# Deprecated / superseded site material

_Last updated: 2026-07-08_

Out-of-date or retracted material from masonborchard.com, kept here so it stays
referable but clearly fenced off from anything live. Nothing in this folder is
current. Each entry says what it is, why it is void, and where the corrected
version now lives. This mirrors the Ionosphere vault's own `deprecated/` store.

The corrections these trace back to (authoritative):

- `ionosphere-vaults/05_analysis-methods/verified-numbers/CORRECTIONS_BRIEFING_2026-07-08.md`
- `ionosphere-vaults/05_analysis-methods/verified-numbers/SINGLE_SOURCE_OF_TRUTH.md`

## What changed (July 2026 correction cycle)

Three Ionosphere findings were retracted or corrected after they had already
gone up on this site:

1. **NYE 2025/2026 temporal clustering was a data-duplication artifact.** The
   parser that built the combined NYE series ingested the same periods twice
   under two naming conventions, double-counting 1,326 of 4,149 periods. Sorted
   by time, each duplicate sat next to its identical twin — the strongest
   possible false input to every *sequential* statistic. Deduplicated (n=2,823),
   NYE is temporally null on both units. Every NYE clustering number is void:
   runs test p < 0.0001 (dedup: z=+0.35, p=0.73), Markov 36.27% vs 3.11% (7.8x),
   lag-1 autocorrelation +0.326, 12 clusters, change-point Jan 1 09:09, Fisher's
   g periodicity p=0.011 (dedup: p=0.93), and the meditation-session alignment
   read off that structure.
2. **The "12.9-fold" ACORN headline overclaims.** It was the single tightest
   7-minute Autumn Lights meditator window — 14 periods, p=0.30. An underpowered
   magnitude, not a result.
3. **Variance-ratio and DFA inferential framing is void/weak** corpus-wide (VR
   used the wrong standardizing formula — ~22x z inflation at k=20; DFA alpha is
   descriptive only, its p-value comes from an anti-conservative fit).

The honest current picture, now live on `ionosphere.html`: the corpus is
aggregate-null and, after the NYE correction, shows **no endogenous temporal
clustering**. The only surviving signals are exploratory covariate-window burst
enrichments — the Autumn Lights meditator window (post-peak binomial p=.011) and
the July 4 2026 focus windows (Fisher p=.0066, flat on a same-machine PRNG
control) — which motivate a preregistered follow-up.

## Snapshots (verbatim, as they stood on 2026-07-08 before correction)

`2026-07-08/snapshot/` holds the pre-correction copy of every file that was
edited, so the original wording is never lost:

| File | What's void in it | Corrected version |
|---|---|---|
| `ionosphere.html` | The NYE section (results table: runs p<0.0001, Markov 36.27%, Fisher's g p=0.011, change-point, DFA, extreme values, Q2 Stouffer), the `nye_burst_clustering.png` figure + caption, the meditation-session-alignment paragraph; the "temporal structure — bursts, clustering, autocorrelation, periodicity" thesis framing; the "strongest temporal-burst signature" limitation; DFA inference in the Autumn Lights section | Live `ionosphere.html` (NYE reframed as a clean null control; thesis reframed to aggregate-null + exploratory window enrichment; July 4 2026 added as the controlled anchor) |
| `articles.json` | `pres-002` (ACORN 2026) title + abstract carrying the 12.9-fold claim and the void NYE clustering numbers | Live `data/articles.json`, same entry, now stamped with a dated erratum (title kept as the record of what was delivered) |
| `articles_safecopy.json` | Mirror of the above (the 12.9-fold title on `pub-008`, the void NYE abstract on `pres-002`) | Live `data/articles_safecopy.json`, same erratum treatment |
| `why-patterns.html` | The "12.9-fold burst-detection effect reported at ACORN 2026" body claim and the ref [22]/[23] citations presented without correction | Live `articles/why-patterns.html` (body reworded to the current picture; refs [22]/[23] carry correction notes) |
| `nye_burst_clustering.png` | A plot of the void duplication artifact | Removed from the live page; no replacement (the result it plotted does not exist post-dedup) |

## On the ACORN 2026 record (why some void tokens remain live)

The ACORN 2026 talk — *"Immersive Audiovisual Feedback Amplifies
Consciousness-RNG Effects by 12.9-Fold Through Temporal Burst Detection"*
(Borchard, Beach & Caputi) — was **already delivered on March 1, 2026**. Per
Mason's decision (2026-07-08), the delivered title and abstract are kept on the
live articles page as the historical record of what was presented, with a dated
**erratum** attached rather than being rewritten or removed. As a result, the
strings "12.9-fold", "Fisher's g p=0.011", "4,149", etc. still appear live —
but only inside that one fenced, erratum-stamped historical entry (and its
`why-patterns.html` citations). Everywhere else on the site they are gone or
moved here.

The WRF winter-storm coherence-sampling claim was **not** re-audited in this
cycle (winter_storm is the Dani-only run excluded from the RNG corpus); it is
kept live with a "pending re-audit" caveat, not asserted as verified.