Combat casualties

[Creepy Old Man]

Has anyone figured out yet how many troops can be expected to die from each attack? In particular, I'm baffled why I seem to always take much higher casualties than my opponent.

For example, in my most recent battle, the two sides were very evenly matched. There were many rounds in which our riflemen were attacking each other. Army strengths were staying around 5750-vs-5550. When my heroic riflemen attack their heroic riflemen, 13 die. (12 later in the fight, when numbers are lower). When their heroic riflemen attack my heroic riflemen, 21 die.

After watching many fights, it's clear that casualty rates are not random. They're always substantially higher for me. Anyone know why?

Incidently, it's not just riflemen, nor does it have to do with the fact that rifles are stronger defending than attacking. In that same fight, the first turn they killed all 53 of my horsemen. Later on, when I attacked their horsemen, (both cases the horsemen were heroic), I killed a whopping 16.

I'm hoping to figure out how to improve my battle efficiency.

 

[Glinda]

Heroic stance always takes massive losses. As far as I can figure out, the losses are dependent both on total strength, and the strength of the stacks involved.

If your opposition is defending, then his DEFENSE numbers are what count, even when he is attacking your guys. So his riflemen are attacking with their defense number against your attack number. Modified by stance (heroic), and total forces.

So when your rifles attack his rifles they are attacking with a lower number (attack) against a higher number (defense). When his rifles attack your rifles, ditto. He attacks with the defense number and you defend with the attack number, since you are the attacker.

No, this doesn't make intuitive sense.

The best way to improve your battle efficiency is to take your guys OFF heroic when they are about to be attacked.

 

[Creepy Old Man]

In that case:

Why do horsemen (attacking at strength 1) inflict the same casualties as knights (attacking at strength 4)?

Why do I lose an entire stack of 53 horsemen, but the defender loses 16 out of 100?

You suggest not using heroic stance, but as I mentioned, this was a very close fight, so it was necessary. Furthermore, from what I understand, there is no way to predict "when they are about to be attacked".

 

[Glinda]

I don't know entirely how it works, but stacks do matter, so his 100 do more damage to your 53 than your ?? do to his 16. Weather was maybe a factor? Was it snowing?

 

[Pinstar]

From What I've seen it does this:

1. Add up total strength on each side and choose a number. The side that owns the chosen number gets to attack this round.
(IE a 200 power army vs a 100 power army will, on average, be twice as likely to get the attack in any given round)
2. Once the attacking side is chosen, choose a random stack in the attacking army's stack.
3. Once a stack is chosen, choose a random enemy stack of the same type (melee vs melee).
3b. If the opponent has none of that type, choose a random opponent stack type.
4. Inflict damage on opponent stack. Damage inflicted is based soley on the stance of the defending unit. The unit stack used to attack does not increase or decrease the damage at all. (a stack of 1 knight and a stack of 100 knights will both inflict the same damage).
NB: Changing the defender's stance during the round does NOT affect its casualties.
5. When round ends, push the attacker's team up one notch on the victory scale.

Repeat


Heroic is only good if you have inferior numbers and are trying to get the power scores up above 50% in your favor. In that case you have a chance to actually beat a stronger opponnent if the RNG is on your side. Fortified is best if you have a very small number of a type of unit and don't want your opponent to get the free bonus in case they wipe yours out. If you've got only 5 horses split among 4 mobile slots, fortify them.

Ideally you want your units on 'normal' while still maintaining a >50% attack chance at the start of the battle.

 

[nemhauser]

I've had the following battle today:

Aztecs (force of 10.000) were attacking Americans (force of 500) while me and my buddies from Brittish Commonwealth were helping Americans defend (force of 15.000)

This was a double screen battlefield. Americans had most of the slots taken by their units with stacks of 1-10. I had a stack of 1100 riflemen, other friend had a couple stacks of 200-400 mounted and ranged.

Opponent was mostly in heroic so my stack of riflemen was as well (without it we would not have the upper hand).

Now, the weird thing I've noticed was that almost ALL of their attacks were against my riflemen (killing 19 units per attack).

If this is random then I must be the unluckiest guy in the game. (out of 24-26 stacks my stack was targeted 6-7 times in a row... probability of that is really low)

So, I'm thinking that it might be random picking but with a weight based on the power of the stacks.


