On War: How to understand what your generals are doing

Explanation of what is going on under the hood to make the combat mechanics operate, like why your general is only attacking with 3 of 300 battalions or why you only captured one province after a battle.

Introduction

14 Nov 2023 Update

Before reading, note that this guide has not been updated to account for the new war mechanics in the 1.5 patch. I may update it once I've had a chance to dig into those mechanics.

So you bought Victoria 3, you started playing, you got into a Diplomatic Play that escalated to war, and now you're wondering why your generals aren't doing what you want them to do.

The developer diaries and the Victoria 3 Wiki explain the general thrust of the mechanics but they don't reveal the specifics. This guide will cover some additional information from the game files that sheds a bit more light on what's going on.

After you're done here, check out my mod BattleCalc to see some of the key numbers discussed here in the game.

Fronts and Armies

There are two rules for how fronts are created:

• Fronts must be between two contiguous blocks of territories. When Mexico fights the USA there is one front between the USA and Mexico and one front between the Indian Territory and Mexico. The USA-Mexico front will be discontinuous because the Indian Territory-Mexico front is in the middle of it, but it's still one front and generals will operate on both sides of it. If you advance and cut off a pocket of enemy territory, a new front forms around the pocket.
  
• Fronts are between exactly two countries. If you're fighting two allied enemies there will be two fronts, one against each. If the Canadian colonies wind up helping Mexico fight the USA, the USA will have a front with each of the Canadian colonies.

In this example I only have one front with the East India Company, even though I have two borders with them. That's because their land all connects together.


Each general can be assigned to exactly one front. If they advance or retreat and that creates a second front, then the generals on both sides will pick one of the fronts to fight on. If the enemy had two generals on that front and you only have one, then one of their generals will be able to push unimpeded. In large wars, then, it's always better to have more generals than fronts in each theatre, so you always have someone to reassign if new fronts emerge.

Troops cannot currently be destroyed in encirclements in the game; if a pocket is cleared out, the generals defending it will move to the nearest remaining front. This is annoying when it saves an AI army, but helpful when it saves your army from a boneheaded mistake by one of your own generals.

How Battles Happen

If at least one general on a side of the front is set to "Advance Front" and there is no current battle on the front, then that side will gain progress toward starting a new battle. This is almost always +10 progress per day, reduced by 50% if the defending side has a general set to "Defend Front".
When the meter reaches 100% then the battle starts. By default every province on the defender's side of the front is equally likely to be the site of the battle. Provinces get their probability of being selected modified:
Capital Bonus is 10. Wargoal Bonus is 10 if the province is in a wargoal state; 6 if it's in a state bordering a wargoal; and 3 if it's in a state adjacent to one that borders a wargoal. This means that if you're trying to take the enemy capital in the war, provinces in the capital state would be 100 times more likely to be the site of the next battle than other provinces with the same terrain and infrastructure.

Generally speaking, then, the battle is more likely to happen in high-infrastructure provinces with better terrain, and much more likely to happen in a province that's either part of the wargoal or gets the front closer to a wargoal.

If neither side has a general set to "Advance Front" then the two sides will just stare at each other and take attrition. If your units are fully supplied and assigned to a front then they take an average of 2% casualties per week due to attrition. If they're on "Stand By" then they're living in barracks and don't take attrition.

How your general picks how many troops to bring

First the game picks which general is going to command each side. The generals are weighted by the number of troops they command, and the weight is increased if they have the corresponding attack or defence order.

Each side starts with a "baseline" number of troops.

• If it's an offensive battle on a regular front, this is the total number of battalions with the Advance Front order on that side of the front.
  
• If it's a defensive battle on a regular front, this is the total number of battalions assigned to that side of the front (regardless of their orders) plus the number of battalions Standing By or garrisoned in the strategic region's HQ. If multiple fronts are present in the same strategic region, then the battalions garrisoned or Standing By in that region's HQ may contribute to defensive battles on all of the fronts!
  
• If it's a naval invasion, the attacker's baseline is the number of mobilized battalions under the attacking general and defender's baseline is the number of battalions Standing By or garrisoned in the defending Land HQ.

The baseline is capped based on the terrain and combat width. The formula is:
Terrain Combat Width is a value between 0 and 1; it's 1 in plains, 0.3 in mountains, and somewhere in between for other terrain types. So if the battle is in a state with 10 Infrastructure (common in low-population areas) and it's happening in mountainous terrain, the cap is (5 + (10 / 2)) * 0.3 = 3 battalions. If these low-infrastructure mountains are the only place battles can happen on that front, then the baseline will always be 3 battalions for battles on that front, until it moves to better terrain/infrastructure.

