as early as possible to get some of your investment returned.

 

[StigOHara]

Doesn't this question depend on what country you start as and to an extent which playstyle (tall/wide)?

I wouldn't strictly build mills at the start of the game unless I'm going for an aggressive war/ world conquest at early game. Also I dunno what im even reading past the first page.

 

[personwithhat]

Doesn't this question depend on what country you start as and to an extent which playstyle (tall/wide)?
I wouldn't strictly build mills at the start of the game unless I'm going for an aggressive war/ world conquest at early game. Also I dunno what im even reading past the first page.

The point here is, if you are planning on having a war before 1944, then you should build Mils instead of Civs in most cases.
The exceptions being when you can't support the Military factories later on
In which case you should build some civs (as few as possible) and then spam Mils as per usual.

Past the first page:
Just finding the break-even point where Case #1 (Building 1 civ that then builds X mils) and Case #2 (Building X mils instead) have the same total military output.

 

[blahmaster6k]

Ok, consider case #1 then.

The main concept behind a Civ is that it pays for itself in the future by speeding up future military builds.
Ofc, you can build 1.5 military factories quite a bit earlier than the Civ can build 1.5 factories.
To make up for the lost time, the civ would have to build more than 1.5 Mils right?

These require more resources, and each successive Mil will also penalize your consumer goods.
Since the whole idea behind Case #1 paying off hinges on the fact that you need to build more factories than in Case#2, consumer goods are actually detrimental.

=========================
Addendum:
Regardless, if you are building Civs for resources..... you are building them for resources.
You are not building them to kickstart your Economy to have a economic boom later on.
That would be working within a resource limitation which is country specific.
You would predict how many resources you would need, that you cant get from infra, and build that many Civs as early as possible to get some of your investment returned.

Building infrastructure at all is a waste except for when you've exhausted all building slots, because it's time spent not building factories. The thing about CIC is they produce exponential growth, where building MIC produces linear growth. Every CIC you build is able to start building even more CIC as soon as it's finished. Compare the factory count in June 1941 of a Russia player who does no ahistorical conquests and builds MIC from day 1, and a Russia player who builds CIC until Molotov (july 1939ish). The latter will have 200+ CIC available to crank out military factories by the dozen when he switches over, and can have 300+ MIC by 1942. The former will be using his starting factories building 2 military factories at a time the entire game, and be stuck on around 60 CIC and maybe 150 MIC if you're pushing it. It's almost never worth building military factories from day 1, unless you have (a)very few building slots, (b)very few factories, or (c) plan to go to war in 1936. Regarding (c), early wars are way more punishing now because of the resistance/compliance mechanics that it's typically better to build up your own country rather than get the nerfed factories via conquest and the equipment drain of resistance that comes with it.

I don't want to pull the "experienced player card" here, but it's universally agreed upon by anyone who's played significant hours, especially in multiplayer, that it's better to build civilian factories for a few years, typically until 1-2 years before you expect to go to war. Theoretical exercises can only get you so far, collective hours of many players playing the game at a high level are what you want to look at to determine what the optimal strategy is.

 

[personwithhat]

Factory count is a poor indication of total output. Just because you have more mils at year X doesn't mean you've produced more guns by year X.

Infrastructure in (most cases) does not pay for itself, but that's a separate subject (Mostly since infrastructure does not get infrastructure construction multiplier)
The only reason I brought it up was to address resource limitations. If you need (and can get) 30 more aluminum by March, you start building 2+ infra in January.
It's just cheaper than selling Civs for resources.

I would advise you to read more carefully and spend some time internalizing the information before responding with nonsense.
Try doing a few test runs to refute my theory, so far mine support it.

 

[Roland Traveler]

In addition to the other points being made, this also ignores other important points, like reduced CIC capacity means building up airfields and ports (if necessary) will be slower later on, as will any repairs while also gimping long-term production of equipment.

Let’s say that you build a total of 60 mils before CG costs cripple your construction capacity. Congratulations, you’ve now got 60 miles constructing equipment that will be obsolete by the time the war starts and that will have to cannibalize itself to either increase production of certain equipment or start production of new ones. Meanwhile someone that invests in CIC will be able to keep building after 60 mils, allowing them to increase production on individual lines or start new lines without needing to cut into other ones. And considering that medium tanks are considered among the best, being able to invest in that while also being able to build up your air force without sacrificing one for the other certainly makes it seem like planning for the medium to long term is superior to planning for the short term.

