Since 1.2, Victoria 3 has not let you march an army into a province. You assign armies to a front, give the commanding general an advance or a defend order, and then watch the arithmetic decide what happens. That one decision is why the classic Paradox move — pile every battalion into a single stack and push — stopped working, and why losing a war you did not want is usually a supply problem rather than a numbers problem.
What this covers: Victoria 3’s post-1.2 warfare model — how fronts are drawn, how battle size is decided, what generals and orders do, and how supply decides who wins. Why it matters: Past a province’s force limit, extra battalions do nothing but feed attrition: the surplus stands on the front without fighting. Who should pick this: Players who returned after 1.2 and lost a war to something that looked like bad luck rather than bad play.
The version this guide describes
The live branch is 1.13.11, released on 18 August 2026 — a hotfix for a crash when sway offers were pending at a month tick, plus Banana Plantation values and one event.
The war layer is being rewritten in parallel. 1.14, the free “AI Update”, entered open beta in early September 2026 and reached open beta update 4 on 29 September. It ships publicly as part of 1.15 on 22 October 2026, alongside the State and Revolution immersion pack.
| Build | Status on 1 October 2026 | What it does to warfare |
|---|---|---|
| 1.13.11 | live | Fronts, orders, generals and supply as described here |
| 1.14 | open beta | War support rework, war goal flipping and expiry, early troopships |
| 1.15 | 22 October 2026 | 1.14’s changes, plus the Cold system and Copper |
One caveat: the wiki’s Warfare overview describes the front system in the shape 1.2 gave it, while its naval sections have been rewritten around 1.13’s ship rework, and the page still flags its own combat-stats section as needing empirical verification.
What a front actually is
A front is not somewhere you move into. It is generated. Once a war — or a diplomatic play that has not resolved — puts you in opposition with another country, fronts appear at each border between allied and enemy territory, and there is one front per pair of contiguous theatres.
An interruption such as a sea or a neutral country does not count if both sides continue around it — the wiki’s example is Russia and the Ottoman Empire around the Black Sea — but if it divides different countries on the same side, you get separate fronts. Fronts are also joined when only one state region separates them, which can put a single front down both sides of the same state.
Three consequences follow.
- War is fought state by state. Each won battle adds occupation score, which takes control of a share of a state. When a state is fully occupied, the front moves to incorporate it — and may then merge, split, or vanish. Armies on a front that vanishes travel to another front or return to an HQ.
- A coalition war is several narrow wars. Each enemy gets their own front, so a broad alliance obliges you to be strong in more places than a one-on-one war would.
- Direction is a setting, not a destination. The battle province is picked from the highest-scoring candidates, scored by distance to the nearest Strategic Objective, or to the nearest war goal when none is set.
Since 1.13.9, the front panel also tracks which side, if any, has made gains over the last few months — the closest thing the interface has to a scoreboard you can act on.
Reaching a front is its own cost. Armies must be mobilised to be stationed at one, which takes five to twenty days depending on state infrastructure, and orders do nothing unless the army is assigned to a front. Demobilising one by hand incurs a 90-day decaying expense and blocks remobilisation for 90 days; being at peace demobilises everything free. An army you cannot remobilise is a front you cannot reinforce for a quarter of a year.
Why you cannot doomstack
A battle draws from the front, then gets capped by the province. Paradox set out the sequence in Dev Diary #77, which shipped with the 1.2 military rework.
| Step | What the command script does |
|---|---|
| 1 | Starts with the total number of battalions on the front |
| 2 | Subtracts a random number of battalions proportional to the front’s length |
| 3 | Caps the result with a force limit set by the battle province’s infrastructure and combat width |
| 4 | Increases the side with more battalions by a random value proportional to its numeric advantage |
Step 3 decides your game. Terrain and infrastructure set how many battalions are in the fight, and everything above that figure is not a reserve in any useful sense — it is on the front, subject to attrition. Base attrition is 10%, with weekly casualties ranging from 4% to 12%, and only formations stationed at their home HQ get the −50% attrition reduction. The one exception is that battalions actively engaged in a battle temporarily stop suffering attrition.
This is also the oldest claim about the design. Before release, Paradox’s position as reported at the time was that winning a war “isn’t simply a matter of ‘doomstacking’ troops”: generals confer bonuses on the battalions they command, and an army can be concentrated along a short front or spread along a large one. The mechanic now matches that framing.
The same diary added a second restraint: battalions in a battle are never reinforced by their barracks hiring new soldiers, so new manpower joins only after the unit disengages and battles end. It also made province selection steerable: battles closer to a Strategic Objective, or the nearest war goal, score higher, and an objective persists until the war ends or the targeted state is conquered.
