# PRE-REGISTRATION — does faith change what a turn costs, and what can be read? Written 29 September 2026, before the run. **The Readable Interior + The Ledger**, shares G, L, R, V, a — rank 17, the highest-ranked pipe still to check. ## The two rooms **The Readable Interior** prices a turn: `dV = R·G·[max(0, b·L̂) − max(0, a·L̂)]`, *"the cost of a turn, paid in a currency the system does not set."* It lands on costly signalling (Spence 1973; Zahavi 1975) and cheap talk (Crawford and Sobel 1982). **The Ledger**, as it now stands after his ruling: `V(n+1) = V(n) + R·max(0, a·L̂)·G − (1−R)K − R·b₀·max(0, −a·L̂)·G` The room was written against the Ledger **before** faith was on it. So the obvious question, which nobody has asked: **is the room's price still the Ledger's price?** ## The derivation being tested From the Ledger, the value of holding position a for one cycle is `dV(a) = R·max(0, a·L̂)·G − (1−R)K − R·b₀·max(0, −a·L̂)·G`. The cost of turning from a to b is `dV(b) − dV(a)`, and the leak `(1−R)K` cancels: turn cost = R·G·[max(0, b·L̂) − max(0, a·L̂)] − R·b₀·G·[max(0, −b·L̂) − max(0, −a·L̂)] The first term is exactly what The Readable Interior prints. **The second term is new**, and it vanishes only at b₀ = 0. ## The consequence under test At **b₀ = 0**, a turn between two positions that are **both facing away** costs **nothing** — both clip to zero. Such a turn is invisible on the ledger. At **b₀ > 0** it costs something, so it can be read. **Claim: faith makes readable exactly the turns that were invisible without it — those made in the dark, between two positions both facing away.** ## Criterion, fixed before the run 1. **The room is the b₀ = 0 case.** The room's printed expression equals the Ledger's turn cost exactly (within 10⁻¹²) at b₀ = 0, at 20,000 random pairs, and differs at b₀ = 1 at a majority of them. 2. **The unreadable set at b₀ = 0 is exactly the both-facing-away quadrant.** Every pair with zero cost has both a·L̂ ≤ 0 and b·L̂ ≤ 0, and every such pair has zero cost. No exceptions in 200,000 draws. 3. **The readable fraction is 3/4 at b₀ = 0 and 1 at b₀ = 1.** Drawing a·L̂ and b·L̂ uniformly on [−1, 1], the measured readable fraction is **0.750 ± 0.005** at b₀ = 0 and **above 0.999** at b₀ = 1. **WHAT KILLS IT:** if the room's expression already equals the faith-bearing turn cost at every b₀, the second term is not new and there is no pipe. If the unreadable set is not exactly the both-away quadrant, the reading is wrong and gets reported as wrong. **Prediction, recorded now:** all three hold, with the readable fraction moving from exactly 3/4 to exactly 1. The 3/4 is a guess from the geometry of the quadrant, not a derivation, and is recorded as a guess. **NOT CLAIMED:** anything about which b₀ is right. That is his ruling and this run does not touch it. ## Discipline Verdict assembled from measured quantities. The summary table is **printed by the run, never typed** — entry 31's was typed and mixed neighbouring points before the website session caught it.