[Grok / deep retrieval] Literature for location-to-rate: inverse-CDF windows, DKS descendants, F1 connecting lemma

[LightChainr]

Standing

Literature officer. Does not enter docs/STATUS.md. Do not close this ticket. Do not close #276 / #321 / #613 / #566. Primary quotes, PDF-stable. NOT FOUND is a valid answer. Do not prove F1 on square site. Do not fit productions. Do not mix this PR with #602 / #615 / the three computational probes.

The user cannot fetch arXiv.

Full brief (source of truth):

https://github.com/LightChainr/Matching-One/blob/docs/probes-full-law-20260907/notes/probes/probe-LIT-retrieval-quantile-rate-20260907.md

Packet PR: #616

Submit back

Open a separate PR (frontier or a docs/literature-officer-… branch off frontier) with the Q1–Q14 note + bibliographic records. Comment the PR URL here. If you find a square-site RSW or a torus four-arm theorem, one comment on #276 and #321 quoting it — do not close those tickets.

[LightChainr]

Literature packet is on #623 (branch docs/literature-officer-quantile-rate-20260907, off frontier).

• Q1–Q14: notes/literature-officer-quantile-rate-20260907.md
• Biblio: notes/literature-officer-quantile-rate-biblio-20260907.md

Headline for the rate probe: there is no named theorem that the inverse-CDF of {r_G ≥ 1} / {r_G = 2} / E[r_G]/2 on a torus inherits a window of width L^{-1/ν} or N^{-3/4}. DKS “sharp” is λ□(N)±ε of fixed ε (and λ□(N) is the single u=1/2 quantile of {ϕ_* ≠ 0} on the plaquette / square-bond torus). Friedgut–Kalai is 1/log n. The connecting lemma “four-arm ⇒ M' ≍ L^{3/4} on a torus” is NOT FOUND. GPS convert four-arm to expected pivotals in a planar quad on triangular site; the torus/wrapping conversion is not literature. α_4=−5/4 and ν=4/3 are theorems on triangular site (LSW / Smirnov–Werner) and remain open on square site and square bond.

Square-site RSW is FOUND (Zeng Thm 1.1; KST Comment 1). Square-site conformal invariance remains OPEN. These are not the same statement and are not in tension with #606. Pointer comments quoting Zeng are on #276 and #321; those tickets are not closed. This ticket is not closed. docs/STATUS.md is not edited. Not mixed with #602 / #615 / #617 / #618 / #619.

[LightChainr]

Queued under #650, not started — Mac RAM is tight (M4 16GB, three omen already live). ASSIGNED_MACHINE will be commented when a CLI slot opens. Retrieval only, no Huawei, no production.

[LightChainr]

Not re-dispatched. Literature packet already landed and merged as PR #623. Mac RAM is the bottleneck; wrapping-line follow-ups from #668/#669 take the open CLI slots. Ticket stays open.