Quantitative Analysis of Civ Traits

[nato]

One of the most important decisions is made before the game even starts: what civ traits to take. This post attempts to quantify the 6 traits and find which are best.

Two warnings:
it is very long
it has to make some assumptions and approximations

That said, I do think it is worthwhile if you are interested.

* If you hate long stuff, skip to the bottom where the nice list is. *

I tried to find the total value of each trait. Generally that value comes down to gold or shields.

You can buy shields at a rate of 1 shield for 4 gold, or you can produce gold at a rate of 1 gold for 4 shields (after economics). Therefore, I am counting 1 gold as of equal worth to 1 shield. When added together I will call these gold/shields.

Tech advances cost beakers, which is essentially gold. I have used the info found by regoarrarr and David Weldon to calculate the gold cost of advances. I have used 24 as the base number.

The conditions of the game are very important. I am guessing the game is 400 turns long on Regent / Monarch. I am assuming a large continents map. I am assuming you eventually get 25 good, low-corruption cities. When it matters, I assume this averages 12 cities of size 6 during the first 100 turns, 12 cities of size 12 and 12 more cities of size 6 the second 100 turns, and 25 cities of size 12 the last 200 turns.

Finally, it is very important that things that come later are less valuable than things that come sooner. Things that come late are discounted in value. How much they are discounted counts on your play style! If you are a rusher, late is useless. If you plan to win late, late coming advantages are discounted less.

Without knowing Civ3's interest rate, I have made no accounting for time. It is very important to remember some of these shields and gold are coming much later, and so are much less valuable.

Ok now the good part:

1. Commercial

a) Corruption: No one seems to know how much corruption is reduced. It seems the effect is rather low, maybe 5%. However, in Civ there are no percents, there is either a whole coin / shield saved or not. My guess is corruption is reduced by about 1 coin per city. If you have 12 cities first 100 turns, and 25 last 300 turns, this is (12 gold * 100 turns) + (25 gold * 300 turns) = 8700 gold.

b) Extra Commerce: You get 1 extra gold for size 6 cities, 2 for size 12. Using what I described above, this is (12 * 100) + ((12 * 100) + (12 * 200)) + (25 * 200) = 9800 gold.

A lot of the benefit of both bonuses comes late game, so keep that discount in mind.

So the total value of Commercial is 8700 + 9800 = 18,500 gold.

2. Industrious

a) Extra Shields: This is identical to the extra commerce in commercial above. Therefore it comes out to 9800 shields. However, extra shields are often irrelevant. Sometimes a city producing one or two less shields will take exact same time to build something. Therefore some of the bonus is lost. It is impossible to know how much. On a guess, I'll reduce the value to 8500 shields. Further, a lot of these shields come very late.

b) Faster Workers: Having double speed workers is somewhat the same as getting a free worker when you build one. What is the value of 1 worker? It costs 10 shields and one population point. City growth would replace the population point; say growth averages 10 turns. Ignoring food, say that pop point would make 1 gold and 1 shield each turn during those 10. Therefore the total cost of a worker is 10 gold and 20 shields. If you make 40 workers in a game, then double speed workers is worth 40 * ( 10 gold + 20 shields) = 400 gold + 800 shields, or 1200 gold/shields.

So the total value of Industrious is 8500 + 1200 = 9700 gold/shields.

3. Religious

a) 1 Turn Anarchy: This saves you the production otherwise lost during anarchy. Non-religious anarchy seems to take 4,6, or 8 turns, so average 6. Religious anarchy takes 1 turn, so each government change saves 5 turns of production. I am assuming 2 government changes, 1 early and 1 later. I looked at a 25 city game and added up the commerce and shields made at an early time and a late time. Early (which wasn't 25 cities yet) was about 30 commerce and 30 shields each turn. For 5 turns that adds up to 150 gold + 150 shields. Later was about 300 gold + 275 shields. For 5 turns that adds up to 1500 gold + 1375 shields. Therefore the anarchy bonus adds up to 1650 gold + 1525 shields, or 3175 gold/shields. More or less revolutions will change this value, as will size of your empire.

b) Half Price Temples and Cathedrals: Ah so straightforward.
25 Temples saving 30 shields each = 750 shields
25 Cathedrals saving 70 shields each = 1750 shields.
Total bonus = 2500 shields. Due to time discount, those temple savings may well equal the cathedral ones.

