Retinues in 2.5.2?

[Kardo]

So what are the ideal retinues (besides pikeman) to use? I know that pure light cavalry and horse archers were nerfed with the new tactics, but was anything elselse nerfed?

Edit: I'm also interested in horde units as well.

 

[Kumicho]

I've just given up on retinues, and RP whatever cultural ones I get...

Between the tactics nerf and their drastically higher cost compared to your own levies, I just castle-stack in my capital and use the retinues as flavor. The couple hundred I get as a small Kingdom certainly aren't going to make any difference anyway.

Maybe if we're done with patches I might dig into the files and redo them for my own games, but I just haven't bothered yet.

 

[Shebaloso]

So what are the ideal retinues (besides pikeman) to use? I know that pure light cavalry and horse archers were nerfed with the new tactics, but was anything elselse nerfed?

Edit: I'm also interested in horde units as well.

The pure heavy infantry ones (norse, saxon, etc.) are very good also, but must be coupled with one "shock" retinue for the archers in order not to fire 100% generic skirmish in the skirmish phase

Pure light cavalry isn't that bad, it's only not extremely overpowered anymore. You have to have <75% light cavalry in order to fire standard harass AND a 12 martial commander in order to fire standard Harass. This translates into 3 heavy vanguard:4 pure LC for hordes, and i believe the west european knight retinue also does reasonably well.

Everything else is garbage. Horse archers are mega garbage and actually lose very frequently to levies. I believe this is due to horse archers also counting as LC for the purpose of tactics eligibility. It's possible have a 100% horse archers flank and have Harass and Raid (which are light cavalry tactics and actually cripple the horse archers) fire. I don't know if this is a bug, but, knowing the developers' laziness with retinues and combat in general, can't say for certain

 

[Ragingnorman]

I want to build a retinue focused on light cavalry (Caballeros), you suggest that I keep it 100% LC or mix it with something else to avoid the disorganised tactics?

Also, are Andalusian Cavalry just a better version of Caballeros for no apparent reason?

 

[Shebaloso]

The main point of mixing retinue types is not only avoiding bad tactics, but also having good tactics becoming eligible with a reasonable probability. For example, you could mix pure LC retinue with others such as pikeman, heavy infantry etc. , so you would go below the 75% threshold. You would not fire disorganized harass, but by having HI/pikemen in that flank, you would make shieldwall, force back, etc eligible, and some tactics that would actually be bad for LC.

Essentially, you have to see what tactics would become eligible with the mix you are going to use. That's why Heavy infantry and archers, for example, are a good mix. A pure heavy infantry flank would only have generic skirmish tactic in the skirmish phase. By adding some archers, you would make "shieldwall" eligible, thus decisively reducing the amount of casualties in the first phase. But add too many archers, and they will die by the dozens in the shock phase.

You can see in this page the tactics in ck2: http://www.ckiiwiki.com/Combat_tactics

As for your question, Pure LC is superior to any mix of LC AND standard retinues. Be advised, though, that a pure "defensive" retinue is superior to every other possible mix in the game, so only use LC if you want to rp.

 

[Pode]

a pure "defensive" retinue is superior to every other possible mix in the game

Everyone keeps quoting this, nobody shows the math to back it up. If battles were pure melee I'd buy it, but I'm seeing good melee retinues get broken in skirmish before they can close to melee more often than not. Since pikemen can't charge and force melee, I'm skeptical enough that I've started doing the math myself. Unfortunately I don't have the game files at work and the CKII wiki is seriously out of date.

 

[szmik]

From my experience most fights is decided before melee phase actually fires, so mixing good skirmishers in other retinues is actually good.
Also my pure LC/HA horde army beats anything with maxed out capital buildings.

 

[Kumicho]

The pure heavy infantry ones (norse, saxon, etc.) are very good also, but must be coupled with one "shock" retinue for the archers in order not to fire 100% generic skirmish in the skirmish phase

Pure light cavalry isn't that bad, it's only not extremely overpowered anymore. You have to have <75% light cavalry in order to fire standard harass AND a 12 martial commander in order to fire standard Harass. This translates into 3 heavy vanguard:4 pure LC for hordes, and i believe the west european knight retinue also does reasonably well.

