How does the Power graph work, or intimidating Power?

What exactly goes into the AI's estimate of how militarily powerful my civ is? Is it based on the power graph?

Either way, what exactly contributes? I assume with confidence that Unit Strength of all military units contributes to this, this thing we might call the Power Estimate or Apparent Military Power, but is it linear?

Does one Longbowman of Strength count for as much as two Archers each of Strength 3? Or is there some exponential formula used so that one S6 Longbowman counts the same as 2.5, 3 or even 4 S3 Archers? (Strength squared would make 4 Archers give the same Power as 1 Longbowman.)

Do unit Promotions factor in? Just the number of Promotions, or the speciic ones chosen?

What about city buildings, especially Barracks, Walls, Castles and Stables?


I'm asking because part of my strategy is to (try to) build a military force sufficient to deter the AI from attacking me, and I want to optimize this approach, so that I use my ressources in the best possible way.

Right now I just put one Archer in each city as it is founded (or before its founding, or at the least a few turns after it is founded, especially if I have The Great Wall) built from my capital city (which has a Barracks and may also have a Military Instuctor), and then later on I build a bunch of Longbowmen stationed in my capital, and may also build one in each of the other cities (after they've got Barracks). I also aim for upgrading the Archers to Longbowmen, although that costs a lot of gold, and I also try to have one Crossbowman (S6 becomes S9 vs melee units, better than S6 Longbowman +25% city defence giving 7.5, and also they upgrade nicely to Machine Gunners) per city eventually, and maybe one Spearman/Pikeman with Combat I and Medic I (two Crossbowmen and two Pikemen in my capital, and 3-4 Longbowmen plus a couple of old Archers).

Also I build a small offensive force of Swordsmen, later adding some Macement to it, and if I can afford it I eventually upgrade the Swordsmen to Macemen. (These - of course - get City Raider Promotions, whereas the Archers, Crossbowmen and Longbowmen get City Garrison.)

This seems to work, keeping my Power graph fairly high (usually not the highest, but enough to deter the AI), but it does cost a lot of hammers.

Can I utilize my hammers better? Should I prioritize Walls more? Does a Barrack in every city help? Should I build more Longbowmen and fewer Swordsmen, or fewer Longbowmen and more Swordsmen?

Would half a dozen Knights help a lot, out of proportion to their hammer cost? I tend to skip Catapults and wait until I can make Trebuchets, then build 3-4 of those to add to my offensive stack, but would building a couple Catapults ASAP boost my Power graph out of proportion to their hammer cost?

The key here is that I'm not a warmonger, I don't want to fight.

What I want is to discourage the AI civs from declaring war on me, so that I can focus on inventing new techs and upgrading my cities, then maybe once I get a really serious science advantage, I might decide to hit the AI civs with the big out-of-context stick.

You can always figure out how much power someone has, so long as you are in contact with them. The power graph shows the amount of soldiers in your army.

fdgsgds said:

You can always figure out how much power someone has, so long as you are in contact with them. The power graph shows the amount of soldiers in your army.

Click to expand...

The Power graph is based on the number of soldiers I have?

Peter Knutsen said:

The Power graph is based on the number of soldiers I have?

Click to expand...

Every military unit as well as certain buildings (Walls, barracks etc) some Wonders (i.e. Great wall), your population and some techs have a certain number of soldiers attached to them, the value for everything is given in this article. Your power is the total of your soldiers.

I would advise against trying to use power to prevent wars however as its really not very effective. Each AI has a certain power ratio above which it will not attack, unfortunately this value is often high (To stop Monty you need double his power) and if you don't reach it then your power means absolutely nothing to them.
Diplomacy is the way to go to prevent unwanted wars!

I have not heard of any fix for it, so I assume that every single military unit is counted, including all of those stupid cheap ships and Zeppelins that the AI always builds.

I have seen power graphs that were 3x mine, yet with modern units I easily smashed them into oblivion. The AI seems very poor about upgrading units - it just keep adding more.