So the total Religious bonus is 1650 gold + 4025 shields, or 5675 gold/shields.

4. Scientific

a) Three Free Advances: Advances cost a number of beakers = tech cost * base number. I am using base = 24 and picking the lower tech cost advances, since that seems to be what the computer gives you. Also I am ignoring the effect of other civs having the advance. Therefore the value is (24 * 26) + (24 * 84) + (24 * 180) = 624 + 2016 + 4320 = 6960 gold.

b) Half Price Libraries, Universities, and Research Labs:
25 Libraries * 40 = 1000 shields
25 Unis * 80 = 2000 shields
25 Labs * 80 = 2000 shields
Adds up to 5000 shields.

Therefore the total value of Scientific is 6960 gold + 5000 shields, or 11960 gold/shields.

5. Militaristic

a) Higher Chance of Promotions: This has two effects. First is that your units have more hit points. Having an elite spearman instead of a veteran one is somewhat like having 20 % more spearmen, since you go from 4 to 5 hit points. This is therefore somewhat like having 25 % more of the shield cost of your promoted units. Say during each of the 4 ages you get 10 such promotions, and the shield costs of the units for each age averaged 20, 40, 60, and 100 for each age. Therefore the value is (10 units * 0.25 increase * 20 shields) + (10 * 0.25 * 40) + (10 * 0.25 * 60) + (10 * 0.25 * 100) = 50 + 100 + 150 + 250 = 550 shields.

The second effect is getting more leaders. Say militaristic civs get an average of 2 more leaders than normal civs. These will be used to construct wonders, for say a value of 600 + 800 = 1400 shields. Therefore the total value of higher probability promotions is 550 + 1400 = 1950 shields. Obviously if you fight more the value goes up.

b) Half Price Military Buildings: Normally not all 25 cities get all these improvements, so some are for less than 25.
25 Barracks * 20 = 500 shields
10 Walls * 10 = 100 shields
20 Harbors * 40 = 800 shields
20 Coastal Fortresses * 40 = 800 shields
25 SAM Batteries * 40 = 1000 shields
10 Airports * 80 = 800 shields
Total value is 4000 shields.

Therefore the total value of militaristic is 5950 shields.

6. Expansionistic

a) Free Scout, Can Build More: The free Scout is valued at 10 shields. Early on, building Scouts instead of Warriors gives about double the exploring power. Therefore each Scout built is worth the price of an additional Warrior, or an extra 10 shields. Say 4 extra Scouts are built, this is worth 40 shields. Total value of Scouts is then 50 shields.
It is important to note the intagible value of knowing the map earlier.

b) Better Huts: This is the most random factor of all the traits. My worst expansionistic game gave me 1 Warrior and 1 tech advance. My best gave me 3 Warriors, 25 gold, 1 map of region, and 10 tech advances. What is a good amount for an average take? My arbitrary guess is 1 Warrior, 25 gold, and 4 tech advances. The Warrior is 10 shields and 25 gold is, indeed, 25 gold. The tech advances will all be low end ones. They will be worth perhaps (24 * 4) + (24 * 4) + (24 * 5) + (24 * 6) = 96 + 96 + 120 + 144 = 456 gold.

Therefore the total value of expansionistic is 50 shields + 456 gold, or 506 gold/shields. Keep in mind this is all right at the start, so there is practically zero discount due to time.


Total Values in Gold/Shields

1. Commercial = 18,500
2. Scientific = 11,960
3. Industrious = 9,700
4. Militaristic = 5,950
5. Religious = 5,675
6. Expansionistic = 506

A lot of assumptions and outright guesses were made to produce this list, but they are all pretty reasonable assumptions. They must be kept in mind though; I am in no way pretending these values are at all definitive.