Everything else is garbage. Horse archers are mega garbage and actually lose very frequently to levies. I believe this is due to horse archers also counting as LC for the purpose of tactics eligibility. It's possible have a 100% horse archers flank and have Harass and Raid (which are light cavalry tactics and actually cripple the horse archers) fire. I don't know if this is a bug, but, knowing the developers' laziness with retinues and combat in general, can't say for certain

I could be wrong (probably am), but Disorganized Harass *can* fire with anything more than 40% LC, and it's chance of firing is increased by the % of LC you have. So even at 70% LC, you still have a ridiculously high ratio of rolling DH. No clue why they have the first "light_cavalry = 0.4" when they have the second bit about the flank's martial ability.

disorganised_harass_tactic = {
    days = 18
    sprite = 4
    group = harass

    trigger = {
        phase = skirmish
        OR = {
            light_cavalry = 0.4
            camel_cavalry = 0.4
        }
        OR = {
            flank_has_leader = no
            NOT = { leader = { Martial = 12 } }
            OR = {
                light_cavalry = 0.75
                camel_cavalry = 0.75
            }
        }
        NOT = {
            horse_archers = 0.5
        }
    }

    mean_time_to_happen = {
        days = 2
        modifier = {
            factor = 1.5
            OR = {
                light_troops = { 
                    who = light_cavalry
                    value = 0.5
                }
                light_troops = { 
                    who = camel_cavalry
                    value = 0.5
                }
            }
        }
        modifier = {
            factor = 1.5
            OR = {
                light_troops = { 
                    who = light_cavalry
                    value = 0.7
                }
                light_troops = { 
                    who = camel_cavalry
                    value = 0.7
                }
            }
        }
        modifier = {
            factor = 1.5
            OR = {
                light_troops = { 
                    who = light_cavalry
                    value = 0.9
                }
                light_troops = { 
                    who = camel_cavalry
                    value = 0.9
                }
            }
        }
    }

    light_cavalry_offensive = 1.0
    camel_cavalry_offensive = 1.0
    horse_archers_offensive = -1.5
    archers_offensive = -1.5
    light_cavalry_defensive = -0.6
    camel_cavalry_defensive = -0.6
    enemy = {
        group = swarm
        factor = 1
    }
}

 

[Dragatus]

I want to build a retinue focused on light cavalry (Caballeros), you suggest that I keep it 100% LC or mix it with something else to avoid the disorganised tactics?

If you insist on using Caballeros mix in 1 Light Skirmish retinue for every 3 Caballero retinues and make sure your commander has 12+ Martial.

You'll be getting either Harass or Volley Harass in skirmish phase and Raid in melee.

Also, are Andalusian Cavalry just a better version of Caballeros for no apparent reason?

Yes, they are.

Everyone keeps quoting this, nobody shows the math to back it up. If battles were pure melee I'd buy it, but I'm seeing good melee retinues get broken in skirmish before they can close to melee more often than not. Since pikemen can't charge and force melee, I'm skeptical enough that I've started doing the math myself. Unfortunately I don't have the game files at work and the CKII wiki is seriously out of date.

It was true in 2.4 when morale loss factor was 1 and armies fought to death. But in 2.5 they increased it to 3, so armies break much faster. This is something I only realized now, so I failed to account for it in previous threads about retinues. I'm not sure if the effect is big enough to dethrone Pikemen though. I'd have to do a new analysis.

The wiki seems to be more or less up to date on unit stats, tactics, and retinues. The combat formula is a bit off, but through experimentation I've been able to figure it out.

1. Total attack values of all soldiers in a flank are added together and multiplied by 0.01 to get damage.
2. Damage is divided by number of soldiers in enemy flank to get damage per soldier.
3. For each unit type damage per soldier is divided by that unit's defence value and multiplied by the number of soldiers of that type. This gives us number of deaths per unit type.

