Re-grades: a new basis, both numbers disclosed

A re-grade is the rarest move this desk makes. A score is not re-argued because a second look felt different; it is re-graded only when the basis itself changes, when the record a product was first graded on turns out to be the wrong ground for it, and then both numbers are disclosed on the card, in the ranking data and in the prose, so no reader has to wonder what the product used to score or why it moved. The October pass made three.

Dr. Bronner’s Peppermint bar: 80, then 86. On Card No. 69 the bar stood at 80, graded as a guest among liquid body washes, on a liquid shelf’s ground. On the bar soaps Card it was graded again on its home basis, against bars, and earned 86, first on that shelf. The re-grade is printed on the bar’s card and in the Card’s prose, in so many words: both numbers are re-grades and both are printed as such. The bar did not change. The company it was graded in did, and the page says so.

The Honest Amish Original Beard & Body Soap: 75, then 78. The same Card, the same shape. The soap first stood at 75 on Card No. 114, graded as a different species among liquid beard washes, its cleansing row discounted for the company. Re-based to home ground on the bar soaps Card, one row moved, the species row, 13 to 16, and the total followed to 78, second on the shelf. The rows that did not move are printed unchanged, and the one that did is named, which is what a re-grade on a new basis means in practice: the arithmetic is shown, not summarised.

The Remington MB4700 trimmer: 48 on one record, 61 on the fuller one. The beard trimmer Card first graded the MB4700 at 48 on its maker’s European record, the record that stood open when the shelf was built. The maker’s American page, opened at the dead-door backfill, prints a materially fuller record: the full range in print, 0.4 to 18 millimetres, a complete specification block and a five-year limited warranty where the European record printed two years. The shelf grades American maker records throughout, so the trimmer was re-graded in the open on the fuller one: 61, rows re-derived and printed, both numbers disclosed on the card, in the ranking data and in the prose, in the Card’s own words, 48 on the European record, 61 on the American. It stays seventh on its shelf, one point behind the seat above it, and the re-grade is entered here because a number that doubles its warranty term on a fuller record is exactly the kind of move a ledger exists to make visible.

Audit verdicts: derived becomes audited

Most brand scores on this site begin as derivations: built in the open from the garments of that brand the desk has already scored, with the arithmetic printed where it can be argued with. A house audit re-tries the number against the maker’s own current pages, read entire, and returns one of two verdicts. It holds the number entire, or it moves it, usually by a point, usually on evidence the derivation never weighed: a repair bench, a resale programme, a take-back scheme standing behind the garments. Either verdict flips the status word from derived to audited, on the brand’s own page and on every page that cites the number, because a status that is true on one page and stale on another is a small lie told at scale. The October pass audited twelve houses. Five totals moved. Seven held entire.

Moved

Asket: 92 derived, 93 audited. The audit on the Asket page re-derived every row from the maker’s own current pages and stood the number one point above the derivation set in our Scandinavian ranking: the permanent collection’s receipts, published entire, weighed at house level for the first time. The verdict stands on the anchor’s page.

Buck Mason: 90 derived, 89 audited. Set at 90 in our American ranking and carried as a derivation, the number was audited on the Buck Mason page against the maker’s own record and came down one point: no afterlife of any kind stands behind the garments, and the longevity row is graded on exactly that. The audit prints the whole re-derivation.

John Smedley: 79 derived, 78 audited. Derived in the open on the Sunspel page from the garment record, the score was re-tried on the Smedley page: four rows held, and the longevity row moved one point on the absence of any repair, resale or take-back programme behind a six-month fault remedy. The moved row is named on the page.

Everlane: 67 derived, 68 audited. The audit on the Everlane page lifted the longevity row a single point, on the resale route and the one repair warranty the derivation never counted, and no other row moved: the transparency pitch survives its audit, at a number one point kinder and considerably better evidenced. The verdict is printed in place.

Charles Tyrwhitt: 65 derived, 66 audited. The 65 was derived in the open on the Spier & MacKay page; the audit on the Tyrwhitt page re-tried the house on its own current record and moved one row, longevity, on a take-back programme the earlier passes had not weighed. The chain is printed in full on the page.