Also, once again remember, a lot of these bonuses, especially Commercial coins and Industrious shields, are coming much later; some of them too late to affect the game at all.

This said, if these numbers are anywhere near accurate, they are useful in trying to decide which traits are best.

Thanks for reading!

[Sevorak]

Great post. An excellent effort.

One thing though. When you calculate Industrious' worker bonus as more workers, and Militaristic's unit bonus as more units, shouldn't you factor in the maintenance cost you save as well?

Just a thought.

-Sev

[Grim Legacy]

Another nice effort. Excellent!

Now for some bragging on my part: with my infallible Civ-intuition I had already chosen Commercial and Scientific as the best traits -playing the French.

[Dis]

interesting indeed

although it hasn't changed my views . I still love my aztecs.

but seriously I like military (it is a significant part of my game). Religious is just convienient, not a necessity. I just hate anarchy , and don't feel like f*cking around with it. I tried the romans, but got wiped off the face of the planet with them.

[sophist]

french are not scientific + commercial

French are commercial and industrial.

[Ray K]

Originally posted by Grim Legacy
Another nice effort. Excellent!

Now for some bragging on my part: with my infallible Civ-intuition I had already chosen Commercial and Scientific as the best traits -playing the French.

I was thinking the exact same thing -- except I was playing the Greeks

[randomturn]

Interesting stuff. I think nato's quantitative evaluation is useful to help understand how good a trait is for building purposes. It's important to remember though that building is only a part of the game, and not the most important part. Grim Legacy is right that the Frendch are a great building civ. I would rate militaristic and religious very high in some strategy areas that cannot be quantified. They are particularly strong at higher levels.

[vmxa1]

I want to say that Scientic value looks good, but in reality is much less. This is due to the large percentage of the labs and university. These are not all that commonly built. Labs are likely to be in few cities as they come so late in the game. Universities will be some, but not most cities and in fact even libraries are not in every city. This really blunts its value, although I still like it.
I see expansonist as one of the weakest, extra value from huts, what does it mean, 50 gold instead 25, so what. Scout is of some use.
Military is not bad, but can be done with out.

Religious giving temple discounts is useful.
Now the best two IMHO is Commercial and Industry.
Industry is the best to me, as faster worker will pay off in so many ways and forever. Quick roads to get workers/settlers/troops going, extra shields for large cities to speed production.

[Ray K]

Originally posted by vmxa1
I want to say that Scientic value looks good, but in reality is much less. This is due to the large percentage of the labs and university. These are not all that commonly built. Labs are likely to be in few cities as they come so late in the game. Universities will be some, but not most cities and in fact even libraries are not in every city. This really blunts its value, although I still like it.

I think it depends on your style.

Right now, I rush-build temples, then build libraries, harbors, aqueducts, marketplace, universities in about that order.
About midway through the conquest, courthouses come right after temples and cathedrals come after aqueducts.

Universities are a must-build for all of my cities by the time they're size 12. With all of the rush-building in the newer cities, most of my empire has universities by the Industrial era.

[Arrian]

One man's library is another man's barracks...

It is, in fact, all relative.

Vmxa1 - I play differently from you. Nearly ALL of my cities have both a library and university (if they lack one, they're building it) and later, research labs. The science improvements were very important in CIV I and II, but even more so in CIV III, due to culture.

Nato - good effort. I think the beauty of the game, however, is that your attempt at quantifying (down to gold/shields) the civ attributes was an impossible task. People play differently, each game develops differently, etc. This is a good thing. Someone who values the militaristic trait is probably going to make use of it, and fight, for example.

Personally, I think scientific and religious are the two best, with commercial and industrious coming in 3rd and 4th, respectively. But that's just me, and I can imagine particular situations where the other traits might just be pretty useful.