Morale is weird. I failed to figure out how exactly it's determined, but I did find that morale loss scales linearly with both damage/kills and the moraleloss factor setting.

A few observations:
1) The power of a unit scales linearly with attack and defence.
2) The power of a unit scales with the inverted square of a unit's cost. If a unit has half cost, you can get twice as many, giving you both double damage as well as twice as many bodies to soak up kills.

I could be wrong (probably am), but Disorganized Harass *can* fire with anything more than 40% LC, and it's chance of firing is increased by the % of LC you have. So even at 70% LC, you still have a ridiculously high ratio of rolling DH. No clue why they have the first "light_cavalry = 0.4" when they have the second bit about the flank's martial ability.

disorganised_harass_tactic = {
    days = 18
    sprite = 4
    group = harass

    trigger = {
        phase = skirmish
        OR = {
            light_cavalry = 0.4
            camel_cavalry = 0.4
        }
        OR = {
            flank_has_leader = no
            NOT = { leader = { Martial = 12 } }
            OR = {
                light_cavalry = 0.75
                camel_cavalry = 0.75
            }
        }
        NOT = {
            horse_archers = 0.5
        }
    }

    mean_time_to_happen = {
        days = 2
        modifier = {
            factor = 1.5
            OR = {
                light_troops = {
                    who = light_cavalry
                    value = 0.5
                }
                light_troops = {
                    who = camel_cavalry
                    value = 0.5
                }
            }
        }
        modifier = {
            factor = 1.5
            OR = {
                light_troops = {
                    who = light_cavalry
                    value = 0.7
                }
                light_troops = {
                    who = camel_cavalry
                    value = 0.7
                }
            }
        }
        modifier = {
            factor = 1.5
            OR = {
                light_troops = {
                    who = light_cavalry
                    value = 0.9
                }
                light_troops = {
                    who = camel_cavalry
                    value = 0.9
                }
            }
        }
    }

    light_cavalry_offensive = 1.0
    camel_cavalry_offensive = 1.0
    horse_archers_offensive = -1.5
    archers_offensive = -1.5
    light_cavalry_defensive = -0.6
    camel_cavalry_defensive = -0.6
    enemy = {
        group = swarm
        factor = 1
    }
}

The 0.4 condition is included because it's a minimum. In order to get Disorganized Harass you need to first have 40% light cavalry and then you also need to fulfill one of three aditional conditions:
1. Flank has no leader.
2. Flank does have a leader, but his Martial is lower than 12.
3. The flank is 75% light cavalry.

 

[Talq]

I could be wrong (probably am), but Disorganized Harass *can* fire with anything more than 40% LC, and it's chance of firing is increased by the % of LC you have. So even at 70% LC, you still have a ridiculously high ratio of rolling DH. No clue why they have the first "light_cavalry = 0.4" when they have the second bit about the flank's martial ability.

It means it can fire if you have >40% LC and no commander or the commander has martial less than 12 and will fire regardless of commander if LC >75%. (ie you need a commander with martial >= 12 if you want to use lots of LC without getting disorganised harass).

Edit: what Dragatus said.

. A pure heavy infantry flank would only have generic skirmish tactic in the skirmish phase. By adding some archers, you would make "shieldwall" eligible, thus decisively reducing the amount of casualties in the first phase.

Yes, its not so much 100% generic skirmish, but that adding archers adds shieldwall, which gets you out of skirmish phase quicker.

 

[Shebaloso]

Everyone keeps quoting this, nobody shows the math to back it up. If battles were pure melee I'd buy it, but I'm seeing good melee retinues get broken in skirmish before they can close to melee more often than not. Since pikemen can't charge and force melee, I'm skeptical enough that I've started doing the math myself. Unfortunately I don't have the game files at work and the CKII wiki is seriously out of date.

the math is actually very simple, it just takes some work.

https://www.reddit.com/r/CrusaderKings/comments/4533oo/optimal_retinue_calculator_conclave/

Essentially, a pure pikemen retinue will endure the skirmish phase with some losses, but rape every other retinue in the shock phase. They are good because their damage output in the shock phase is extremely high due to firing good tactics such as stand fast and force back, which also happens to have a huge bonus against common shock tactics such as the "charge" group.

