R1's recorded prior: a sealed bet on a run never executed

A sealed prediction held open for scoring. R1 (the carpet fractal-quantum-Hall iDMRG computation) was never run, but a prior was committed for it — a bet placed against the author's own standing claim about the carpet state ([[CASE-01]], the fractal-FQH thread). Recording it so the prediction survives and can be scored if R1 is ever run, per predict-then-score.

What is held — the exact recorded prior (verbatim, log L19253):

> "R1 is the expensive one and the only one whose outcome I'd bet against my own earlier claim."

In its surrounding context (the recipes paragraph): "Recipes. Seven scripts plus a README with predictions and runtimes for BernoullisBook. R6 is the one to run first: minutes, and it settles whether 'closures don't commute at p' has a rate, with the answer very likely being yes and the rate being the inverse Liouville speed. R1 is the expensive one and the only one whose outcome I'd bet against my own earlier claim. …"

The scored quantity is this: the author committed, before any run, to bet against their own standing carpet claim ([[CASE-01]], the fractal-FQH thread) on R1's outcome. The verbatim sentence above is what gets scored later — a paraphrase would not. This probe holds it open and flags that R1 has no lab/runs/ entry: the bet is placed, the run is not executed.

To close. Run R1 (or its filed successor), then score the recorded prior against the outcome — a committed-in-advance bet that lost or held is worth more than an unrun claim.