Monty is the major pain in the neck, but not the only one. Those civs that tend to be early attackers are much more dangerous, as they often seem to do nothing but improve the minimal amount of squares and then go full bore into building tons of units. I have seen stacks of 20+ horse archers and the like pre-1AD.

Ghpstage said:

I would advise against trying to use power to prevent wars however as its really not very effective. Each AI has a certain power ratio above which it will not attack, unfortunately this value is often high (To stop Monty you need double his power) and if you don't reach it then your power means absolutely nothing to them.
Diplomacy is the way to go to prevent unwanted wars!

Click to expand...

Nevertheless it actually works for me, at least on Warlord difficulty (3rd lowest) level. I'm not advanced enough to play higher.

If I build a fair defensive force, and one offensive stack, then I can keep trading with everyone, and just reject consistently when one of the AIs demand I stop trading with their worst enemies, and only rarely gift them techs when they ask nicely (never when they demand or threaten), without being war decced.

I'm quite certain this is because the AI perceives me as being militarily powerful, but I'd like to know what the best approach is for being perceived this way, if I'm spending my hammers the right way, since if I can do it more efficiently, that means more hammers for non-military projects, and also a bit more commerce going to science because I have fewer military units to maintain.

Peter Knutsen said:

The Power graph is based on the number of soldiers I have?

Click to expand...

Not necessarily. Each unit and building in the game has a power rating. You can find the value of it in the respective XML files. These ratings are added up and contribute towards your overall power rating. You need to take that rating with a grain of salt though, as it doesn't always give you the big picture. An enemy could have alot of really crappy units and the graph would indicate that it is quite powerful, even though your Riflemen etc. could cut through his forces like butter. Or it could be just the opposite, with your Macemen getting slaughtered by his Infantry.

Willem said:

Not necessarily. Each unit and building in the game has a power rating. You can find the value of it in the respective XML files. These ratings are added up and contribute towards your overall power rating. You need to take that rating with a grain of salt though, as it doesn't always give you the big picture. An enemy could have alot of really crappy units and the graph would indicate that it is quite powerful, even though your Riflemen etc. could cut through his forces like butter. Or it could be just the opposite, with your Macemen getting slaughtered by his Infantry.

Click to expand...

Of course; I was mocking the other guy. I never believed the formula didn't take unit Strength into account.

Where can I find these XML files?

Search for an old thread called "The Inner Workings of the Demog Screen Explained" by Robi D, which does exactly what the title says - including how power rating is worked out. There may have been some changes since it was posted.
Where are the XML files ? Try looking among the Civ4 or Civ4 BtS Assets folders, wherein you'll find sub-folders and eventually files clearly labelled XML.

Bushface said:

Search for an old thread called "The Inner Workings of the Demog Screen Explained" by Robi D, which does exactly what the title says - including how power rating is worked out. There may have been some changes since it was posted.
Where are the XML files ? Try looking among the Civ4 or Civ4 BtS Assets folders, wherein you'll find sub-folders and eventually files clearly labelled XML.

Click to expand...

Thanks, that's very useful!

I was aware of the existence of that article, but I thought it only dealt with the actual Demographics screen, like the stuff back from Civ 1, Literacy percentage, life expectancy, length of militayr service and so forth.

I've compiled this quick reference list of fairly early units and buildings, and their Power graph contribution, along with what the units can upgrade to (although I don't remember all upgrade paths), but I've divided all values by 1000:


2 Walls
2 Stable
2 Forge
3 Barracks

2 Warrior
3 Archer (Lbow 6, Xbow 7)
4 Spear (Pike 6)
4 Chariot (Kni. 10?)
5 Catap. (Can. 12)
6 Sword (Mace 9)
6 Axe (Mace 9?)
6 Pike
6 Lbow
6 H. Archer (Kni. 10?)
7 Xbow (M. Gun 14)
8 Trebu. (Can. 12)
9 Mace
9 Musket
10 Knight
12 Cannon
14 Rifle
14 Machine Gun

Later on I'll probably do some kind of analysis of hammer cost vs Power contribution, and maybe also upgrade gold cost vs Power contribution difference, e.g. exactly how much oomph one gets for paying the gold cost to upgrade one Archer to a Longbowman, or one Swordsman to a Maceman.