Generals, command limits and the political bill
An army is a formation with a home HQ and room for up to four commanders. New units and commanders are recruited through formations, and a formation’s command limit is the sum of what its commanders’ ranks allow.
| Rank | Command | Interest-group political strength | Bureaucracy |
|---|---|---|---|
| Brigadier General | +30 battalions | +2% | −10 |
| Major General | +60 battalions | +5% | −15 |
| Lieutenant General | +80 battalions | +10% | −20 |
| General | +125 battalions | +15% | −25 |
| Field Marshal | +200 battalions | +20% | −30 |
| Commander-in-Chief | +50 battalions | +25% | −10 |
Two rules matter more than the numbers. First, promotion is political: the interest group a commander supports gains political strength as their rank rises, and male rulers of monarchies or dictatorships can be made generals at the Commander-in-Chief rank and have it revoked freely while not mobilised. Second, exceeding the command limit is not a soft failure. Maximum organisation drops when a formation lacks command limit, carries more than 50% special units, or suffers a supply penalty — and organisation multiplies everything: at 0% a formation loses 75% offence, 75% defence, all morale recovery, and half its recovery rate. It recovers on the order of 1% a day, so an over-limit formation is a months-long repair rather than a bad week. That political half is the same machinery that governs law reform, which is why the interest groups guide describes the same characters from another angle.
1.13.9 changed how the limit is counted: regulars or raised conscripts, whichever is higher, is compared against the command limit instead of the two being added together, which the patch note says “removes the need to mass-promote commanders when you raise conscripts”. An army with no generals at all can still defend a front, though only up to 10 battalions before gaining penalties.
Where you park an army is itself a mechanic. Formations at their home HQ take half attrition, and since 1.10.3 armies stationed at an HQ also defend fronts inside that HQ’s strategic region, counted in the same front-advantage calculation as armies sent forward. Generals stationed in an HQ as a garrison join defensive invasion battles rather than merely sending their troops, and armies at a region HQ do participate in the defence of a naval invasion — with one caveat the wiki states plainly: when several invasions land at once, a single stationed army may commit all its battalions to one battle and leave the others uncovered.
Orders: advance, defend, and nothing else
Advance orders push progress toward the next battle: while no battle is running, progress accumulates daily, and at 100 the army engages whatever opposes it or advances occupation if nothing is there. If both sides advance, the usual result is alternating attacks. Defend orders prevent enemy advances; a defending general can recapture friendly territory but can never occupy enemy territory.
| Order | Effects | Gate |
|---|---|---|
| Advance Front | +5 advancement speed, +10% offence | — |
| Reckless Advance | +7 speed, +15% offence, +10% morale loss, +150% blunder chance | general is Reckless |
| Cautious Advance | +3 speed, −5% morale loss, +50% careful manoeuvre chance | general is Cautious |
| Pillage | +20% morale damage, +10% kill rate, +100% devastation, −20% occupation gain | general is Cruel, Wrathful or Pillager |
| Defend Front | +10% defence | — |
| Adamant Defense | +15% defence, +20% supply consumption, +100% dug-in chance | defensive strategist traits |
| Last Stand | +20% defence, +30% casualties taken | Stalwart Defender, Trench Rat or Defense-in-Depth Specialist |
| Delaying Tactics | +10% defence, +10% morale loss, +200% camouflaged chance | Pillager, Cruel, Wrathful, Bandit or Social Bandit |
Read the defensive column and the pattern is uncomfortable: the strongest defensive multipliers are locked behind specific traits, and one of them belongs to bandits and pillagers rather than to a defensive specialist. Defend Front, the only ungated option, is worth half of Adamant Defense. A defensive war is therefore a recruitment problem before it is a deployment problem — hire the right general before the war, not during it. The remaining advance variants gate on composition instead, rewarding armies built around one artillery type or 30% of a single cavalry type.
The wiki lists the naval equivalent: interception, convoy raiding, convoy escort and blockade. One entry there matters directly to land defence: naval invasions carry a −25% attack penalty for every invading battalion that is not of Marine type. If your plan is to let an invasion land and counter-attack, that penalty works for you. Land invasions are the other route in; they require military access through every neutral country along the way and cannot be started at all if the army would need more than 30 days to reach the destination.
Supply, attrition and the arithmetic of waiting
Everything above describes who fights. Supply decides whether they can.
