Combat

Ship combat, which is a systems problem; and on-foot combat, which is deliberately scarce.

The design's success metric is not "is fighting fun". It is "is being hunted terrifying". Combat's primary job in this game is to put unbanked survey data at risk, which multiplies the tension of every other system. A player who escapes has had a good encounter.

What combat is for

There is no bounty economy, no loot, and no money, so combat cannot be a livelihood. It has exactly three jobs:

Escape is a legitimate and usually correct outcome. The design should make running away skillful and satisfying rather than a failure state — because a player carrying four hours of data should run, and a design that punishes them for it is a design that punishes correct play.

Ship combat

⬜ Designed, not built.

The shape of an engagement

   [ contact ]──→[ assess ]──→┬──→[ ESCAPE ]───→ mass-lock break → charge → jump
                              │      the usual, and the skillful, answer
                              │
                              └──→[ ENGAGE ]──→[ disable ]──→[ disengage ]
                                     pips · heat · subsystems

Everything mechanically distinctive happens in the two named systems that already exist for other reasons: power and heat. Combat adds almost no new machinery, which is exactly why a small team can make it good.

Escape — the primary skill

Four things stand between a ship and a jump out, and mastering them is the combat skill that matters most.

The heat gamble is the good one. Silent running drops your signature to near nothing but your heat rises with nowhere to go, and you have perhaps forty seconds before modules start failing. Running silent to break a lock, then venting at the last possible moment, is a genuine skill with a genuine cost and it uses no new systems at all.

Combat under burn

The situation the travel model makes possible, and the one most worth building well: an engagement while both ships are under acceleration.

The best encounters in the design are the ones where a pirate matches your burn and you have to fight, run and decelerate at the same time, with one drive and one thermal budget between the three.

Resolved: M6 opponents engage at rest or on maneuver thrust only. Matched-burn engagements need AI that can plot, hold and adjust a transit solution under fire, and that is the likeliest thing in M6 to overrun. The player can still be intercepted mid-burn — they simply have to decide whether to abort, which is already a good decision.

Weapons

Four classes. Deliberately few — depth comes from subsystem targeting and heat management, not from a catalog.

Travel time is the fidelity that matters. Kinetic rounds cross 3 km in about 1.2 seconds, which means leading a maneuvering target at range is a real skill and a fast, erratic ship is genuinely hard to hit — with no accuracy stat anywhere in the system. The physics is already there; the weapon just has to be an entity.

Long TTK is a deliberate choice. A 60-second engagement is long enough for pips, heat, subsystem targeting and the decision to run to all matter. A 10-second engagement is a reflex test, and reflex tests are where a browser game's input latency and a solo team's netcode lose to the competition.

Defensive systems

No consumables anywhere — every defensive system is a module on a cooldown. See flight for why: one resource, no ammunition economy.

Subsystem targeting — why combat is about disabling

With a lock inside 8 km, individual modules on a target can be selected and damaged. This is the design's answer to "how is combat interesting with four weapon types":

Destroying a ship yields nothing. Disabling one ends the encounter. The design should make disabling the obvious, satisfying resolution and destruction the wasteful one, because that is the behavior a non-commercial persistent universe wants.

Opponents

Honest scoping: a solo pipeline cannot produce great AI, so the design targets readable AI rather than clever AI. Every opponent's intent must be legible from its behavior within three seconds.

The Hunter is the only one that should be genuinely frightening, and it should be rare enough that encountering one is a story. Resolved: a Hunter can follow one jump and only one, and the player sees a matching charge signature before committing. Being followed once is a story; being followed indefinitely is a punishment, and it would break the principle that escape is always viable.

On-foot combat

⬜ Designed, not built.

Scarce, lethal, and slow. See onfoot for the scoping decision and its reasoning.

The rules

The suit is the health bar, and it is the same suit that keeps you alive outside. A firefight that costs 30% integrity on a vacuum world is not a setback, it is an emergency — you now have a compromised suit and a walk back to the ship. That coupling is what makes scarce combat carry weight without needing gunfeel to compete with a dedicated shooter.

What is not built

No on-foot PvP in the MVP. No infantry AI beyond patrol, alert and engage. No melee. No cover system in the mechanical sense — cover is geometry, and the player's use of it is not a mechanic the game has to implement.

Resolved: environmental threats only, through M5. This removes humanoid rigging, animation and infantry AI from M5 entirely — the largest single scope saving available — and loses very little, because the environment is the more interesting antagonist. Hostile humanoids, if they arrive at all, arrive with M6.

Related

• ships — pips and heat, which are most of this page's depth
• flight — flight assist off, which is where combat flying lives
• onfoot — the suit, which is the on-foot health model
• modes — PvP consent and where combat can happen at all