(In general I find upgrade costs very high, though, and the AI probably does too, since it very rarely seems to upgrade its units, at least on the low difficulty levels that I can play at.)

Here's the hammers-to-Power ratio of those units I mentioned before, plus a few more units that are also fairly early (although I don't want to bother including naval units), and also City improvements.


I've divided the units into categories, based on how I use them, and then sorted for most favourable ratio (most bang for your hammers), and included Walls twice, once with the discount for Stone (yay!) and once without.

Clearly, if you want to look intimidating to the AI, just build lots of Axeman. For a more useful offensive force that is still somewhat intimidating, build Swordsmen instead, but build relatively few siege units. Both Catapults and Trebuchets contribute little to Power rating, so only build as many as each offensive stack needs to be able to shoot down city walls.

If you intend not to attack, you might want to not build any siege units at all; if you realize you do need to attack, you can build a few siege units quickly, if you already have the "meat" of the offensive force, in the form of Swordsmen or Macemen. Macement themselves are also somewhat worth building, although less attractive than Swordsmen.

Defensive units, Archers and so forst, don't help much, so just build one per city, and then only build more if you're using Heriditary rule, although even if you are, it may be better to build Swordsmen and station them in your cities, although personally I prefer to have my offensive units in one big stack at some tile outside my cities.

Mounted units don't help much towards Power. Chariots are fairly cheap (7.5 hammers per Power, i.e. per 1000 "soldiers"), but I don't recall them as being particularly strong. Eveyrthing else is less favourable, although Cavalry, a late unit, comes close and is probably well worth building.

As for buildings, you want a Barracks in every city, clearly (besides the Power effect, it's a permanent structure barring spy sabotage, unlike units which die, and having lots of Barracks means you can switch to unit production in most cities in emergencies), and if you have Stone you definitely want Walls. Without Stone you might only want Walls in your capital, depending on how much you expect to be attacked. (IIRC Walls are not a prereq for The Great Wall.)

A Forge helps very little with regards to Power, but its other effects makes it very attractive anyway. Stables do not appear to be worth building, unless you want to field a mounted force, in which case you should still only build Stables in a few cities.


Note that I have not yet looked into unit upgrade gold costs. I tend to not have much gold in most games, but I'll probably do some kind of analysis of that today, anyway, then post my findings.

First the units sorted into various categories (as I use them) and then within each category sorted according to most favourable ratio between gold upgrade cost and the difference between the old and new Power, Power being the number of soldier-equivalents divided by 1000.


DEFENSIVE
Archer to Longbowman isn't very favourable, but it's the best you're gonna get early in the game, so if you have a bunch of gold, upgrade your Archers to Longbowmen to further intimidate the AI civs against declaring war on you. Later on, Longbowman to Rifleman is cheap, and Rifleman to Infantry is even cheaper, although of course later in the game, each point of Power counts for less.

Upgrading Crossbowmen to Machine Gun doesn't give much intimdiation per gold coin, but it does give you Machine Gunners which have a high defence-only Strength, and also your Crossbowmen can have the City Garrison promotion, unlike Machine Guns you build directly, so building one Crossbowman per city is, to my mind, attractive, especially if it can be built with 2 promotions (Barracks plus Vassalage).

OFFENSIVE
Warrior to Axeman is fairly cheap, so that can probably give you some good early-game intimdiation agains the AI, for not a whole lot of gold (60 gold each, and early game you can perhaps upgrade them one by one since you're likely to have few cities). Swordsman to Maceman is really expensive for what it gives in intimidation, even though Macemen are better city attackers (S8, sometimes S12, vs the S6.6 of a Swordsman). The question is, is it worth upgrading Swordsmen to Macemen? AI city defence is often, like mine, based on Archers or Longbowmen or Crossbowmen, and the Macement don't get any bonus against those, although the AI might use Pikemen for city defence too, where the Macemen come in handy (I use Pikemen sometimes). After some thought, I'll change my playstyle and from now on not upgrade from Swordsmen to Macement, instead keeping them around until I can make them Riflemen.

