Soldiers rating in demographics

[Nefliqus]

How units powers is transformed into Soldiers rating in demographics. Dose this rating include only units power or some technology and building also? Is it Combat strength or Basic Units strength? Does it include city power also?


NOVAL invert mirror map script

 

[southpw11]

If I understand your question, you're asking what goes into the soldiers demographic rating. I think this question was asked once long ago, and I believe it includes combat units only, and takes into account what type of unit (ie..warrior is worth a lot less than infantry).

 

[Tabarnak]

With hoplites and companion cavs, Greece seems to be on top soldiers pretty often in the beginning. Both UUs have higher strengh than regular units of their type.

 

[Nefliqus]

If I understand your question, you're asking what goes into the soldiers demographic rating. I think this question was asked once long ago, and I believe it includes combat units only, and takes into account what type of unit (ie..warrior is worth a lot less than infantry).

It is not so easy and clear, for example
one worrior with combat str 6 results with 8246 but
2 worriors with combat str 6*2 give 11662 result


NOVAL invert mirror map script

 

[Tabarnak]

I'm not sure, but i think workers and settlers have some value.

 

[CrazyHorseSouix]

this is something I've always been curious about. Not just soldier count, but how are all of the demographics figured?

It has also always bugged me that pop. counts always seem low. If it is 1950, and my empire owns most of the world with most of the cities thriving, shouldn't I have more than 30 million people?

 

[vexing]

it's units alone,
the values increase over time,
level is included in unit rankings,
the formula is not very clear. for example, having two warriors might only be a 50% value increase over having one. i could probably figure out the exact formula but the demographics screen is only updated between turns which is annoying for testing purposes.

 

[Lexicus]

Yes, the population demographic is stupid. I've played games where I own half of a Huge map, with many 30+ cities particularly in my core, and I don't think I've ever seen population exceed 50-60 million.

Soldiers takes into account number of units and tech. I'm pretty sure workers and settles aren't counted but I could be wrong.

Demographics are largely self-explanatory...Approval Rating is the only one that doesn't make much sense to me.

 

[Nefliqus]

Population formula is very well known:
DEMOPOP = 1000 * City1POP ^ 2,8 + 1000 * City2POP ^ 2,8 ... + 1000 * CitynPOP ^ 2,8
rounded down to nearest 1000

Here you have table with list of demo pop and corresponding number of total citizens and cities

demoCIV5.pdf:


NOVAL invert mirror map script

 

[Nefliqus]

Spoiler :

http://forum.civilization.org.pl/img/Nefliqus_phpVsvJO5_civ5soldierscore.jpg

It is simple method how to calculate score of units group.
I have tried to make mathematics formula for this but it was to hard for me. Maybe some of you can use this score step list.

Spoiler :

Military score steps
4000
4472
4899
5292
5657
6000
6325
6633
6928
7211
7483
7746
8000
8246
8485
8718
8944
9165
9381
9592
9798
10000
10198
10392
10583
10770
10954
11136
11314
11489
11662
11832
12000
12166
12329
12490
12649
12806
12961
13115
13266
13416
13565
13711
13856
14000
14142
14283
14422
14560
14697
14832
14967
15100
15232
15362
15492
15620
15748
15875
16000
16125
16248
16371
16492
16613
16733
16852
16971
17088
17205
17321
17436
17550
17664
17776
17889
18000
18111
18221
18330
18439
18547
18655
18762
18868
18974
19079
19183
19287
19391
19494
19596
19698
19799
19900
20000
20100
20199
20298
20396
20494
20591
20688
20785
20881
20976
21071
21166
21260
21354
21448
21541
21633
21726
21817
21909
22000
22091
22181
22271
22361
22450
22539
22627
22716
22804
22891
22978
23065
23152
23238
23324
23409
23495
23580
23664
23749
23833
23917
24000
24083
24166
24249
24331
24413
24495
24576
24658
24739
24819
24900
24980
25060
25140
25219
25298
25377
25456

it is log function f(x) = loga(x − p) + q

Military score is depended on:
-unit type (unit base power determine first unit score which is positive modified by unit promotions and current civ gold)
-unit promotions
-units number
-civilization gold amount

e.g gold steps for warrior and immortal:

Spoiler :

warrior; first score 8246 (without promotions)
0
35
139
312
554
866

immortal; first score 11489 (without promotions)
0
37
83
147
230
331
450
587

If civ rich gold amount step unit score move up by 1 level in score list, 2 gold steps = go up by 2 levels.
Gold steps are connected only with first unit score, not with unit power e.g gold steps for warrior with one promotion:

warrior; first score 8944 (one promotion);
0
26
100
226
400

and jaguar with base power = warrior
first score 9165 (without promotions)
0
23
91
205
363

Military score is NOT depended on:
-cities number
-city power
-unit combat power
-technology score
-turn number

If someone can use this data for preparing mathematics formulas pls send me info how u did it.
I need 2 formulas: military score steps and unit gold steps.