Now if you were playing China, Poland, France, or a Czechoslovak planning to tell Germany to go screw itself, then you would have a point about prioritizing mils over CIC, but for most countries the war doesn’t end in 1940, meaning they have to think long-term instead of short-term

 

[personwithhat]

I can concur on the reduction in Airfield and port/etc. capability, as well as repair. In that case the math is quite different, as an airport does not produce goods, and has fixed value indeterminate of the stage of the game.

By 'long-term' I assume you mean 1944? Otherwise, as noted before, the total output would be lower if you build Civs (case1) as compared to Mils (case 2)
You would not be 'Sacrificing' anything. Your total production output is higher than any production you would have had if you had built Civs, until the break even point. Which is significantly past 1942, especially if you build Civs for years as recommended in an earlier point.

Let’s say that you build a total of 60 mils before CG costs cripple your construction capacity.

Run some test cases before throwing numbers around with no rhyme nor reason.

 

[KubiG37]

I've done a similar testing (For different purpose) on the soviet AI, in my mod the AI builds only CIVs at first until certain threshold, vanilla just keeps 1:1 ratio.
I was testing the overall effect of Soviet production on its army and war against Germany. (Just once though)

Differences:
Main soviet template in my mod is also different, it is 10 INF + support arty (vanilla is 9 INF + 3 Art)
SOV AI builds more airplanes
SOV AI builds infantry weapons 2 (not just 1 like vanilla)

My test results (historical focus on):

======== CASE ONE ========
Vanilla AI, 1:1 CIV/MIL Ratio

April 1938 - 63 CIV 64 MIL // ___Aircraft: 2003 Infantry EQ: 135000 Art: 3180 // Army size(eq.level): (?) _________//Surrender progress: 0%
April 1939 - 72 CIV 70 MIL // __ Aircraft: 2700 Infantry EQ: 157000 Art: 5560 // Army size(eq.level): ? (?)________// Surrender progress: 0%
April 1940 - 100 CIV 100 MIL // Aircraft: 4002 Infantry EQ: 179000 Art: 7910 // Army size(eq.level): 1,9M (?)____// Surrender progress: 0%
April 1941 - 124 CIV 136 MIL // Aircraft: 5850 Infantry EQ: 220000 Art: 11900 // Army size(eq.level): 2,4M (100%)// Surrender progress: 0%
April 1942 - 110 CIV 139 MIL // Aircraft: 2400 Infantry EQ: 275000 Art: 8250 // Army size(eq.level): 3,1M (88%) // Surrender progress: 0%
April 1943 - 109 CIV 114 MIL // Aircraft: 1200 Infantry EQ: 137000 Art: 8000 // Army size(eq.level): 2,6M (56%) // Surrender progress: 52%

======== CASE TWO ========
AI first builds to a total of 90 CIVs, when at war AI keeps 1,3 MIL for every CIV,

April 1938 - 89 CIV 49 MIL // ___Aircraft: 2130 Infantry EQ: 144000 Art: 1200 // Army size(eq.level):1,4 M (99%) // Surrender progress: 0%
April 1939 - 93 CIV 74 MIL // ___Aircraft: 2950 Infantry EQ: 175000 Art: 1680 // Army size(eq.level): 1,8M (99%) // Surrender progress: 0%
April 1940 - 128 CIV 110 MIL // Aircraft: 4650 Infantry EQ: 216000 Art: 2120 // Army size(eq.level): 2,2M (99%) // Surrender progress: 0%
April 1941 - 131 CIV 140 MIL // Aircraft: 7260 Infantry EQ: 319000 Art: 3350 // Army size(eq.level): 3,3M (99%) // Surrender progress: 0%
April 1942 - 146 CIV 181 MIL // Aircraft: 4000 Infantry EQ: 325000 Art: 2000 // Army size(eq.level): 4,0M (82%) // Surrender progress: 0%
April 1943 - 137 CIV 227 MIL // Aircraft: 2430 Infantry EQ: 218000 Art: 983 // Army size(eq.level): 3,5M (62%) // Surrender progress: 5%


Note: They results are by no means exact, it's affected by trade, different templates requiring different amount of equipment, Lend Lease, war going bad/good etc... but I think the result is still pretty clear.

You can see a lot of things there, mainly how bad the vanilla AI is. But one thing - the AI already has more CIVs and MILs in 04-1939 , after just 3 years.

 

[personwithhat]

The number of factories at a given point is irrelevant. The total production output is the comparison point.
I appreciate the effort, but if you want to do a test case make it in an organized and concise manner.
What you've shown is, unfortunately, pointless due to the inconsistency and the impossibility of replication.

 

[Roland Traveler]

By 'long-term' I assume you mean 1944?

Around there, yes. For me early game is 1936-1939/40, mid game is 1940-1943/44, and late game is 1944+. In my experience, most games end in the mid to late game, meaning that a concentration on what maximizes immediate production isn’t all that useful.

