How do you check the stitching and hardware engravings on a pre-owned bag?

Short answer: Count the stitching and read the engravings. With a millimetre ruler, count stitches per centimetre and check the spacing is even; under side lighting, compare the font, depth and letter spacing of the hardware. Compare both against references of the same model and era, and remember that photos have limits and that repairs and wear change both details.

The site tool gives a visual-consistency probability for reference only. It is not an official authentication certificate and is not a verdict on authenticity; for authentication, use the brand or a professional authentication service.

Stitching: numbers beat impressions

Lay a ruler flat along the seam and count how many stitches fall within one centimetre. Most models keep a fairly stable count, so a large gap from the reference value deserves attention. A difference of a single stitch can easily be measurement error and is no cause for alarm.

Then look at evenness. Hand and machine stitching both show small variations, but normal spacing is broadly consistent. Spacing that alternates between tight and loose, slanted stitches and loose thread tails are all worth a second look.

Repaired areas may be stitched differently from the original. If the seller has disclosed a repair, recording a slight variance is the fairer choice.

Hardware: read engravings in side light

Engravings should be viewed with light coming from the side, which throws small shadows along the letter edges and reveals depth and spacing. Against a reference of the same model and era, compare three things: the shape of the letter strokes, the spacing between letters and the depth of the engraving.

Wear makes engravings shallower and rounds their edges, which is normal. Some hardware may have been replaced, so it is fair to ask the seller. Only a clearly different font or clearly uneven spacing should be recorded as clearly inconsistent.

How much weight these two carry

In the tool we built, hardware and stitching are the two more heavily weighted features, because both involve many small details that are hard to get right everywhere at once. Every likelihood ratio, though, is an illustrative parameter awaiting calibration with real inspection data.

If you can see only one of the two clearly, that is fine. Record what you can see and mark the rest as unobservable. The result will stay closer to the prior, which is an honest sign that evidence is still thin.

When to stop and send it for authentication

If both stitching and hardware look clearly inconsistent, stop studying photos and send the item for authentication. If only one shows a slight variance, ask for more photos and read it together with the other four features. A prescreen helps you choose the next step; the conclusion belongs to the brand or an authentication service. Organising the stitching and hardware photos, labelled by location, makes their job easier.

Follow-up questions

Can stitch counts change between production years?

They can, which is why references should match the model and era and come from a reliable source.

A slightly blurred engraving: does that count as inconsistent?

Not necessarily. Wear softens engravings, so a slight variance may fit better; ask for a side-lit macro shot.

I do not know the reference stitch count. What then?

Count it yourself from reliable photos of the same model, or record it as unobservable rather than guessing.

The prescreen we built combines your stitching and hardware observations with the other four features and returns an appearance-consistency probability with a range. It is a reference probability based on appearance, not an official authentication certificate and not a conclusion on authenticity; for authentication, please use the brand or a professional authentication service.

About this tool

Sources

  1. Hand, D. J. & Yu, K., 2001, Idiot's Bayes: Not So Stupid After All? International Statistical Review 69(3)
  2. ISO, 2020, ISO 22383: Security and resilience, Authenticity, integrity and trust for products and documents, Guidelines for the selection and performance evaluation of authentication solutions