NOVAL invert mirror map script

 

[vexing]

how did you get these numbers? in f(x) = loga(x − p) + q, what are the x, p and q?

 

[Nefliqus]

how did you get these numbers? in f(x) = loga(x − p) + q, what are the x, p and q?

it is log function formula moved by vector [p,q]
but
it is not important because you can use this formula (2000 * (3 + n)^0,5) to calculate n position from score list, n starts with 0

NOVAL invert mirror map script

 

[thectexperience]

Another problem with the demographics that I've found is that they will offer clash with the updates that appear between turns: "most advanced Civ"; "most powerful Civ" etc...
I pay close attention to the techs and I've seen fairly dramatic differences between number of techs on the demographics screen vs the one-off info update
I find the advisor screen to be similarly inaccurate. I recently attempted a sword rush on the tech-crazed Sejong. My military advisor told me he just saw the only unit the Koreans had, but when I arrived a few turns later they had at least one pike for every one of my swords, plus catapults and some auxiliary archers etc
The bottom line is that the lack of info sucks, but you best be aware of it

 

[craig123]

it is log function formula moved by vector [p,q]
but
it is not important because you can use this formula (2000 * (3 + n)^0,5) to calculate n position from score list, n starts with 0

I see you've also posted this on CivFanatics.

A minor point but n starts with 1.

I've been thinking about this some more. Are you sure that 4000 is definitely the lowest possible rating. The reason I ask is that I wonder if the actual formula might be 2000 * n^0.5, and therefore the sequence would start: 2000, 2828, 3464, 4000, 4472, ...

Perhaps the starting warrior takes the score up to 4000. If he is disbanded does it fall to 2000?

I have no evidence for this, but I just can't work out why they would use n + 3 in the formula rather than just n.

 

[vexing]

the formula is just 2000*n^.5
that gets rid of the funky -8 group subtraction that nefliqus was doing. multiple units just get their n values added together. also there is a lua command "getmilitarymight()" which returns the n value. i've been trying to figure out how the n value is gotten, here are some n values if someone else wants to give it a go:

american minuteman 78
aa gun 432
at gun 277
arabian camel 69
archer 16
aztec jaguar 21
barb archer 19
barb spear 27
barb sword 44
barb warrior 17
battleship 115
cannon 120
caravel 13
carrier 199
catapult 39
cavalry 173
chariot archer 24
chokunu 39
crossbow 61
destroyer 103
egyptian warchariot 25
english longbow 61
galley 2

 

[chazzycat]

I wonder if ships having such low comparative values has anything to do with the AI not building enough of them...

 

[craig123]

the formula is just 2000*n^.5

Wow, my hunch was right. Thanks vexing - how did you find that out btw?

 

[vexing]

two ways: first just looking at the values and the grouping math, it's a lot clearer to not have the magic 3 thrown in there. without it, groups can just be added together.

second, i found the in game lua command getMilitaryMight. it returns a value that can be plugged directly in the formula 2000*n^.5 to get the demographics value. i listed the results of that command for a subset of units.

 

[vexing]

the formula for n seems to be (1.15 * combat strength) ^ 1.5
this isn't exact, there are maybe two levels of truncation, but it fits the basic curve.

there are additional percent modifiers for some subset of innate promotions, which i haven't figured out exactly. for example the jaguar and mohawk warrior both start at 25% higher than their counterparts, an anti tank gun starts at 25% above it's base 222 (277), while an anti aircraft gun starts at 95% above (432).

additionally the % damage a unit has suffered is cut directly from its strength (ie an aa gun at 1 health is either 43 or 44)

 

[Tabarnak]

the jaguar and mohawk warrior both start at 25% higher than their counterparts

That explains why in mp game Iroquois and Aztecs are always on top in demographics even if you really try to build a bigger army. Especially Aztecs in the first 20 turns.

Thanks vexing great info.