Otherwise, as noted before, the total output would be lower if you build Civs (case1) as compared to Mils (case 2)
You would not be 'Sacrificing' anything.

Most games don’t end at 1940. Do you not understand that?

Your total production output is higher than any production you would have had if you had built Civs, until the break even point. Which is significantly past 1942, especially if you build Civs for years as recommended in an earlier point.

As I said, it’s useful if looking at countries where you’re planning on either winning or losing early, but that is not most countries. Just Barbarossa and the Pacific War start in 1941, putting them entirely in the mid-game. Unless you’re doing gamey strategies and/or rushing war, Britain will still be fighting the Axis after 1940. D-Day will rarely be happening before 1942, and the African campaign will take place in 1940 to 1942 while Italy will be invaded shortly after an Allied victory there. The fact is concentrating on what gives you more early game production as compared to mid to late game production is missing out on when most of the game actually takes place. The fact is the game will most likely drag on into the mid-game and will reward planning accordingly instead of forcing yourself to rely on occupations to increase your factory count due to over investing in mils early on.

Run some test cases before throwing numbers around with no rhyme nor reason.

How’s 4,226 hours of playtime for “test cases”? I know what I’m talking about because I’ve played a short-term game, long-term game, and everything in between. I’ve had wars in every game year between 1936 (Rhineland war) and 1951 (USSR vs max buffed Axis dominating Europe, Asia, and Africa). You got a problem with my hypothetical? Then say what it is. Point out where it’s flawed, don’t just go “Play some games first, then come back.”
Besides that, it’s a hypothetical, it’s not based on a strict in-game model but an abstracted version that controls for everything. It’s the exact same as your model since both are based on in-game experiences but neither 100% accurately simulate it. Mine just happens to be based off a lot more in-practice data than a few highly controlled tests that only look at a set amount of data points.

 

[blahmaster6k]

As @Roland Traveler said quite well above, what's the point of having higher total production output in the early game if you haven't even researched good tech yet?

Total production output is not the comparison point, it's a meaningless stat. If you have 60 military factories in 1938 and I have 30, we've produced the exact same number of 1941 medium tanks - zero. If I have 150 military factories in 1939 and you have 75, and we both started producing 1941 medium tanks somewhere in early 1939, I'm going to have twice as many 1941 medium tanks as you. If we both start producing 1940 fighters in 1939 and I have twice your factory count, I'm going to have twice as many 1940 fighters as you, regardless of whether you had more military factories in the early game, because production output in the early game doesn't matter very much. Sure, have all the 1936 fighters and guns you want, you're not going to be very effective with them once you fight someone with twice your industry producing more modern equipment. And I have no idea where you get the idea that it's more worth it to build infrastructure than trade away factories for resources. You're spending months wasting 15 CIC to boost infrastructure in one state which will typically give you anywhere from 1-20 more resources when built up to level 10. Or, you could've been spending that time building even more factories with your CIC and trading away a single CIC for 8 of a resource.

Also, please cut it out with the holier-than-thou attitude. You can't just dismiss other people's analysis because you disagree with it or because they didn't use any differential equations and meet your ridiculous standards. People don't have to cite someone's masters thesis when they've played the game and have experience doing both of the things you're comparing. I've built mils only as Germany before, and because I've done that I'm aware that it's extremely suboptimal compared to a CIC buildup. Same for the other people commenting in this thread, I'm sure. Have you played for thousands of hours? Many of the people on this forum have. @Roland Traveler has 4226 as evidenced by his above post, I have over 2000 myself. We know what we're talking about.

 

[personwithhat]

In my experience, MP matches tend to pitter out/end by 1942 and seldom go past that.

Building a civ in 1936 will benefit you by 1944 or later.
Building a civ in 1937 will benefit you by 1945 or later

In a competitive scenario I presume you'd want more guns/etc. at 1941-1942 than more guns at 1945+
Besides, you could always do AA_I, support equipment, motorized, scout planes, light tanks for armored recon, etc.

Infrastructure is 3000. A Civ is 10800. Even counting the different bonuses applied, 2 infrastructure (properly placed) grants you more than 8 resources, and is hence more cost effective. (Also affected by extraction tech, while traded goods are not)

Regarding technological differences balancing out production worth, this needs to be handled on a case by case basis and relevant to your game plan.
Sometimes you need hundreds of crappy garrisons using any gun. Other's you want an elite cohort. Having more units is generally favorable compared to having less.
Since the scope and effect is situational, I've discounted it from the current calculations, and addressed every other possible variable.
I do agree that it's a factor that makes Civ's more favorable, however the current payoff is so bad that it's not enough.

