Day 01 · Phase 1 — Foundations & Audit

Site Audit& Epistemic Label System

A complete inventory of every substantive claim currently published on this site, together with a four-label discipline that will govern all future pages.

Purpose of Day 01

Before any new computational or theoretical work is published, the existing content must be audited. Every assertion is classified under exactly one of four epistemic labels. This discipline is the foundation of the 90-day program: readers must be able to distinguish at a glance what has been proved, what has been computed, what is taken from the literature, and what remains heuristic or open.

1. The Four Epistemic Labels

From this day forward every substantive claim on the site carries exactly one of the following labels. The labels are mutually exclusive and exhaustive for the purposes of this project.

Kanıtlandı · Proven

Mathematical proof

A statement for which a complete mathematical argument exists. Example: the cycle inequality \(0 < m\ln 2 - k\ln 3 \le k\ln(1+1/(3B))\) is a theorem of elementary analysis.

Hesaplandı · Computed

Machine-checked computation

A finite computational verification that has been executed and is reproducible. Example: “every \(n<2^{36}\) eventually descends” is a computational fact, not a theorem about all positive integers.

Bilinen literatür · Literature

Attributed prior result

A result taken from the published literature and correctly attributed. Example: Tao’s 2019 almost-all theorem; Barina’s verification records; the classical Steiner cycle equation.

Varsayım / Sezgi · Heuristic

Heuristic, open, or informal

Probabilistic arguments, geometric-mean heuristics, informal intuition, or statements that remain open. Example: the claim that the average multiplicative factor is \(\sqrt{3}/2\approx 0.866\).

Usage rule. A single sentence may contain only one label. If a paragraph mixes a theorem with a computation, the two claims are separated and each is labelled independently.

2. Inventory of Current Claims

Every substantive assertion that appears on the public pages as of 2026-10-10 is listed below and assigned exactly one label.

Claim (paraphrased)PageLabelNotes
The map \(T\) is the accelerated Collatz map.index, allKanıtlandıDefinition.
Every integer \(2\le n<2^{36}\) eventually falls below its starting value under \(T\).index, verificationHesaplandıFinite computation (157 s). Not a theorem for all \(n\).
The Collatz conjecture holds for all positive integers less than \(2^{36}\).verificationHesaplandıFollows by induction from the computation; still computational.
Any nontrivial cycle with elements \(>2^{36}\) has ≥190 537 odd terms (length ≥492 531).index, cyclesKanıtlandıConditional on verification bound \(B\); the inequality is analytic.
The cycle inequality \(0<m\ln 2-k\ln 3\le k\ln(1+1/(3B))\).cyclesKanıtlandıElementary analysis.
Table rows for \(B=2^{68}\) and \(B=2^{71}\).index, cyclesBilinen literatürExternal records (Barina et al.). Not results of this project.
Geometric-mean factor ≈ \(\sqrt{3}/2\approx 0.866<1\).indexVarsayım/SezgiStandard probabilistic heuristic.
Three negative cycles (smallest |n| = 1, 5, 17).negativesBilinen literatürClassically known; recovered on the site.
Every start in [−200 000, −1] enters one of the three negative cycles.negativesHesaplandıFinite scan; does not prove uniqueness on all \(\mathbb{Z}^-\).
Parity vectors and tail density statistically indistinguishable for positive vs negative maps.negativesVarsayım/SezgiEmpirical / heuristic observation.
Parity walks forever above the critical line: measure zero yet nonempty in the 2-adics.negatives, stopping-timesBilinen literatürStandard 2-adic analysis.
Whether any positive ordinary integer is divergent remains open.negativesVarsayım/SezgiCorrectly stated as open.
Stopping-time density tends to zero (Terras).stopping-timesBilinen literatürTerras 1976.
Union-Find clustering partitions long trajectories into families.familiesHesaplandıComputational construction.
Verification completed in 157 seconds.verificationHesaplandıTiming of a specific run.
A genuine proof must exploit positivity.index, negativesVarsayım/SezgiMeta-mathematical observation, not a theorem.

3. Critical Corrections Required

Correction A — Table rows \(2^{68}\) and \(2^{71}\)

These rows are Bilinen literatür. They rely on external verification records (Barina et al.). Future versions of the cycles page will separate “this project” from “external records”.

Correction B — Language of “verification”

The phrase “the Collatz conjecture holds for all \(n<2^{36}\)” is true only as a computational statement and must never appear without the label Hesaplandı. The conjecture remains open for the infinite set of positive integers.

4. Label Discipline Going Forward

  1. Every new page opens with an explicit statement of which results are new computations, theorems, or literature.
  2. One of the four tags is attached to every numbered claim or highlighted box.
  3. The verbs “solve”, “prove the conjecture”, or “almost solved” are never used.
  4. When a computation is reported: hardware, flags, wall-clock time, and content hash whenever feasible.
  5. When a literature result is cited: precise bibliographic pointer.

5. Summary of Day 01

ItemStatus
Four-label system definedDone
Full claim inventory of current siteDone
Identification of presentation risksDone
Forward discipline for Days 02–90Done

Next (Day 02): Notation and definitions page — \(T(n)\), Terras map, Syracuse map, stopping time, total stopping time, glide — with every definition labelled and referenced.

← No previous day Overview Day 02 → (scheduled)