01
Begin with evidence-rich photographs
Start with a sharp, evenly lit view of the dial and case rather than a dramatic wrist shot. Keep the camera square to the watch, reduce reflections, and capture the full outline. Add separate views of the case sides, crown, clasp, caseback, and any paperwork the seller is willing to show. A single blurred image can suggest a family; a deliberate set gives you details to test.
Example: if the first photograph hides the date window beneath glare, request a second dial view before deciding between two similar candidates. Photographing the crown and bezel straight-on may expose differences in shape or markings that the original angle concealed. The objective is not more images for their own sake; it is a set in which each image answers a specific identification question.
02
Read the visible cues as a system
Work from broad traits toward small ones. Note the case material and shape, dial layout, hour markers, hand style, bezel markings, date position, crown guards, bracelet or strap attachment, and visible text. Then ask whether those traits belong together on one known reference. A familiar logo or color is only a starting clue because related models and later replacements can look similar at first glance.
Treat inconsistency as a prompt to investigate, not as instant evidence of wrongdoing. A service dial, replacement bracelet, aftermarket part, poor photograph, or incorrect seller description can each produce a mismatch. Write down what supports the candidate and what conflicts with it. That evidence list is more useful than silently choosing the closest-looking result and forgetting the unresolved details.
03
Test the candidate beyond visual similarity
Once you have a candidate, compare its expected specifications with the watch in front of you. Check case proportions, dial configuration, movement family when legitimately visible, bracelet or clasp references, and the placement and wording of engravings. Prefer manufacturer archives, official manuals, and records supplied with the watch when they are available; marketplace titles can repeat one another’s errors.
Example: two watches may share a black dial and steel bezel while differing in case diameter, date treatment, or bracelet end links. A candidate that matches the dial but conflicts with several structural details is weak. Record the candidate in Patina AI as something to review, then keep the disputed cues attached to the decision rather than presenting the reference as settled fact.
04
Separate identification from authentication
A photo-based identification candidate does not replace professional authentication. Photographs usually cannot establish the condition and originality of internal components, validate provenance, or rule out sophisticated substitutions. Serial and reference markings may also require physical examination. For a purchase where authenticity materially affects the decision, use a qualified professional who can inspect the watch directly.
Use the identification result to improve the next conversation. Ask for the missing views, compare the physical watch with the expected configuration, and state uncertainty plainly when you share your conclusion. Patina AI can organize visible cues and reference context, but the final decision should remain proportional to the evidence and the consequences of being wrong.