Four Players, Four Duties, One Failure Mode

Co-op is the version of this game that most people will actually play, and it is easy to see why. With up to four players, the group can divide the work: one searches buildings, another gathers food, while the others build or defend the shelter. Resource gathering stops being a chore and combat gets room for teamwork. That is the promise, and it holds — as long as somebody actually does the dividing.

The failure mode is not a horde. It is four players independently deciding what needs doing, arriving at four different answers, and each doing a partial version of a different job. At the end of the session the base has four unfinished additions, nobody has run the food pipeline, and the ammunition bench has been touched by three people who each assumed someone else was keeping count.

The duplicated queue problem

Crafting is where duplication gets expensive. When two players work the same bench, you consume materials at twice the rate for the same output, and neither player notices because each one only sees their own half of the ledger. The shortage does not surface until the first real firefight, at which point the squad discovers it has half the ammunition it believed it had.

The fix is ownership rather than effort. A bench should have one owner — not because one person is faster, but because throughput is a counting problem and counting requires a single source of truth. This is the same reason the manual insists ammunition is a base resource rather than a loot resource: once you accept that rounds are manufactured and not found, someone has to be responsible for the manufacturing, and that someone cannot be everybody.

Nobody guards during a build

Building generates noise. That is not a flavour detail — it is the same acoustic model that governs gunfire, applied to hammers and saws. If all four players are inside the compound working, the squad is broadcasting continuously and has nobody positioned to answer what the broadcast attracts.

The usual rationalisation is that the work will be finished faster with four people on it, and then everyone can defend. That reasoning holds right up until the moment it does not, and the moment it does not is when the mist is thick and the group is standing in an unfinished structure with no walls on one side.

A standing rule that at least one role is always outside the wall, facing out, costs a quarter of your construction speed and buys back the ability to notice a problem before it is inside the base. Given how the fog amplifies exactly the noise a build makes, that trade is not close.

The scout's value is information, not kills

Reconnaissance is the role most often played wrong, because it looks like a combat role and is not. A scout who returns with three dead infected has produced nothing the squad could not have produced by staying home. A scout who returns knowing where the hostile camp is, and which approach to it is unwatched, has changed what the squad can do tomorrow.

The reason this is hard to hold onto is that shooting is immediately satisfying and mapping is not. The manual's role table is structured to make the informational value explicit, because the alternative is a squad where everyone gravitates toward the fun job and the map stays dark.

Why the roles are shaped the way they are

Four roles cover four queues: information, food, firepower and infrastructure. Each of them fails differently and none of them substitutes for another.

Information fails slowly — you do not notice a dark map until you walk into something. Food fails on a delay — hunger does not bite the moment the pipeline stops, it bites two days later when the stores run out. Firepower fails instantly and loudly, which is why it gets over-invested in. Infrastructure fails structurally — a base without a maintained perimeter does not degrade gradually, it gets entered once and is no longer a base.

Assigning one role per player is not about efficiency. It is about making each of those four failure modes somebody's specific concern, so that the quiet ones — the dark map and the empty larder — get attention before they become emergencies.

The hosting corollary

There is a hardware version of the same argument. In peer-to-peer hosting, the host's machine simulates the world for everyone, which means the host's specification is a squad-wide ceiling. A host running single-channel memory will stutter for all four players whenever the mist rolls in — precisely when the squad needs reliable input. If anyone in the group has a dual-channel machine, they should be the host, even if someone else has the better connection.

A stable session is not a networking achievement. It is a hardware decision made before anyone connects, in the same way that a stable squad is a planning decision made before anyone starts building.

The caveat

None of this is enforced by the game, and that is deliberate rather than an oversight. The systems are loose enough that a group of four friends can ignore all of it and still have a good session — right up until the session where they cannot. Duty separation is a way of spending a little coordination up front to remove a class of failure that is otherwise invisible until it is expensive.

End of Dispatch 02  //  Filed from Fireteam Operations  //  Revised with the hosting model