Further modifiers are then applied to the baseline:

• If the attacking troops have average offense higher than the defenders' average defense, then the defenders multiply their baseline by a random value between 1 and 3 that's also limited by how big the difference in stats is:
• If the attacking troops have average offense lower than the defenders' average defense, then the attackers multiply their baseline in a similar way, but limited to 2 instead of 3:
• These multipliers still can't increase the number of troops beyond the number available on the front. On the attacking side, this cap doesn't include troops under generals set to "Defend Front".
  
• Finally, the attacker's number of troops is reduced to a random value between 33% and 100% of the previously-calculated amount, and the defender's total is reduced to between 50% and 100% of the previously-calculated amount.

See the Appendix to this guide for some more math on this subject.

So as an example, for this front there's a battle happening in some Plains:
It's happening in the state of Hebei, which has 113 infrastructure right now in my game. So the terrain baseline cap for both sides of the battle is:
My Canadian troops have an average offense of 143 and the defending Chinese troops have an average defense of 43, so they increase their baseline:
This means China might have been able to bring up to 61.5 * 1.998 = 122.9 battalions depending on how well they rolled.

The attackers have better stats so they don't get an increase.

Lastly, the attackers get reduced by up to two-thirds and the defenders get reduced by up to half. So Canada could have brought anywhere between 0.33 * 61.5 = 20.5 and 61.5 battalions, and 52 as shown in the screenshot was a pretty good roll for my generals. Meanwhile China could have brought as many as 122.9 (if they got the maximum value for both rolls) or as few as 30.75 (if they got the minimum for both).

The commanding general will always draw from their assigned troops before borrowing any from other armies on the same front. When they're deciding which allies to borrow from, they're supposed borrow more troops from their own country than from allies and they're supposed to prefer borrowing troops with higher morale; however, the morale part of this appears to be bugged and the part about allies might be too.

How the battle is won

Each battalion targets at most one battalion on the other side. The battalions change targets in each round/tick of combat. It appears to be the case that multiple battalions cannot target the same enemy; see for example this battle where the single defending battalion has inflicted over 400 casualties while taking only 129 casualties from the thirty-two battalions attacking it:


Each round of combat inflicts casualties on both armies. The attacking battalions inflict more casualties and receive fewer casualties if their Offense is higher than the defender's Defense, and vice versa. The damage inflicted is scaled based on the remaining manpower of the battalion doing the shooting.

In this example I'm inflicting disproportionate casualties despite having been outnumbered over three to one at the start of the battle; this is due to the stats advantage.

Casualties can be either Dead or Wounded. By default 75% are Wounded and the rest are Dead. Some of the Wounded will rejoin their unit while others will become Dependants of a pop somewhere in your country. The number of Dead is reduced by having better Medical Aid production methods for your units, and it goes up if the opposing army has a higher Kill Rate (from better artillery PMs, Machine Guns, Flamethrowers, or Chemical Weapon Specialists).

In the screenshot above, 85% of my casualties are surviving the battle, and only 74% of the Chinese casualties are surviving. This is because, while we both have good medical branches attached to our armies and giving a boost to Recovery Rate, my Kill Rate is a lot higher:

When a battalion takes casualties, it also loses morale. Morale is effectively a multiplier for the manpower of the unit; if it has 50% morale then half of its troops won't be taking part in the fighting. Absent any modifiers for morale damage or resistance, each casualty received (killed or wounded) will also demoralize one surviving soldier in that battalion. Therefore, without morale modifiers on either side, the battle will be over once one side has taken 50% casualties, with the other 50% of its troops demoralized.
Demoralized troops cannot be recovered during a battle.
Morale is also reduced by having low supply for the army. This can keep an army stuck at 0% morale, in which case it auto-loses battles as soon as they start. You can push a front very fast when this happens, so convoy raiding to cut off supply to a front can be very powerful.

Battalions leave the battlefield if they are reduced to zero morale or zero manpower. The battle ends only when one side has no battalions left on the field; there doesn't appear to be any mechanic for retreat before this point.

Battalions can reinforce during the battle, if the building they come from is able to hire people to replace the casualties.

What happens after the battle

If the attacker wins then they advance a certain number of provinces into opposing territory.