-Arrian

[mharmless]

I'd have to argue that millitaristic is undervalued here. Having an elite unit is not the same as having 20% more of that kind of unit. First off, I'm not supporting .2 more units, and more importantly I don't lose that unit when it takes four hits.

Next, in regards to leaders. Lets assume you get four leaders over the course of the game. If you get your second arround 1 AD and use him to rush a forbiden palace, then you will be increaseing the output of 10-15 cities from nothing to 20-100% production value. Call it 300 turns earlier than otherwise, and by your numbers they would be all size 12 cities. You list production as 1 sheild 1 gold per citizen when abstracting this, so lets do that.

15 cities X 12 production X 2 (to count gold too) - 2 (existing 1 gold 1 sheild) = 358

358 (gold/sheilds) x 300 (turns) x .6 (average yeild after corruption) = 64,440


And above all, lets not forget the time/value of precious precious wonders. If you get the pyramids as a direct result of a leader, then that leader has cause growth rate to double. Assumeing to goes from 10 turns to 5 (20 food needed to only 10, +2 more than needed to support population), and that you get the pyramids with your first leader, and that you rush build with citizens.

In this case, you are getting some 500 turns of 10-15 cities being able to rush build (40 sheilds) every 5 turns instead of 10. We go from 50x15 rush builds to 100x15 rush builds. This means the pyramids have yeilded you an extra 30,000 sheilds because you got them first as a direct result of being a millitaristic civ.

Or hell, colossus. If you get that from a leader than you are getting 12 extra gold in that city for about 300 turns (average shelf life). 3600 value.

You can see the point. Leaders are extremely valueble and a millitaristic civ will get them more frequently and earlier than a non millitaristic.

[rah]

Good work, but of course I'll add my 2 cents.

Industrial, only is real useful early. Most workers in my empire are chain gang, while my workers are adding to pop in conquered cities.

Religious, I will change goverments quite a few more than 2 times in a game if I only have one turn of anarchy. Great for taking advantage of those war lulls and slipping into democracy, then switching back to com for fighting, or rushing those temples and units after a major offensive. Better to rush buy with those foreign citizens then just wasting them by starving them.

Militeristic, Yes those extra leaders are big. and field promotions early in the game can sometimes spell the difference.

In all, the QUANTIFICATION is good, but some of the assumptions (like the game) could be tweaked depending on how you play. But it is most important that you take max advantage for any characteristics that you choose.

RAH
Good effort.

[nato]

Sevorak - Thanks, I'll try to add in the upkeep costs saved tonight or tomorrow. Good catch.


Grim Legacy - Thanks, though the "infallible Civ-intuition" may be in a bit of trouble!


randomturn - Well I don't generally think of gold for building, I usually use it for science ... I guess commerce would be the more technically correct word than gold, but 1 beaker = 1 gold.

There are definitely unquantifiables ... but I think most things essentially have a gold/shield value. Like in real life, many things are not for sale ... but realistically speaking, you can use dollars as a rough measure for their value.

What I mean is, there must be some amount of gold/shields which I would rather have than the unquantifiable values of a trait ... that amount is their value in gold/shields. Its a decent unit of account.

You're right of course though, this whole thing is an attempt to put a number on unquantifiables, and it's nowhere near perfect. But by trying to translate the two bonuses each trait gives, its not unreasonable.


vmxa1 - You're right, play style counts big time. However I made an assumption that most people would put all 3 science buildings in all their GOOD cities (not the hopeless outliers). At least I would. If you don't, then you are right, Scientific's gold/shield value would be lower. Happily, that particular adjustment would be quite easy to make. (On a sidenote, give Universities a shot - 4 culture!)


Arrian - You're right, playing to the trait will increase it's value. More fighting increases Militaristic, more Temples Religious, and so on. I was trying to find which was worth more if all else was equal, or ceteris parabis as they say. This would help give an idea which trait gave you more return in a generic situation. If it gives more in that case, then when you play to the trait, it will give even more.