 

[Pode]

It was true in 2.4 when morale loss factor was 1 and armies fought to death. But in 2.5 they increased it to 3, so armies break much faster. This is something I only realized now, so I failed to account for it in previous threads about retinues. I'm not sure if the effect is big enough to dethrone Pikemen though. I'd have to do a new analysis.

The wiki seems to be more or less up to date on unit stats, tactics, and retinues. The combat formula is a bit off, but through experimentation I've been able to figure it out.

1. Total attack values of all soldiers in a flank are added together and multiplied by 0.01 to get damage.
2. Damage is divided by number of soldiers in enemy flank to get damage per soldier.
3. For each unit type damage per soldier is divided by that unit's defence value and multiplied by the number of soldiers of that type. This gives us number of deaths per unit type.

Morale is weird. I failed to figure out how exactly it's determined, but I did find that morale loss scales linearly with both damage/kills and the moraleloss factor setting.

A few observations:
1) The power of a unit scales linearly with attack and defence.
2) The power of a unit scales with the inverted square of a unit's cost. If a unit has half cost, you can get twice as many, giving you both double damage as well as twice as many bodies to soak up kills.

So I finally did that new analysis based on the mechanics Dragatus described. I assumed that morale damage is treated like regular damage, just with some extra modifiers possible, the total morale damage for a flank is multiplied by the morale loss multiplier of 3, and the defenders' morale stat determines casualties instead of defense.
If that's true, the ratio of enemy casualties to friendly casualties for a flank size of one, no tactics or other bonuses, and a token enemy with all stats = 1 works out to be 4 * Attack / (1/Defense + 3/Morale). Which fits with Dragatus' observations above.
Since we don't often field armies of 1, I computed kill-to-loss ratios for each retinue unit versus 1000 of the all-1's baseline troops using generic tactics. I also ran those calcs for the most beneficial tactic for each retinue unit (regardless of whether / how often that tactic triggers for a pure flank).
Casualties are computed every day, so this kill to loss ratio gets compounded each day. Assuming that combat lasts for 18 days of skirmish and 18 days of melee, actual casualties would be (skirmish loss ratio)^18th * (melee loss ratio)^18th. So the relative power of a unit can be compared by multiplying the skirmish and melee daily loss ratios together. I chose to take the square root of this product to keep the units the same, enemy losses / friendly losses.

I come up with pure cavalry (of every type except horse archer retinues) being by far the way to go for field battles and the consensus pike units only being worthwhile in assaults. The pikes are the best killers in melee (about 4 times better than the best killers in skirmish), but they're SO BAD in skirmish that they don't even break the top ten overall. I looked at loss ratio, loss ratio per cap point, & loss ratio per man (for attrition). Defense retinues showed up tenth in the top ten lists for loss ratio and loss ratio per cap, the other 28 spots on the 3 top ten lists went to cav units. Charges on the 12th day vs the 18th will bias things even more away from pikes and towards cav. Pikes will make that up in prolonged melee, but how often does that happen?

Combat power (root of skirmish loss ratio * melee loss ratio) top 10:
Berber (27.2), Andalusian (23.0), Camel (20.9), Caballero (20.7), Cataphract (19.5), Knight (19.2), Gusar (18.8), Cavalry (15.5), Hussar (14.9), Defense (7.3)
All using Harass and Heroic Countercharge, except Cataphract (Harass Swarm and Couched Lance Charge) and Defense (Volley and Schiltron).

Combat power per man:
Same units, except Horse Archer retinues using Altaic tactic replace Defense in the number 10 slot. Camels are best. Cataphracts, Knights, and Gusars are pretty much tied for runners up.

Probably most important, combat power per cap:
Camel (27.8), Berber (22.6), Andalusian (19.2), Gusar (17.9), Caballero and Cavalry (17.2), Hussar (16.6), Knight (16.0), Cataphract (15.0), Defense (12.2)

A dedicated unit of assaulters should be Italian or Scottish pike unless you're attrition limited, where you want Knights.