Formations fighting overseas depend on the supply network. If enemy fleets raid it, the effect is direct and linear: the wiki’s example is that a 2% reduction in supply network efficiency leaves morale recovery of units overseas at 98%. If supply falls below 95%, an equal penalty applies to organisation — a unit at 80% supply carries a 20% penalty to maximum organisation. Shortages in the buildings that produce your units reduce a formation’s supply the same way. Those buildings sit on the same input chains as everything else, so a supply collapse is often an economic collapse in uniform; the loops in the construction and GDP guide apply to barracks too.
That is the defensive case in two numbers. Convoy raiding does not need to sink your army; it needs to knock morale recovery and maximum organisation down far enough that defensive multipliers stop mattering. Escorting convoys and protecting supply lines are front-line decisions, not chores to delegate.
Attrition sets the clock. Base attrition is 10%, attrition casualties run from 4% to 12% weekly, and the wiki’s own illustration of an early Anglo-Spanish war shows losses to attrition far outstripping losses to battle. Add mobilisation at five to twenty days, a 90-day lockout on any army demobilised by hand, and conscription scaling with Army Model laws and state population — unincorporated states carry −50% — and the rule is blunt: the army you can keep in the field for a year is far smaller than the one you can raise for a month.
A defence-in-depth playbook for a war you did not want
1. Decide which front you are actually fighting. A coalition gives each enemy their own front. You need one front where you are strong and enough elsewhere that nothing moves.
2. Keep the reserve at home HQ until you need it. Half attrition, and since 1.10.3 an HQ army still defends fronts in its own strategic region. If you cannot hold a front, thinning it is not a mistake — feeding it is.
3. Put a defensive order on a general with the traits for it, and check the order panel rather than assuming. Orders have no effect unless the army is assigned to a front, and only the advancing side generates progress toward the next battle: a stray advance order converts your defence into their attack.
4. Do not reinforce a bad-terrain front. The force limit belongs to the province: infrastructure and combat width cap how many battalions engage, and the surplus takes attrition in place. One mountainous, low-infrastructure province buys you nothing with extra battalions.
5. Attack the supply network instead of the army. A 2% raiding cut is a 2% cut to morale recovery; past 95% it also cuts maximum organisation. If you can win only one thing, win the sea node next to their overseas front.
6. Play the war support clock; the clock ends wars. Every participant starts at +100 and loses war support weekly to exhaustion. A country cannot go below 0 unless an enemy occupies all its war goals or its capital, and it capitulates automatically at −100. Subjects cannot capitulate, and a negotiated peace needs every country with or targeted by a war goal to agree.
| War exhaustion driver | Weekly gain |
|---|---|
| Base | +0.2 |
| Enemy controls all war goals (scaled by percentage) | +1 |
| Per 1% of population that are radicals | +0.015 |
| Per 10% casualties, scaling, with all battles lost | +0.4 |
| Occupation above 50% | +2.0 |
| Occupation above 90% | +10 |
7. Hold territory, not battles. Since 1.13.9, a state at least 50% occupied counts as occupied for war goal control rather than requiring full occupation, and a war goal is generally contested if you control the enemy capital or 50% of their incorporated states. The 1.2 rework had already reweighted exhaustion so casualties matter more than a small land grab, so holding your capital state and a slice of your opponent’s incorporated states beats a battle win you cannot exploit.
8. Know when the rules change underneath you. On 22 October the war support rework arrives with 1.15. Open beta update 4 introduced war goal expiry — in wars lasting over two years, war goals uncontested for 12 months are dropped — and AI countries now capitulate when clearly defeated, so wars should no longer drag on at 0% war support. Open beta update 2 had already rescaled war goal infamy against a state’s share of global GDP and population rather than a flat value, and early troopships are now available from the start. A plan built on outlasting the AI is being weakened; a plan built on holding war goals is not. What else ships alongside those changes sits in the Volume 3 roadmap.
One last piece of weather: the 1.15 Cold system slows ships in icy sea nodes, and land units supplied through an affected sea node need 50% more supply ships. A defensive plan that assumes sea supply through a northern winter has a season.
What this means for the current build
The front system is not a movement system with a different interface. It is a contest over how much of your army is allowed to matter at once, and the levers are terrain, command limits, organisation and supply. On live 1.13.11 the strongest defensive position is usually a home-HQ army under a defensive specialist, a front you have deliberately chosen to be weak on, and a raid on the other side’s convoys. When 1.15 lands on 22 October, the layer underneath — war support, war goal control, infamy — moves with it, so re-test any playbook that leans on the AI’s reluctance to capitulate.