The number of provinces taken appears to depend on:

• the width of the front (the wider the front the farther the attacker advances)
  
• the number of battalions the attacker had left at the end of the battle (larger battle means more territory won)
  
• the density of the front (the greater the ratio of battalions to provinces is, the fewer provinces taken)

Some barracks Production Methods have an effect too:

• Trench Infantry and Chemical Weapon Specialists reduce the number of provinces captured by their army
  
• Mechanized Infantry, Mobile Artillery, all of the Reconnaissance PMs, and Infiltrators increase the number of provinces captured by their army
  
• Trench Infantry, Squad Infantry, and Machine Gunners reduce the number of provinces lost in a defeat

So if you're confident that you can win battles easily against your foes, selecting Chemical Weapon Specialists is counterproductive because you will still win with Infiltrators and you'll capture land faster.

Naval Invasions

• You can't start planning a Naval Invasion until the war breaks out.
  
• The fleet doing the invasion can transport exactly one army, and it needs to be an army commanded by a general who's from a HQ in the same region as the fleet's base.
  
• The fleet takes time to prepare (AFAIK it's always 50 days) and then launches the invasion.
  
• The army needs to travel back to its home HQ before the invasion can commence. This may take longer than 50 days.
  
• The invasion will not start until the battalions assigned to the general are fully mobilized. This may lead to a considerable delay if there are partiallly-mobilized conscripts in the HQ that get assigned to him.
  
• If an enemy has a fleet set to defend the coast of the strategic region targeted by the invasion, then your fleet has to fight their fleet first.
  
• If your fleet wins, then the army being transported starts the land battle. It will be facing any and all armies currently Standing By or garrisoned (not mobilized) in the same strategic region as the invasion target state.
  
• If you haven't researched Landing Craft (Military, Tier IV) then your army takes a 25% penalty to offense during this battle.
  
• If your army has more battalions than the supporting fleet has flotillas, then the army takes another penalty to offense, scaled based on how many more battalions that ships there are.
  
• If your army wins then you capture some land and a new front opens. Then you can assign additional armies to this front.
  
• It is possible for the invading army to start a second battle in another province while the first one is still happening. I have no idea what conditions are required to make this happen, but I've seen it twice.

Tips

• Having all your troops under one general or splitting them up to be under several generals doesn't affect the number that join a battle. The only reason you'd want to concentrate troops under one general is if he has really good traits, since a general doesn't give any boosts to troops he borrows from another commander.
  
• Having multiple generals assigned to the front means that you can more easily handle unexpected front line splits. If your generals are of equal skill to each other then this is a good idea.
  
• If you keep having battles with only a tiny number of troops engaged, it's probably because you're fighting in bad terrain and/or bad infrastructure. You cannot fix this by adding more troops, only by moving the front line out of the bad terrain/low infrastructure. Consider focusing on another front, or launching a naval invasion to open up a new front that's easier to push. If you're the war leader, you can also use Violate Sovereignty to demand passage through a country that borders you and the enemy war leader, and invade them if they refuse; this will create a new front either way.
  
• If you know the whole front is bad terrain and low infrastructure, don't send hundreds of battalions to it; they'll just be sitting around dying of typhoid or whatever it is that causes attrition.
  
• Having more troops than your opponent does provide a second advantage beyond possibly being able to outnumber them in battle: it means battalions that lost morale will have a chance to recover it between battles, rather than constantly getting selected for another battle less than a month after the last one ended.
  
• Don't use Chemical Warfare Specialists unless you absolutely need the extra offense and morale damage that they provide compared to Flamethrowers or Infiltrators. They slow your advance down significantly.
  
• To do successful naval invasions, you want to have some large fleets and you want to make sure that the HQ the fleet is based in also has an army that's got roughly as many battalions as the fleet has flotillas. If the army is much bigger, it'll have bad offense. If the army is too small it might be too weak to get a beachhead. If all your fleets are small, none will be able to transport a significant invasion force.
  
• You should also be prepared to have a lot of trouble with naval invasions until you research Landing Craft, as the offensive penalty is pretty severe. This is less of an issue if you're confident that the defender doesn't have many troops left Standing By or Garrisoned in the target HQ though.

Comments welcome

Please let me know in the comments if you think I got anything wrong. Some of this info is a little speculative as combat is not very well documented.

Good luck, and may the odds be ever in your generals' favour.

Appendix: Detailed Math for How Many Battalions Fight

Feel free to skip this section unless you're interested in the exact numbers for battle size.