Your 4.2k hours of gameplay have no bearing on the discussion. They are not test cases, the actions and hence results taken differ match to match (as they should in a strategy game). The ability to reproduce a test result is necessary to corroborate/disprove any theory.

Sorry if I come across as 'holier-than-thou'. I've attempted to make my PDF as readable for any casual viewer, and am very open to any suggestions.
I've yet to see an analysis. Lots of opinions, and disparaging comments sure.

 

[Zauberelefant]

Heyoo,

Attached is a quick and (I hope) logical breakdown of Civ's in Hoi4.

Lemme know if anyone see's logical issues or has other tips to keep in mind.

I welcome ya'll to do test runs and compare.

The basic result, is that building Civs then Mils (Case 1) is worse than building only Mils (Case 2).
Well, until 1944 or thereabouts, but most wars end before then.

==========================
Addendum, quick demonstration
According to graph, should break even in about 5.5 years if Mil construction bonus is 1.2

Scenario:

• Japan
• 90 Civs
• All infra at 100%, own China/Machuko.
• Basic tools and up to Concentrated Industry IV (For the factory slots)
• Partial Mobilization as economy law
• Limited Exports as trade law
• Total of 10% construction bonus for Civs
  Total of 20% construction bonus for Mils
• All Military factories on Gun_I

======== CASE ONE ========
Build 3 Civs, 15 builders on each. Done by March 11, 1936
Everything else (and after) is building Mils.
January 1938 - 478K guns
January 1939 - 799K guns
January 1940 - 1177K guns
January 1941 - 1607K guns
January 1942 - 2082K guns

======== CASE TWO ========
Build only Mils
January 1938 - 490K guns
January 1939 - 815K guns
January 1940 - 1193K guns
January 1941 - 1620K guns
January 1942 - 2091K guns <--- Still 9K guns behind.

(FYI 2147K is largest stockpile, it caps out)

RESULT: Case Two wins, as expected.

You are, surprisingly, not the first person to discuss this.
Secret Master and myself, amongst others, had the latest long threads on the topic, and the conclusion is:

Building CIC is clever if you're planning to still fight in 1942. How long exactly depends, but for Germany, 2 years is enough.
Foregoing CIC is smarter if you're conceivably done conquering by 41.

I can assure you that these tests were done in live Games and discussed by the Community.
One glaring issue with your metric is that inf Equipment can Rack Up max efficiency way earlier than say 1941 mediums or 1940 fighters. But these win wars, while the production of rifles has little predictive value.

The next is the neglecting of military infrastructure, like air fields, radar, AA, fuel silos.

Then you only Talk about domestically available resources, steel and alu, but neglect the important rubber and oil (and tungsten). But these are needed for a war Industry.

So, given the obvious shortcomings of your test Setup, you can cut the Attitude and start actually listening to people with experience. Then, please apply your skills to contribute for a better collective understanding of the issue.

 

[Praetori]

Heyoo,


RESULT: Case Two wins, as expected.

Nice PDF. It helps to illustrate how you've reasoned. If the goal is to simply build as much equipment of a specific type by a specific date early 40's you're spot on. Going heavy on MICs can put you ahead in terms of equipment in the field. However you also need A LOT of other things that MICs cannot get you.

Japan which you used as an example example needs to conquer resources in order to put MICs to good use (or even keep the armies, fleets and aircraft flying) since she won't be able to trade for them once war is in full swing. She'll need fuel-silos to stock up pre-war on or risk ending up empty for her fleets and air-arm (and then all the infantry-equipment in the world won't help). She'll need NICs to build her a fleet capable of battling the USN etc.

This has all been discussed, at length, before and countless simulations in the live game has been run. The main point is that you can't achieve complete victory/world-domination or whatever goal you're pursuing by 1941-1942 by just building MICs unless you play very specific countries and cheese/exploit the mechanics.
For any "normal" or semi-historical SP or MP game you'll start losing the IC/resource-game somewhere in 1941-1942 if you just build MICs (and you're rarely that much ahead in the strategic sense at that point that it pays off).

 

[personwithhat]

I would think that having more equipment all game up to 1944 would be enough of a reason to motivate most nations to build Mils first.

Again, I agree that your military infra (Fuel silos, air bases, etc.) will be worse off.
Ofc Civs will be better at building air bases......my point here is to prove that it's an economic trade off, more air bases in exchange for less military output.

The general community understanding is that CIC's "Pay for themselves" in 2-3 years and cause economic booms.
A misunderstanding arising from the higher factory numbers, despite having a significantly lower output over the course of the game.

Having fewer factories reduces the resource requirement.