Phantom Brigade

Most tactics games ask you to think about space. Phantom Brigade asks you to think about time, and that single reframing is why it deserves attention over a dozen better-funded competitors.

Developed by Brace Yourself Games — the Vancouver studio best known for the rhythm-roguelike hybrid Crypt of the NecroDancer, which should tell you something about their appetite for structural experiments — Phantom Brigade casts the player as the commander of the last intact squad of a defeated resistance. The occupied nation of Fyrland is under the boot of the Kaskad regime. The squad has a handful of scavenged mechs, a transport, and a map full of enemy patrols that outnumber and outgun them. It is, structurally, an insurgency simulator dressed in 40-ton armor.

The market position is unusual, and slightly awkward. Phantom Brigade sits in the gap between three well-established things: the limb-targeting mech tactics of Front Mission, the campaign-and-salvage economy of BattleTech, and the moment-to-moment cinematic weight of MechWarrior. It borrows the vocabulary of all three, then throws out the one thing they share — the discrete turn. Comparisons to XCOM show up constantly in press coverage and player reviews, and they’re the least useful ones. XCOM is a probability game. Phantom Brigade is a prediction game. The difference matters enormously in practice.

The reception reflects a game that is loved for one thing and forgiven for several others. The core combat converts nearly everyone who engages with it, while the layers surrounding it lose a meaningful slice of players by the mid-campaign. The 2.0 update — a broad overhaul of progression, factions, map generation and quality-of-life — was the studio’s answer to exactly that critique, and it moved the needle. It didn’t close the gap entirely.

• Timeline-Based Command System: Instead of taking turns, you scrub through a short predicted future — a few seconds of projected enemy movement and fire — then draw your own movement paths, dodges, melee lunges and weapon shots along the same timeline. Both sides then execute simultaneously in real time. It is choreography, not chess.
• Physics-Driven Destructible Combat: Buildings actually collapse. Limbs and weapons are shot off individual units and stay off. A mech that loses its right arm loses the gun attached to it, and the wreck you leave behind is salvage you can strip.
• Deep Mech Fabrication: Chassis, limbs, internal components, paint and decals, plus recovered exotica — railguns, beam weapons, energy swords. Loadouts reward building around a concept rather than climbing a stat-stick ladder.
• Pilot Traits and Squad Synergy: Pilots earn experience and draw from over a hundred traits, which is enough breadth to build a squad around a doctrine — brawlers, snipers, mobility harassment — instead of just leveling everyone identically.
• Procedurally Shifting Overworld Campaign: A province-by-province liberation war with enemy patrols to evade, repair and refit downtime, and a front line that moves whether or not you’re ready for it.

The Four-Second Chess Match

Here’s what actually happens in a Phantom Brigade battle, and why the description undersells it.

The game shows a five-second slice of the immediate future: ghost outlines of enemy mechs walking, turning, raising weapons, and the projected paths of the shots they’re about to take. You can scrub this timeline like a video editor’s scratch track — forward, back, frame by frame. Then you author your own five seconds. You drag a movement spline across the ground. You place a dodge at second 2.4 because that’s when the enemy railgun fires. You queue a shotgun blast at 3.1, timed to land the instant your target’s cover finishes collapsing from a teammate’s missile at 2.9.

Then you hit execute, and everything resolves at once, in real time, with camera work good enough to reward replays.

The design achievement here is that the enemy prediction is honest but incomplete. You see what they intend to do based on current conditions. Change the conditions — flatten the building they’re behind, sever the arm holding the weapon, move somewhere they weren’t planning for — and their plan degrades in ways the projection couldn’t have shown you. The result is a tactical layer with something almost no turn-based game achieves: genuine anticipation. You are not calculating hit percentages. You are reading intent and then invalidating it.

This is a strategy game about latency and commitment. Every order is a bet placed four seconds before it resolves, and the skill ceiling lives entirely in how precisely you can sequence actions inside that window. When it clicks — when a mech threads through a collapsing street to shear the legs off a heavy at exactly the moment its escort is mid-reload — nothing else in the genre feels like it.

The Campaign Layer Is the Weak Limb

And then you return to the overworld, and the temperature drops.

The campaign is the part of Phantom Brigade that has always been outrun by its own combat, and it’s the source of nearly every credible criticism in the player base. Mission variety is thin. You attack bases, defend positions, ambush convoys and hunt patrols, and the deltas between these are smaller than the four-second sandbox deserves. The tactical engine can generate a hundred unforgettable situations; the mission designer keeps asking for the same six.

The 2.0 update did real work here — reworked progression pacing, better faction structure, improved map generation, and a long list of quality-of-life fixes that returning players noticed and appreciated. The province-capture loop has more shape than it did at launch, and the escalating regime response gives the map a pulse it previously lacked. But the fundamental rhythm is unchanged: extraordinary combat, adequate connective tissue. By the late campaign, the strategic layer reads less like a war and more like a mission dispenser with an inventory screen attached.