Lastly, ignoring tactics (like if you want some cushion to cover for bad tactics rolls), Berbers are best, Camels are best per cap, and Cataphracts are best per man

 

[Dragatus]

So I finally did that new analysis based on the mechanics Dragatus described. I assumed that morale damage is treated like regular damage, just with some extra modifiers possible, the total morale damage for a flank is multiplied by the morale loss multiplier of 3, and the defenders' morale stat determines casualties instead of defense.

If that's true, the ratio of enemy casualties to friendly casualties for a flank size of one, no tactics or other bonuses, and a token enemy with all stats = 1 works out to be 4 * Attack / (1/Defense + 3/Morale). Which fits with Dragatus' observations above.
Since we don't often field armies of 1, I computed kill-to-loss ratios for each retinue unit versus 1000 of the all-1's baseline troops using generic tactics. I also ran those calcs for the most beneficial tactic for each retinue unit (regardless of whether / how often that tactic triggers for a pure flank).
Casualties are computed every day, so this kill to loss ratio gets compounded each day. Assuming that combat lasts for 18 days of skirmish and 18 days of melee, actual casualties would be (skirmish loss ratio)^18th * (melee loss ratio)^18th. So the relative power of a unit can be compared by multiplying the skirmish and melee daily loss ratios together. I chose to take the square root of this product to keep the units the same, enemy losses / friendly losses.

I come up with pure cavalry (of every type except horse archer retinues) being by far the way to go for field battles and the consensus pike units only being worthwhile in assaults. The pikes are the best killers in melee (about 4 times better than the best killers in skirmish), but they're SO BAD in skirmish that they don't even break the top ten overall. I looked at loss ratio, loss ratio per cap point, & loss ratio per man (for attrition). Defense retinues showed up tenth in the top ten lists for loss ratio and loss ratio per cap, the other 28 spots on the 3 top ten lists went to cav units. Charges on the 12th day vs the 18th will bias things even more away from pikes and towards cav. Pikes will make that up in prolonged melee, but how often does that happen?

Combat power (root of skirmish loss ratio * melee loss ratio) top 10:
Berber (27.2), Andalusian (23.0), Camel (20.9), Caballero (20.7), Cataphract (19.5), Knight (19.2), Gusar (18.8), Cavalry (15.5), Hussar (14.9), Defense (7.3)
All using Harass and Heroic Countercharge, except Cataphract (Harass Swarm and Couched Lance Charge) and Defense (Volley and Schiltron).

Combat power per man:
Same units, except Horse Archer retinues using Altaic tactic replace Defense in the number 10 slot. Camels are best. Cataphracts, Knights, and Gusars are pretty much tied for runners up.

Probably most important, combat power per cap:
Camel (27.8), Berber (22.6), Andalusian (19.2), Gusar (17.9), Caballero and Cavalry (17.2), Hussar (16.6), Knight (16.0), Cataphract (15.0), Defense (12.2)

A dedicated unit of assaulters should be Italian or Scottish pike unless you're attrition limited, where you want Knights.

Lastly, ignoring tactics (like if you want some cushion to cover for bad tactics rolls), Berbers are best, Camels are best per cap, and Cataphracts are best per man

I think you made an error in your daily loses ratio formula. Morale loss is calculated from casualties suffered, rather than directly from damage.

Your formula: ( 4 * Attack ) / ( 1 / Defense + 3 / Morale )
Corrected formula: ( 4 * Attack ) / ( 1 / Defense + 3 / [ Defense * Morale ] )
Alternative formula: ( 4 * Attack ) / ( 1 / Defense * [ 1 + 3 / Morale ] )

I also have some doubts about simply adding together both casualties and morale loss, but it's nearly 4 AM for me at the moment so I'll let it be for now.

Another thing that I noticed is that you seem to have listed Volley as the optimal skirmish tactic for a Defense retinue, when Shieldwall is much better (+300% HI defence, +240 Pikeman defence, +60% Archer offence) and now only requires 1% Archers to trigger.

