PIVOT NOTE — The Muster

Written 2026-09-14 after the original PROCEED verdict was withdrawn.
Read this before building the next muster prototype.

Original hypothesis

If the player chooses, weeks before any threat, whether iron becomes ploughs or spearheads — and that allocation is the only cost of readiness, with the muster itself free — then when the horn sounds the player feels the weight of the earlier choice.

Why it pivoted

The prototype could not have falsified its own hypothesis.

• Dominant strategy. 15 spears (one per settler) beat any 2d6 raid, and 15 iron arrived by turn 5 against an earliest raid on turn 7. Optimal play was "spears to 15, then ploughs" — a step function with a computable knee. The felt tension came from the tester not having done the arithmetic.
• Unfalsifiable plough. Harvest Score had no consumption, threshold, or failure state. Regret was mechanically impossible, so "were ploughs a trap?" could only ever answer no.

What to KEEP

• The displaced-cost premise itself. Nothing disproved it. Pay at the forge, muster free — still the most promising version of the hook.
• The turn-7 look-back as the target beat. "I can see exactly where I under-invested" is the reaction to design toward. It was real, even if the decision behind it was solvable.
• Secret raid timing. Uncertainty about when generated more pressure than magnitude did. Now recorded as Core Mechanic #6 in the concept doc.
• The free muster. No objection from the tester, and it keeps the Manor Lords coupling trap at bay. The weakest-supported keeper, but nothing argues against it.
• Paper as the path. Right call. The defect was in the rules, not the medium, and paper is where it should have been caught.

What to CHANGE

• Remove the guaranteed-win threshold. Raid strength must be able to exceed any reachable defence, or must scale with settlement progress, so "enough spears" is never a fixed number the player can memorise. Without this there is no decision, only an execution problem.
• Give ploughs teeth. Harvest Score must feed something with a failure state — food stores, population growth, next-season capacity. If an under-fed settlement loses settlers, a plough skipped is a cost the player can feel, and "was it a trap?" becomes answerable in both directions.
• Add a recurring military sink. Spears that never degrade mean military spending stops forever once the pool is filled. Gear decay (e.g. a spear lasts 2–3 raids) keeps the allocation live across seasons instead of front-loading it.
• Resolve the dead-settler gear question. Does a fallen defender's spear return to the armoury? Recovered undermines displaced cost; lost compounds raid severity. Pick one deliberately.
• Fix the inert opening. Threat must be readable from turn 2–3, not the midpoint. (Carried from the original report; still valid.)

Revised hypothesis for the next prototype

If military readiness has no computable safe threshold (raid strength scales and
gear decays) and the economic alternative has a real failure state (ploughs feed
a population that can starve), then the iron allocation stays a live decision
across a whole season — and the player can articulate, unprompted, what their
allocation cost them in both directions.

Evidence required this time: at least one tester who did not write the rules, plus an explicit pre-session adversarial pass asking "what is the optimal play, and does it make the decision disappear?" If that question has a clean answer, the prototype is not ready to run.

Process lesson

Compute the dominant strategy before running any balance prototype. It took four lines of arithmetic and would have caught this before a session was spent. A verdict written by the same person who wrote the rules, in the same sitting, with no adversarial pass, is not evidence.