Held entire

Quince: 59, held entire, audited. Re-tried at house level on the Quince page, every row confirmed: no programme of any kind stands behind the garments, and the number the derivation set is the number the audit keeps. Verdict on the page.

Sunspel: 85, held entire, audited. The audit on the Sunspel page re-derived every row from the maker’s own current pages and held the number entire, in the page’s own words. Verdict on the page.

Todd Snyder: 79, held entire, audited. Tested on the Todd Snyder page against the maker’s own pages and held entire, rows and all, on the page where inflating it would be easiest. Verdict on the page.

Spier & MacKay: 76, held entire, audited. The standing score was audited in the open on the Spier & MacKay page and confirmed by the house audit that follows the ranking, against the strongest garment record any anchor has brought to the series. Verdict on the page.

L.L.Bean: 87, held entire, audited. Re-tried at house level on the L.L.Bean page, where the audit after the ranking holds every row: a repair service stands, and the guarantee as printed today carries its one-year term, both weighed and both printed. Verdict on the page.

Barbour: 81, held entire, audited. Derived in the open at 81 on the Barbour page and audited at house level after the ranking, where the number holds, repair and re-wax ledger and all. Verdict on the page.

mfpen: 84, held entire, audited. The anchor’s 84 was re-tried at house level on the mfpen page and held entire, rows unchanged, the status flipped to audited on every surface that states it. Verdict on the page.

Price and record corrections, printed in the open

Prices drift faster than scores, and a stale price is an error with a date on it. The October pass corrected the record in four places a reader is entitled to know about, each printed on the page that carries the figure, not whispered into a quiet edit.

The Asket sweatshirt is $180, and the capsule totals were re-based. The sweatshirt Card verifies the garment at $180 in the maker’s US store. The capsule wardrobe page had printed that $180 in its sweatshirt row all along, while its totals were still carrying the garment at $95, a figure that traces to a 2018 review, so the totals ran $85 under the sum of their own columns. The correction is printed in the open on the capsule page itself, under its own heading: the totals are re-based to add the printed columns exactly, and they now stand at $1,460.38 for the starter build, $2,020.79 for the recommended build and $2,697.98 for the forever build, with the pages that cite those totals conformed to match.

The Asket tee stands at $65. The maker’s US price for the tee is $65 today, and the tree prints it so, on the white tee ranking and wherever the garment is cited. Earlier figures standing in older copy were conformed in the pass; the $65 is the checked record, and it is the figure every total that includes the tee is built on.

The Buck Mason tee stands at $48. The curved-hem tee prints at $48 at the maker’s store, and the tree now prints $48 tree-wide, including on the American brands ranking, where an older figure stood before the pass. A two-dollar drift on a fifty-dollar garment is a small thing, which is precisely why it is entered: a desk that corrects the small figures in the open can be believed about the large ones.

The Todd Snyder + Champion pocket sweatshirt: the $168 record stands, flagged. The American ranking prints the Todd Snyder + Champion pocket sweatshirt at $168 at the official store, and that record stands as the garment’s dated check. The audit on the Todd Snyder page prints the flag at the strength it is entered here, no stronger: the named collaboration piece does not stand in the maker’s current published inventory under the collaboration’s name; the pocket sweatshirt the maker sells today is the mainline Foundation garment at $98, and the collaboration’s standing fleece at the audit is a Champion henley sweatshirt listing at $198. The two records are kept apart, nothing from the collaboration line is blended into the mainline record, and the fuller re-check of the garment’s standing is owed at the November pass. Until it lands, the flag stands here and on the audit page, in the open, where a reader pricing the garment will meet it.

The standing rules this ledger enforces

Rule one: a re-grade needs a new basis, and both numbers. A score is never re-argued on a second feeling. It moves on a new basis, a fuller record, a home ground, a programme the first pass never weighed, and when it moves, the old number is disclosed beside the new one on the card, in the ranking data and in the prose. A reader should never have to take a change on trust, or discover one by comparing screenshots.

Rule two: a status flip is conformed tree-wide. When a derivation becomes an audit, the word changes everywhere the number is cited, on the anchor’s page, on its neighbours’ pages, in the rankings and in the summaries machines read. A status that is true in one place and stale in another is an error the site is telling about itself, and the pass is not finished until the sweep is.