I would also recommend using Force Back as the melee tactic for any sort of Pikeman retinue as that is the one that is realistically going to trigger in most situations.

EDIT: OK, I've gotten some sleep now and can think more clearly.

And I'm sorry, but your method of calculating combat power is wrong. That is probably easiest to demonstrate by imagining a hypothetical unit that has 0 skirmish attack and a ridiculously high 10.000 in all other attributes. Following your method of calculation this unit's combat power would be rated as 0 because it doesn't do any damage at all in skirmish phase and both ratios are simply multiplied. However, what would actually happen in a battle is that the unit would indeed do no damage in skirmish but it would also take virtually no losses and as soon as the battle enters melee phase it would instantly rout the enemy.

Pikemen and Heavy Infantry are essentially doing a less extreme version of the same thing. They take relatively low loses during skirmish (at least if they trigger Shieldwall) and then do very well in melee.

However, while it is relatively simple to point out that the method is flawed by providing a counter-example, it is more difficult to actually come up with an alternative formula that works better. That's something I still need to work on myself.

I'm fairly sure morale is a mostly linear benefit. If you have two units who have identical attributes in all categories, except one has double morale, that one will simply be able to take twice as many loses before it routs. This is messed up somewhat by the fact that units also recover some of their morale on a daily basis, but let's not go into too much detail at this point.

Attack and defence however, are non-linear benefits. If a unit has twice as much attack it will kill enemies twice as fast, but because the enemy is taking higher casualties they will also have less soldiers doing damage on any given day of the battle. A high defence on the other hand will not only keep you longer in the fight but will also help sustain your damage output. What I'm still struggling with is evaluating exactly how non-linear these effects are.

But even if your calculation was incorrect, it was helpful because it showed me some things I hadn't though of myself. Multiplying skirmish and melee performance might still work if instead of multiplying the ratio of loses we multiplied the ratio of survivors. But it would get complicated. Something like this:

performance over 36 days of battle = ( day 1 skirmish ratio between friendly and enemy survivors ) * (skirmish day 2 ratio between friendly and enemy survivors ) * ... * ( skirmish day 18 ratio between friendly and enemy survivors ) * ( melee day 1 ratio between friendly and enemy survivors ) * ... * (melee day 18 ratio between friendly and enemy survivors ) * (ratio of friendly and enemy morale )

In the end it might be easier to simply test units in game with a mod specifically designed for combat testing. Mod in a new special unit and create a generic retinue of it. Play as a kingdom and make 1000 of them. Exit the game and edit the stats of the unit to match the unit you wish to test for if it triggers the tactic you wish to test for. Enter the game and play a hostile neighbour of the original kingdom and attack the retinue with a levy. Write down results. Exit the game, change special unit stats to match the second unit you wish to test for and continue the procedure.

 

[Narvait]

Do you know if unit related leadership traits are now working properly?
I.e. does heavy infantry leader helps heavy infantry in way prescribed?

 

[Dragatus]

I didn't even know they weren't working properly. I might run a test after I get home.

 

[Narvait]

I didn't even know they weren't working properly. I might run a test after I get home.

This got reported back then, not sure on status
https://forum.paradoxplaza.com/foru...no-unit-type-leader-trait-not-working.851585/

 

[Pode]

You hit on the biggest flaw in my reasoning, unless losses are computed differently, more on that thought in a bit.

It's not the ratio of losses that matters but the ratio of survivors. Compound that each day of combat & we've got it.

The new thought I had today was about morale casualties, morale damage, & kills. I think perhaps the reason people have always had trouble figuring morale out is that maybe kills cause some troops to panic & morale damage causes additional troops to panic. So losses = kills + 3*kills from panic + (1.5% * total attack * morale damage bonus) / (total morale * morale defense bonus). Since so few troops have equal morale and defense, this would make the panic "kills" due to morale loss weirdly nonlinear with attack like so many of us have complained about all along. Thoughts?