Note that Musketmen are somewhat worth building, because as Gunpowder units they get to ignore wall and castle defensive bonuses. That's somewhat attractive. Sadly, nothing can be upgraded to Musketmen, although as with defensive units, Swordsmen/Macement upgraded to Riflemen are preferable for the offensive stack(s) because unlike directly built Riflemen, the upgraded ones can have City Raider promotions.

SIEGE
Not much to say here. Trebuchets bombard better (16% per turn per unit off city walls, vs 8% per turn per unit for Catapults) but can only be built later in the game, and cost twice as many hammers.

MOUNTED/FAST
These are really expensive to upgrade. Having a few around may serve other purposes, but building and upgrading these for the sake of intimidation is probably a big mistake.



I've gone by the v3.13 Reference Chart, which IIRC I downloaded from this site (Civ Fanatics) a month or two ago. The one exception is Crossbowman to Machine Gun which isn't in the Reference; I added that one myself based on personal knowledge (although I also loaded a saved game to check (and yes indeed it can be done) and of course to find the gold cost for doing so).

I'm wondering if anything can be upgraded to Grenadier, for instance, so if any upgrade paths are missing, please let me know...



Finally, all the figures, probably including a few that I have left out of the above categories due to not being sure where to place them.

Peter Knutsen said:

I never believed the formula didn't take unit Strength into account.

Click to expand...

It doesn't though. A unit's power rating is a different value from it's strength, and they aren't directly connected.

Where can I find these XML files?

Click to expand...

Game Folder/BtS/Assets/XML. If you're planning on altering anything, make sure you copy the file over to your Custom Assets folder first.

Another quite related question:

It seems that for a period of time after I can build Macemen, I can still build Swordsmen, but then later I become unable to build Swordsmen. Why is that? Clearly the tech that allows me to build Macemen doesn't prevent me from building Swordsmen.

Likewise, once I get Feudalism, there is a period of time in which I can still build Archers alongside Longbowmen, but then after a while (after some more techs) I can only build Longbowmen.

What is this mechanism?

In some cases, I might prefer to continue building Swordsmen, so it bothers me that the game prevents me from making that choice.

Peter Knutsen said:

It seems that for a period of time after I can build Macemen, I can still build Swordsmen, but then later I become unable to build Swordsmen. Why is that? Clearly the tech that allows me to build Macemen doesn't prevent me from building Swordsmen.

Likewise, once I get Feudalism, there is a period of time in which I can still build Archers alongside Longbowmen, but then after a while (after some more techs) I can only build Longbowmen.

Click to expand...

I think you're mistaken about the Swordsman, they shouldn't be appearing after Macemen show up. It has to do with the number of different units that something can upgrade to. In the case of the Archer, it can upgrade to either a Longbowman or a Crossbowman. So as long as either one of those can't be built yet, you'll still be able to build the Archer.

Power means so little in the vast majority of cases. What matters is if your units can let your survive when diplo fails.

TheMeInTeam said:

Power means so little in the vast majority of cases. What matters is if your units can let your survive when diplo fails.

Click to expand...

That is so not true! I can't believe you'd even make a statement like that. I've had plenty of games where I didn't care about diplo at all, if a civ didn't like me too bad. But having a strong power rating made all the difference as to whether I was declared on or not. Diplo is what doesn't matter in this game. Many Civs will declare on you even at Pleased, but they'll think twice about if you have a military that can kick their butt. I've even had plenty of civs that were Furious with me, but they never attacked because my power rating was much higher than theirs. Power is everything in this game. If you don't have it, expect to be declared on frequently.