However some traits may well give more return if you play to them than others. What I mean is, if you go all out war, Militaristic may increase in value more than Commercial would if you went all out gold. Hard to say. I'll give this some thought.


mharmless - As Sevorak first noted, I do have to take lower unit upkeep costs into account.

However, I am not so sure I follow the point "I don't lose that unit when it takes four hits." I think having 20% extra units would simulate this at least reasonably well.

As to the wonders, very good point, and good math! However I think adding in the value of the wonder would be what they call double counting, which is an easy trap. The price of the wonder is the shields it takes to make it, not it's effects. For instance, if I counted the value of the wonder for Militaristic, then I would have to count the value of whatever the extra shields were used to build for Industrious, the extra gold was used to buy for Commercial, and so on. I think that is indeed double counting and not right to do.

Good points though, I'll have to think about it, but I do believe the wonder thing is double counting.


rah - Good points ... I think I have some answers.

For the workers, the double speed has a value ... I tried to calculate it and credit Industrious for it.

For the Religious anarchy, that is correct, if you play to the trait you will get more out of it. Again I was trying to come up with a generic situation. My one early change + one late change setup was to simulate going to either Monarchy or Republic early, and then Democracy (or less likely Communism) later.

I did say "More or less revolutions will change this value, as will size of your empire." ... I think two is reasonable for a normal game ... maybe it should be three though.

Leaders are valuable ... they are worth one wonder apiece. I credited them for the cost of the wonder. I think the 20% extra units measures the value of extra promotions, albeit very roughly.

I think my assumptions on the game, 25 good cities basically, are reasonable. I am definitely open to suggestions if ppl think that is not a good average. It is supposed to be a normal game that gives all the traits a level field to show what value they give under normal circumstances.

"it is most important that you take max advantage for any characteristics that you choose."

You are dead on here - that is a great conclusion to point out. Doing that basically increases the gold/shield value your trait is giving you.


Thanks for reading and taking the time to make comments everyone! I do appreciate it.

[rah]

Please don't kill yourself making it perfect, since I doubt it ever could be. But tweaking a little provides some great food for thought.

It has already served the purpose by making us think on how to quantify it. By attempting to shoot holes in what you've done, it's made us discuss exactly some of the issues and makes us question if we're choosing the right civ, or taking proper advantage of what we choose.

For the record, if you're religious I'd expect at least 6-10 government changes if you want to maximize your advantage.
But the value can't be compared to non-religious since you wouldn't think of doing it that often. Oh... it's so hard to quantify.



RAH

[mharmless]

I counted the full wonder because in the early game (for me at least) the only way I get wonders is if my leader allowed me to get it immediately. Even if it could be had via building, the time it takes to make it should be factored in, and then the time value of that wonder can be added to the pile too. If colosus takes your archtypical city 100 turns to build, than the leader has caused 100 extra turns of effect on that wonder. Same for all wonders.

With the forbiden palace, I didn't caculate it's value for the entire game, but just for the 300 turns it would take a backwater berg to construct it. With the leader, it is done 300 turns earlier and turns that entire area into a powerhouse immediately. Those 300 turns are then all productive in all those cities, hence the insanely large sheild value.

Edit: About the not loseing the unit when it takes four hits. Four elite spearmen are superior to five veteran spearmen. If attacked by five veteran horsemen, ignoreing defensive mods, then the five veteran spearman each have about a 50% chance of dying. The Elites though, they will only lose about 20% of the time. The four elites will hold the ground better than five vets (ignoreing in battle promotions to elite from vet). Particularly against horseman, as if we trade hits equaly once it loses 3 it will be at one hitpoint and retreat, wheras my spearman is at 2 hits still. If it was a vet, then they would both be at 1 hp and the horse would attack untill death (50% chance I lose the vet at this point, 0% I lose the elite). Less units lost, less resources expended. In this case, I save 20-40 sheilds that would have been spent replaceing the extra deaths among the vets.