Difficulty balance also swings hard. Early attrition can spiral — lose two mechs in an ambush you couldn’t have scouted, and the repair economy punishes you for hours. Later, once the fabrication chain matures and pilots have stacked the right traits, encounters that should be climactic get dismantled in one timeline. The 2.0 progression rework smoothed this, not solved it.

Onboarding Friction

Phantom Brigade is bad at explaining Phantom Brigade. This is the single most fixable flaw it still carries.

The timeline system is intuitive to watch and unintuitive to operate. Concepts that turn out to be load-bearing — how action queuing interacts with movement momentum, why a shot whiffs when the reticle looked green, how dodge timing actually resolves against projected fire, what the difference is between concurrent and sequential orders on one unit — are underexplained or left to be inferred. New players routinely spend their opening hours losing fights for reasons the UI never surfaces. Players who push through consistently describe the moment it clicks as a revelation, which is exactly the problem: a revelation gated behind an information deficit is a design failure, not a rite of passage.

Pathfinding compounds it. Mechs occasionally take routes you did not draw and would not have chosen, and because the entire plan is timing-dependent, a two-second navigation hiccup doesn’t cost you a step — it costs you the whole sequence. It’s not constant. It’s frequent enough to matter.

Customization and the Salvage Loop

The fabrication side is quietly one of the best-tuned systems here. Because damage is localized and persistent, salvage is specific: you don’t loot a generic parts pile, you strip the arm you shot off, the chassis you cracked, the railgun whose owner you legged. That creates a direct causal line from a tactical decision to an economic outcome, which is the kind of loop most strategy games only gesture at. Deliberately disabling rather than destroying a target — shooting off the weapon arm to keep the torso intact — is a real, rewarded tactic. That’s excellent design.

Trait-driven pilot builds pair with it well. With a hundred-plus traits in the pool, squads develop identities, and the good loadouts are the ones built around a pilot’s quirks rather than around a spreadsheet optimum.

The art direction is the game’s second-best asset. Fyrland is rendered in a muted, overcast palette — wet concrete, industrial sprawl, low grey skies — that makes the mechs look heavy rather than heroic. This is not anime mecha. It’s cold-war hardware with mud on it, and the restraint is deliberate and effective. Silhouettes read clearly at combat distance, which matters more than it sounds when you’re parsing six units and their projected paths simultaneously.

The destruction sells everything. Buildings don’t play a collapse animation; they come apart, and the debris changes the fight. Combined with the replay-quality camera work, execution phases carry a cinematic weight no turn-based contemporary matches.

The UI is where form and function trade punches. The timeline scrubber itself is a smart piece of interface engineering — legible, responsive, and hard to imagine designed better. Everything else is denser than it needs to be, with critical information distributed across panels that assume a fluency the game never teaches. The mech fabrication screens are functional but cramped.

Performance is the recurring technical complaint, and it’s credible on paper. Physics simulation plus full destructibility plus multi-unit real-time resolution is an expensive combination, and busy urban maps are built to strain mid-range hardware. It’s not a broken game — it’s a game whose ambition occasionally exceeds its optimization budget.

Steam, PC, mouse and keyboard. There is no ambiguity here and no compromise to discuss, because this game could not exist on a controller without being redesigned.

The core interaction is drawing splines and placing timed events on a scrub bar. That is a precision pointing task with continuous, high-resolution input requirements — analog-stick approximation would introduce exactly the imprecision the whole design is built to eliminate. Mouse and keyboard isn’t the preferred input; it’s the only honest one.

That makes Phantom Brigade an emphatically desk-bound game in a market drifting toward Steam Deck verification. Deck-class hardware faces the double problem of input precision and the physics/destruction cost on busy maps. Treat this as a game for a real machine at a real desk with an hour to spare per session.

The Steam ecosystem itself works in the game’s favor. A long Early Access history and the 2.0 overhaul mean the current build is meaningfully better than the one many older reviews describe — which is worth remembering when reading the sentiment split. Sort by recent, and read what returning players say about 2.0 specifically.

Phantom Brigade is a game with one extraordinary idea and the courage to build everything else around it. That idea — collapsing the turn into a shared four-second window you choreograph against a legible but breakable prediction — is strong enough to carry a mediocre game. It doesn’t have to. What surrounds it is uneven rather than bad: gorgeous, tactile, mechanically deep in the fabrication and salvage layers, and undercooked in mission design and onboarding.

The 2.0 update matters. It’s the difference between a fascinating tech demo with a campaign stapled on and an actual game you’ll finish. Brace Yourself Games went back into a shipped title and rebuilt its progression, factions and maps, and the improved standing that followed is earned rather than inherited.

We rate it a solid recommendation for the right audience — tactics players who will forgive the rough edges for the combat. Buy it for that. You will forgive the rest, occasionally through gritted teeth, because there is nothing else that does this. Four seconds at a time, Phantom Brigade is the most interesting tactics game on Steam.