A Bag Can Fail Every Critical Test and Still Score 44
The authentication checklist returns a score out of 100, and the score on its own can be actively misleading. That is not a flaw that slipped through — it is a property of any weighted average, and the tool works around it by returning a second thing beside the number.
Four ways to answer the same checklist
| What you marked | Score | Critical fails | Verdict |
|---|---|---|---|
| Everything passes | 100 | 0 | No red flags on this checklist, which is not proof. Use a professional or the platform's authentication before you pay. |
| All three critical tells FAIL, rest pass | 43.8 | 3 | Red flag: a critical check failed. Do not pay unless a professional authenticator clears it. |
| All three critical tells pass, rest fail | 56.3 | 0 | Serious doubts. Get it professionally authenticated before any money moves. |
| Every answer is 'unsure' | 40 | 0 | Serious doubts. Get it professionally authenticated before any money moves. |
Read the second row again. Every critical tell has failed — the stitching, the hardware and the serial — and the score is 43.8, barely below the 56.3 of a bag with nothing wrong but wear. Until 30 September 2026 the verdict beside that row read "mixed signals"; it now reads "red flag", because any failed critical tell overrides the total. The number itself has not changed, which is the point of this post.
Why the arithmetic does that
The handbag checklist has 7 items totalling 16 points of weight. The three critical ones carry 9 of it; the other 4 carry 7 between them.
So failing all three criticals still leaves 7 of 16 — 43.8% — available from the non-critical items. A convincing fake that gets the lining, the materials, the symmetry and the dust bag right scores in the forties while failing everything that actually matters.
The third row is the mirror image: a genuine but battered bag with correct stitching, hardware and serial but a worn lining and no paperwork scores 56.3 — lower than the fake.
What to read instead
The tool returns failedCritical: a plain list of the critical tells that failed. That list, not the score, is the output.
Any critical failure is a stop. One is enough, and the verdict now says so whatever the score.
The score is for comparing two items that both pass the criticals. That is the only situation in which it is doing useful work.
"Unsure" is not neutral. It scores 0.4 rather than 0.5, so a checklist answered entirely with "unsure" comes out at 40 — deliberately below the midpoint, because not knowing is a bad sign rather than a neutral one.
The three that carry the weight
For a handbag they are the same three every professional starts with.
Stitching. Even, tight, straight, and the same count per inch throughout. Counterfeits are made fast, and speed shows in the seams before it shows anywhere else.
Hardware. Heavy, cleanly finished, correctly engraved. Hardware is expensive to fake well because it requires tooling rather than skill.
Serial or date code. Present, correctly formatted, and consistent with the brand's own scheme for that era.
Everything else on the list is corroboration. The paperwork carries a weight of 1 out of 16 for a good reason: authenticity cards are the easiest part of the package to counterfeit, and a bag that comes with perfect paperwork and poor stitching is a worse sign than one with neither.
What the score is explicitly not
The module says so in its own documentation: a high score is not a certificate of authenticity. It is a structured way of writing down what you observed, which is more than most buyers do and a great deal less than an expert does.
For anything valuable, use a professional authentication service. The cost is small against the item, most serious platforms include it, and it is the only thing that transfers risk away from you.
Using the checklist properly
Score it before you are emotionally committed, ideally from photographs, before any conversation with the seller.
Ask for the photographs the checklist needs — the stitching close up, the hardware engraving, the serial, the interior stamp. A seller who will not provide them has answered a different question.
Mark honestly. The temptation to mark "pass" on something you cannot really see is what the "unsure" option exists to resist.
Then read the critical list first and the number second.
Every category has its own three
Watches, sneakers, apparel and jewellery each have their own checklist with their own critical tells — a watch's movement, a sneaker's SKU, a garment's woven labels, a piece of jewellery's hallmark. The structure is identical and so is the reading order.
And so is the conclusion: the score is a summary, the critical list is the decision, and neither replaces an expert on an item worth real money.
Why a weighted average behaves this way
Any score that sums weighted parts can be pushed up by the parts that are easy to get right. That is not a flaw in this particular checklist — it is what averages do, and it is why safety-critical scoring systems almost always pair a number with a set of hard gates.
The checklist does exactly that. The number is the summary; failedCritical is the gate. Using only the number is like reading an exam total without noticing a failed compulsory paper.
The counterfeit that scores well
Modern fakes are made by people who know which details buyers check. The commonly-checked things — the shape, the logo, the dust bag, the card — are where the effort goes, precisely because they are what the buyer looks at.
The critical items are the ones that stay hard: stitch density, cast hardware, correctly formatted serials. A fake that fails those and passes everything else is not unusual. It is the normal shape of a good fake, and it is exactly the case the second row of the table describes.
What to do with a low score and no critical failures
This is the interesting case — the third row, at 56.3. Correct stitching, hardware and serial, but a worn lining, no paperwork and asymmetric placement.
That profile is far more consistent with a genuine, well-used item than with a counterfeit, and it is frequently where the good buying is: the price reflects the low score, the doubt is lower than the number suggests, and condition is a known, priceable quantity in a way that authenticity is not.
It is still not proof. A checklist can raise doubts but cannot authenticate, so on anything valuable the order stays the same: score it, then have a professional or the platform's own authentication clear it before you pay.