Rule three: a record that converts is never silently edited. A page a reader has already acted on keeps its record. Corrections to it are printed, dated and entered here, with the old figure standing; they are never slipped in under the same date as though the page had always read that way. The stamp on a page is a promise about when its record was last checked, and a quiet edit breaks it retroactively.

What has not changed

For balance, and because a ledger that only lists movement invites the wrong inference: the derivations standing on this site stand exactly as they were built, verbatim, as dated history, including the ones the audits later confirmed and the ones they corrected. The audit sections argue with them in the open; they do not rewrite them. Scores the October pass did not re-read did not move, and are not implied to have been re-verified by appearing anywhere near this page. And nothing in this ledger reaches backward: a figure that was true on its check date and has since drifted at the maker’s store is not an error, it is a date, which is why every page carries one. This page records verdicts. It does not airbrush them, and it does not airbrush the record they were reached from.

Questions, answered

How often is this ledger updated?

Whenever a score, a price or a record changes, and the entry is made in the same pass that makes the change. The desk re-checks its rankings on a rolling cadence, flagship rankings at least monthly and seasonal pieces before each season turns, and every re-grade, audit verdict and material correction the pass produces is entered here with its date, its old figure and its new one. A change that is not in this ledger has not been made; a figure that moves without an entry here is an error, and we want it reported as one.

How can a reader report an error?

Through the contact route printed on this site, naming the page, the figure and the record that contradicts it: a maker’s current product page, a published size chart, a price standing at the maker’s own store. Reports are checked against the maker’s own record, not against other roundups. If the report is right, the page is corrected, its checked date is updated, and the correction is entered in this ledger with the old figure standing beside the new one, so the error stays on the record rather than being edited out of it.

What happens to a score when a maker changes its record?

Nothing, until the desk re-reads the record and prints a verdict. A maker’s new price, new programme or new terms does not move a score by itself, because a score grades the record as the desk read it on its check date, and that reading stands as dated history. When the desk does re-read it, one of two things happens in the open: an audit, which re-tries the number on the maker’s own current pages and either holds it entire or moves it with the reason printed, or a re-grade on a new basis, which is allowed only with both the old number and the new number disclosed on the card, in the ranking data and in the prose. Either way, the change is entered here.

What is the difference between derived and audited?

A derived score is built in the open from evidence the site already holds: the garments of that brand this desk has scored, the published record those garments stand on, and arithmetic a reader can follow and argue with on the page. An audited score has been re-tried at house level against the maker’s own current pages, read entire on the audit’s check date: the cloth, the construction, the fit system, the terms and whatever repair, resale or take-back programme the house actually runs. The audit either holds the derived number entire or moves it, and prints which. Most audits in the October pass held; the ones that moved, moved on evidence the derivation had never weighed, usually a programme standing behind the garment.

Why do the old numbers stay printed at all?

Because a ranking that quietly rewrites its own past cannot be checked, and being checkable is the whole offer. The derivation a score was built on stands verbatim as dated history on its page, the old totals stand in this ledger beside the new ones, and the audit sections print what the earlier pass got right as carefully as what it missed. A reader who cited our 92 for Asket last month can find the 93, the date it moved and the reason, without taking anyone’s word for any of it.

Does a correction ever change which product or brand wins?

It can, and when it does the new order stands on the page with the old one recorded here. In the October pass no re-grade changed a winner: the Bronner’s bar held its seat at its new number, the Remington trimmer stayed seventh at its new one, one point behind the seat above it, and the five audits that moved a total moved it by a single point each, inside fields whose order the point did not disturb. The rule does not promise that outcome. It promises that if a correction ever does re-seat a ranking, the re-seating is printed, dated and explained, here and on the page, rather than discovered.

Next from Young Cotton

The bases this ledger enforces stand at the methodology page: the published criteria, the re-check cadence and the corrections policy this page is the working record of. The commerce terms behind the site’s links stand at the disclosure page. This ledger’s next entries will be made by the November pass, and they will be entered here, with their dates, when they are made.