Aging & Senescence · §13

Reproducibility, grading, and the ledger

Every quantity is reproduced deterministically (SEED 19; the result hash is byte-identical across two runs at 62d5e1eb93db…) and graded honestly: mechanisms and the cross-species null result are [V], the copy-number longevity switch is [L], and absolute rates, incidence and lifespans are [O] with stated obstacles.

The discipline is LOCK → Derive → Gate with no tuning: every constant is a measured input or a derived value, never chosen to hit a target. The deterministic engine emits a byte-identical result on repeated runs (2×sha256), and the irreproducibility ledger collects every [O] item with its specific obstacle.

LOCK → Derive → Gate

Inputs are locked, results are derived by fixed rules, and gates check the derivation by count. The vendored substrate math is never re-derived here — the R19 field and its derived stability and dwell laws are defined once in §1; the promoter γ values are measurements; the dynamics are deterministic simulations seeded at 19.

SEED = 19  |  result sha256 = 62d5e1eb93db89630c4cd288… (2× identical)

The grade vocabulary

Three grades carry the evidence load. [V] is verified — measured or reproduced in-package. [L] is anchored to a cited result. [O] is open — not reproducible in-package, with the obstacle stated. The grading is the framework's load-bearing claim, not decoration.

No tuning

No constant is chosen to hit a target. Noise scales and chronic-stressor magnitudes used in the pathology sweeps are stated, round, and set the relative shape only; the absolute magnitudes they would imply are graded [O]. The irreproducibility ledger lists every [O] item, its obstacle, and its location.