Willem said:

That is so not true! I can't believe you'd even make a statement like that. I've had plenty of games where I didn't care about diplo at all, if a civ didn't like me too bad. But having a strong power rating made all the difference as to whether I was declared on or not. Diplo is what doesn't matter in this game. Many Civs will declare on you even at Pleased, but they'll think twice about if you have a military that can kick their butt. I've even had plenty of civs that were Furious with me, but they never attacked because my power rating was much higher than theirs. Power is everything in this game. If you don't have it, expect to be declared on frequently.

Click to expand...

At least on higher levels it's virtually impossible to maintain a large enough army to deter DoWs by power ratio. A "pretty strong army" has no effect on war decisions - you have to pass a certain (high) threshold for them to opt peace because of your power. Usually it's virtually impossible to even have comparable power to the AIs for the most part of the game, never mind pass the threshold.

Anyway, in most cases it takes about four-five times as much army to deter a DoW by power than deter an invasion. Diplomacy is (relatively) free, units cost hammers and gold.

Willem said:

That is so not true! I can't believe you'd even make a statement like that. I've had plenty of games where I didn't care about diplo at all, if a civ didn't like me too bad. But having a strong power rating made all the difference as to whether I was declared on or not. Diplo is what doesn't matter in this game. Many Civs will declare on you even at Pleased, but they'll think twice about if you have a military that can kick their butt. I've even had plenty of civs that were Furious with me, but they never attacked because my power rating was much higher than theirs. Power is everything in this game. If you don't have it, expect to be declared on frequently.

Click to expand...

Power is a tiny bit relevant below monarch. Monarch+ blathering about power is mostly nonsense.

The worst part, however, is that what you're posting is flat-out wrong. Maybe a tiny look at the code and you'd see that...both the AI targeting code and the power threshold required to stop an AI from declaring.

Some AIs take anywhere from 150% to 195% of their power before they won't declare on you (fun that these tend to be the same guys who plot @ pleased). If you have double your opponent power, yes you will win. And that sounds great...just have double their power! But that's a ridiculous statement in context. Nobody playing at their real level is going to be able to get 2x the AI power when a DoW still matters. To afford that is literally impossible on emperor+. You can't do it, and neither can anybody else (excepting later-game situations where they've essentially won already).

Let's pick this apart a little more:

Civs will declare on you even at Pleased, but they'll think twice about if you have a military that can kick their butt.

Click to expand...

The AI doesn't "think" at all. If you are above a cutoff threshold, it won't declare...too bad this threshold typically requires you to be markedly stronger. If you are below that threshold, power is absolutely 100% worthless for averting a DoW from that AI. Adding to the fun is that if you are at war, anybody else considering a DoW gets to add your target power to its own, further increasing the ridiculous #'s you need.

On the flip side is this trash statement:

Diplo is what doesn't matter in this game.

Click to expand...

And this one:

Power is everything in this game. If you don't have it, expect to be declared on frequently.

Click to expand...

To the first, I'll just point that as long as your mindset remains this way, you will not consistently beat high levels. This mindset does matter in multiplayer, however. But for SP, it's ignorance that will block your improvement and anyone else's who believes it.

As for the 2nd one:

http://www.youtube.com/watch?v=sBphpz3AgVw&feature=PlayList&p=414EEAB9F8CDB010&index=15

Whoops. Guess diplo isn't meaningless after all.

Willem said:

That is so not true! I can't believe you'd even make a statement like that. I've had plenty of games where I didn't care about diplo at all, if a civ didn't like me too bad. But having a strong power rating made all the difference as to whether I was declared on or not. Diplo is what doesn't matter in this game. Many Civs will declare on you even at Pleased, but they'll think twice about if you have a military that can kick their butt. I've even had plenty of civs that were Furious with me, but they never attacked because my power rating was much higher than theirs. Power is everything in this game. If you don't have it, expect to be declared on frequently.

Click to expand...

As the others have said it's mostly about difficulty level. At Monarch and below it's easy to be so strong that the only point of the diplo button is to see what you can extract in tribute every 10 turns and to trade resources.

EDIT: OOPS brought this thread from the dead my bad.