First, some variable definitions to simplify writing out the equations (with gratuitous use of Greek letters because they're cooler):


The value of τ depends on whether the army is attacking or defending:


The minimum, average, and maximum number of battalions that a general can bring to a battle is (click to zoom in, Steam won't let me show it at full size):


For the example battle in Hebei earlier in the guide:

To find the expected average number of attacking and defending troops for this battle, I use the "more troops than terrain cap, no stat disadvantage" equation for the attacker and the "stat disadvantage and many more troops than the terrain cap" for the defender, since τκ<μ for the defender:

So if this battle were to happen many times with the same armies and in the same province, on average Canada would attack with 41 battalions and China would defend with 69 (nice).

In chart format

I've left off the equation for the curved part of the charts for the situation where the side in question has a stat disadvantage; the equations are in the table above, and these charts are just to help get an idea of the general pattern.

Attacker Charts

No stat disadvantage

Stat disadvantage

Defender Charts

No stat disadvantage

Stat disadvantage

[nosquimoriturisumus]

@Lil_Perk and @firelord2121, I will update this guide at some point but I'm waiting until the new combat system is stable. Today's dev diary indicated that the dev team is still contemplating nontrivial changes to how things work, and I don't want to wind up needing to update the guide weekly as they patch things.

[firelord2121]

huh

[LiI_Perk]

Please update this once the new system is finalized, it used to be so helpful!

[Hugo]

@captain5555555555 This happens in other paradox games to, but only sometimes. In eu4 the flanking range is usually only 2. So only a few often damaged units could fight the last remaining units. In Imperator Rome this was even more of an issue. In Victoria 3 it seems like it is designed to happen this way.

[captain5555555555]

This has the very serious and bizarre negative consequence of making the combat law follow the *inverse* of lancaster's laws. Outnumbered forces deal disproportionately more damage to their opponent; again, like every battle is Thermopylae. This is why battles take so damn long.

Paradox put out a post saying they want to reduce morale over time in a battle to fix the length. This fundamentally misunderstands what is wrong.
1) If a fixed number of units were able to deal damage at a given time
2) A battalion could receive damage from more than other battalion (say 3)
3) The number of battalions that could damage an enemy could change over time based on technology (perhaps starting at 1 and moving up to unlimited with siege artillery, follow the transition from lanchester's ancient law to lanchester's square law)

Any of these solutions would change minimal balance and make the combat system work better.

[captain5555555555]

A very weird result of their combat system is that the effective width of the battle is min(battalions A, Battalions D) which means the combat width shrinks over course of the battle. This leads to the bizarre result that ~50% of the time of a battle it is 3-1 hopelessly outnumbered battalions fighting a gentleman army that lines up 1-by-1 to engage the enemy. Armies don't do sensible things like encircle, flank, or defeat in detail. They just patiently wait.

So not only is it true that most battalions sit idle while a battle goes on; *most of the battalions in the battle are idle while 1-3 units are engaging.*

This is different from every combat model I have even heard of and is different from how combat is evaluated in every other Paradox game. Victoria 3 basically evaluates every single battle like it is the battle of Thermopylae where 300 Kazakhs fighting in plains hold up 100,000 Russians for 3 months.

[nosquimoriturisumus]

@captain5555555555 that's an interesting result! I hadn't thought to look at the save games, but in hindsight it makes complete sense that the info's there as the game shows the graphs of troop strength when you reload a save.

I'll add some info about battle casualties if the upcoming 1.1 doesn't change them.

[captain5555555555]

I did some work looking into what offence and defense mean but brute force of looking at dozens of battle outcomes. It seems like offence and defense are are exactly equivalently in terms of combat effectiveness and are called "ce" in code, which I assume to mean "combat effectiveness". attacker_start_ce=33.76575 --> offence stat of 33.7

by opening a savegame file in a text editor following these directions: https://www.reddit.com/r/victoria3/comments/yg4s7e/how_to_editdecrypt_victoria_3_save_files/

Information about battles from ongoing wars can be accessed in the save file. The data for the plot is only collected if the player is in the war. By taking the data from some of my save files I was able to reconstruct most of how battles work. Battles between brigades are 1:1. A brigade with 5 men does the exact same damage as a brigade with 1000 men and equal stats.

This screenshot is my graph:
https://gyazo.com/1a9fef3ed59cf689db555beea87ae3a5

[Hugo]

This confirms my experience. High stats correlate negatively with winning. Lower stats is better.

[makapse]

2/2
Another part of the battle mechanic you didnt discuss much about is the battle conditions that are rolled at the start of each battle that can really impact the battle, but even I am not too familiar with how it is selected at random, just that the generals can affect it a bit. Some traits allow the occurrence of certain conditions while others can prevent them. And not all of them are helpful to you, like the mud condition means 50% provinces captured, etc