Independent Playtest — Muster v2

Date: 2026-09-15
Tester: Jeroen (project owner). Did not write the v2 rules, did not read
the solver findings or the author's REPORT before playing.
Ledger: kept by the author. The tester made every allocation decision; the
author computed state and never suggested a purchase.
Hidden values: rolled with a true-random seed and SHA-256 committed before
the first decision (970801a72a58949ef7c44bbb18534dad). Neither party knew them.
Stopped: mid-Season 2 (turn 6 of 8), at the tester's call.

This closes the condition left open by REPORT.md — that run was played by the rules author and could not count as independent evidence.

What happened

Season 1 — six settlers dead on the earliest possible turn

The tester opened on ploughs (T1), switched to spears (T2), mixed (T3), then committed hard to spears on T4.

The raid landed on turn 4 — the earliest the rules allow. Multiplier 0.95, a mild raid, strength 6.

Defenders: 0. Eight spears were paid for and sitting in the forge. Not one was ready. Six of twelve settlers died.

Everything shrank at once: income halved (6 → 3 iron/turn), the spear cap fell from 10 to 4, and **four of the eight paid-for spears became permanently stranded** — iron spent arming people who no longer existed.

The season was survived. Food was met (10 ploughs for 6 settlers), 2 carried over.

Season 2 — held, but by the dice

Opening position: 8 ready spears against a cap of 4. **The tester could not order a single spear for the entire season** — over cap from turn one.

The raid landed T5. Multiplier 0.75 — snake eyes, the weakest roll in the distribution. Expected 5.9 became actual 4, exactly matching the 4 defenders.

Zero losses. The tester was braced for two. That was variance, not planning.

Hidden values, revealed after the stop

The finding that ends this prototype

At Season 2, Turn 2, after a second consecutive turn where banking iron was the only legal move, unprompted:

"its kinda frustrating. i want more diversity in weapons and tools."

State at that moment: pop 6, spear cap 4, 8 spears already committed (4 stranded forever), the season's food requirement already met. Iron kept arriving at 3/turn with nowhere to go.

This is the economy-designer's predicted failure, reached in play rather than on paper. Their review of the concept said it directly: *"Spears never degrade, so once the armoury is full, military spending stops forever and the late game flatlines by construction."* Here the armoury filled on turn 5 of season 1 — not because the player over-invested, but because **the population collapsed and took the cap down with it.**

The single-axis allocation (iron → spears or ploughs) has a dead zone, and losing settlers drives you straight into it. Once food is met and spears are capped, the central decision of the game does not exist.

A second signal, before play began

On reading the rules:

"is this a turn based game caus i dont like that at all"

The game is real-time; turns are paper scaffolding only. Recorded because the abstraction cost goodwill with the target audience even as scaffolding — worth knowing if a future prototype is handed to an external tester.

Verdict: the mechanic works; the economy around it is too thin

What the playtest confirms:

• Lead time is the real thing. Season 1 is unambiguous: eight spears paid for, zero available, six dead. The regret was specific, immediate, and traceable to a turn. No prior version of this design produced that.
• The scaling threat estimate is load-bearing. The tester tracked it and allocated against it.
• Raid timing uncertainty bites. Turn 4 versus turn 8 was the whole game in season 1, and nothing in the player's control changed it.

What it refutes:

• The allocation does not stay live across a season. It went dead for the entire back half of season 1 and effectively all of season 2. Both of the hypothesis's failure states were reachable, but with one gear type and a population-linked cap, the decision disappears whenever the settlement shrinks.
• Criterion (c) from the concept doc's hypothesis — *the player articulates what the allocation cost them in both directions* — was not met. The tester articulated frustration at having no choice at all, which is the opposite signal.

Conclusion: PIVOT — narrow, not fundamental. The muster mechanic itself is validated. The surrounding economy needs more sinks before the allocation can stay live, and the tester named the fix without being led to it: more gear types.

What the next iteration needs

• More than one military sink. Mail, shields, better tools — each a separate thing iron can become, each with its own cap or none. The concept doc defers gear tiers past MVP; this playtest says at least one more axis is required for the core loop to function, not for polish.
• Decouple the spear cap from population, or let surplus gear do something. Stranding four spears when settlers die is thematically right but mechanically dead. Options: surplus gear as replacement stock for future raids, or scrap value returning iron.
• Give banked iron a cost or a use. Two consecutive turns of "bank it" as the only legal move is where the frustration surfaced.
• Re-test with a real-time framing if possible. The tester disliked turns before play began. The rules translate (lead time is just a smith visibly working), and a Unity spike would test feel as well as arithmetic.

Method notes

Four method improvements over the author's own run held up: hidden values committed with a published hash, a rule-enforcing ledger, the adversarial solver pass run before write-up, and an independent decision-maker.

One ledger bug surfaced during play: the spear-cap assertion blocked a plough-only purchase, because in-flight spears exceeded a cap that had shrunk when six settlers died. Ploughs never touch the spear cap. Fixed live; no design conclusion depended on it.