Computer died today, so I'll be slow in reworking all this in terms of troops remaining each day, but I think we're close. Testing mod is a great idea to sanity check all this.

 

[Dragatus]

Sorry to hear about your computer. My old one died last October so I remember well how that feels.

I've done some combat tests with a mod specifically designed for that a while back. In the mod I set morale recovery to 0, eliminated all tactics except generic skirmish, eliminated all bonuses to attack, defence, or morale, modified traits so all characters would have 0 martial and no combat bonuses, and set all the values for all the troops to 1. Then I ran games as Harold Godwinson fighting against William the Bastard. I'd record data from the fight, then go change a variable and see how that changes things.

One major thing I noticed is that the numbers don't fit. In one case for example the interface showed me that I did 50 damage, the displayed kills were 48, but the actual difference between displayed troops before and after that day of battle was 47. My hypothesis is that this is due to rounding errors. It also didn't seem to affect the general trend too much, so I mostly ignored it.

What I've been able to find out about morale is that a flanks morale scales linearly with the sum total morale of all the troops in it. There seems to be a multiplier of some sort, as the displayed morale is some 12%-20% higher than just the sum total. There may also be a base value, I'm not sure. Defenders also seem to get a small bonus to morale compared to attackers. For example in one of the test battles the attackers had about 13% extra morale while the defenders had about 18% extra morale. Extra in this case means in addition to the sum total of morale of all troops.

But the main thing to take away here is that a flank's maximum/full morale scales linearly with the sum total of the morale values of all the individual soldiers.

After taking damage a flank's morale is reduced by a number that scales linearly with damage/kills, the morale loss multiplier and an unknown factor. In a major test I recorded all relevant morale information for all 6 flanks in a battle and the ratio between how much a flank's absolute morale value (absolute as in absolute numbers rather than percentile value) was reduced and the flanks loses varied wildly between the flanks, but was consistent within a specific flank. The unknown factor appears to be related to distribution of troops among the flanks. It may also differ between attacker and defender, but I haven't done any tests to confirm it.

If you're interested, here is an excel table with all the data I recorded: https://www.dropbox.com/s/70z5hjamu4ugvks/ck2morale.xls?dl=0

You'll want to look at sheet3. The first two sheets are a bit of a mess. But sheet 3 are the records of the first 5 days of two battles between the same two armies.

It may be a little hard to read at first as I made it for my own purposes and not to share with other people. The data was taken from two battles and I was the attacker in both. In the first one I spread my troops evenly between the three flanks. These flanks are marked as A, B, and C from left to right. Enemy flanks opposing them are Alpha (opposing A), Beta (opposing B), and Gama (opposing C). After the battle I reloaded the save from just before the battle started and moved all troops from flank C to flank A and recorded data for that too.

For each flank I recorded how many soldiers there were (marked by the designation of the flank), what was the displayed morale for the flank (the number shows how much morale a flank would have if it were at full morale), and the displayed percentage of morale for the flanks (morale %). The next column (delta m) computes how much morale a flank lost per day of battle and the following three record the displayed damage (damage), displayed kills (killed), and loses computed from difference in the displayed number of troops (loses). In the three columns following that I divided delta m with damage, killed and loses respectively and in the column after that I averaged the previous three values. The final column was an afterthought. It simply divides a flanks morale with the number of soldiers in it to show the general morale multiplier.

All combat values for all the units are 1 (including unit morale) and the morale loss factor was set to 3 in defines (default value). A mistake I did in my testing is that I never checked whether morale loss (delta m) is based on damage taken or loses, but in this test run the two numbers are supposed to be identical (because all defence values are 1).

EDIT: A little update. I tested it and morale loss is based on how many of your troops have died and not on enemy damage.

 

[Narvait]

Dragatus, did you check on unit based leadership traits

 

[Dragatus]

Dragatus, did you check on unit based leadership traits

I hadn't before, but I did just now. And they do work. In fact they work much better than expected. I added a 100% bonus to all troop types to all the martial educations and the damage went through the roof.

By the looks of it a 100% troop bonus will square the damage done.