I also noted that the enemy heroic riflemen were having casualties of 22 which was a bit higher than mine (but the riflemen are better in defence so that was logical to me)

 

[Creepy Old Man]

I've seen similar results. I want to log a couple more battles, and then I'll do some rigourous statistics.

For now, I'm leaning towards the following theory:

[pre]
Logarithm( <Quantity of stack> )
<Probability of stack being attacked> = ------------------------------------------
Sum over all stacks( Logarithm( <Quantity of each stack> ) )
[/pre]

 

[deviant devin]

You suggest not using heroic stance, but as I mentioned, this was a very close fight, so it was necessary. Furthermore, from what I understand, there is no way to predict "when they are about to be attacked".

if it is a very close fight, then you will be alternating attacks at each turn. so move your guys to 'fortify' while they are attacking because you know their attack is up next. then while they are attacking, go from 'fortify' to 'heroic' and you slaughter them if they are still on heroic.

the gunmen have high defense and low attack so that is why they are more effective for the side that is defending. if you are getting atttacked, use phalanx for example, and if you are attacking, use legion.

 

[Garrcia]

I don't know entirely how it works, but stacks do matter, so his 100 do more damage to your 53 than your ?? do to his 16. Weather was maybe a factor? Was it snowing?

I think the shooting stack only matters for selecting what kind of stack is shot (same vs same if possible).

For the shot stack the considerations are its stance and that overkill (i.e. damage>stack total) does not carry over to other stacks - hence the value of single unit stacks.

 

[JaeRoler]

@Creepy Can you tell us who was the attacker in the aforementioned battle? Also, telling us whether Leonardo's Workshop or Himeji Castle was involved would help.

From what I have noticed in my battles, my side managed to chain a series of 4 attacks on the opponent with our attack power remaining the same. The opponent is attacking whilst I was defending. No troops was in heroic. Weather remained the same for the 4 rounds.

4 Knights (Attk 4) died.
2 Cruisers (Attk 8) died.
2 Artillery (Attk 8) died.
5 Riflemen (Attk 3) died.

You will notice that these numbers corresponds to a damage of 15~16.

My theory is that :

1) RNG will choose which side to take the damage.
2) Based on the opposing power fielded, the battle system will calculate the damage dealt (how they do this, I have no idea yet).
3) RNG will choose a stack to take casualties.
4) Casualties suffered = Damage dealt divided by power of unit. (take the corresponding value when you are attacking or defending). This is capped by the max number of units in the stack.
5) Modifiers to the Casualties are then applied depending on the stance that you took.
6) All Casualties suffered is then rounded up so that minimum Casualty is always 1.

Based on this theory, I would say that Creepy is on the attacker's side. Therefore, with near-equivalent power,

21 Riflemen on Heroic = 7 Casualties before stance modifiers are applied.
12/13 Rifleman on Heoric = 4+ Casualties before stance modifiers are applied.

7 Attacking Rifleman = 21 damage suffered.
4 Defending Rifleman = 20 damage suffered.

If we assume that 21 damage is the damage suffered by both sides, then

21 damage = 4.2 Rifleman casualties, modifed by Heroic (x3) equates 12.6 Rifleman casualties, and then rounded up to 13 dead Rifleman actual.

I could be wrong in my theory, but it seems to fulfil most of the battle parameters that I have seen.

Edit :

It occurred to me that it would do wonders to our understanding of the system if someone could do actual battle logs (round by round details), so that we have a database to study from. Unfortunately, I understand that there is no actual battle logging system in CivWorld so it is a very tedious operation to undertake.

 

[Creepy Old Man]

I've been working on acquiring these battle logs, but unfortunately have been plagued with a series of bugged battles lately. (Units attacking friendly units; progress bars moving the wrong way; casualties not being taken). Thus I haven't added any new logs in about two days. Overall I have details for about 150 rounds of combat. However, some of those are old enough they predate the patch in which it was claimed that sides with an overwhelming advantage will attack even more often - and I don't know which ones those are.

So far, I have come to agree with the above. The combination of <unit strength> x <casualties> x <stance modifier> is essentially constant round to round. The value of that constant changes slightly as combat progresses, presumably due to changing total army strengths.

However, the constant is *not* simply a multiple of any of the following: attacker strength; defender strength; total strength. Changing the stance of a unit (and therefore the strength of that side) changes the damage taken by the opposing side, but not in a linear way. Due to my inability to observe unbugged battles lately, I haven't been able to make any guesses as to what the function is.