The information problem with a single photo
A flat dial photo captures the dial typography, the applied logo, the hour markers, and the bezel insert visible from above. It does not capture: the bezel pearl height (only visible at angle), the crown profile (only visible from the side), the case-finishing transitions on the lugs (visible from above only on a polished watch under specific lighting), the bracelet end-link gap at the case, or the rehaut engraving (only visible at a slight tilt). Any of those, alone, is enough to catch a counterfeit. Together they cover the bulk of fake-tells.
Authentication accuracy as a function of input is roughly: 60% on a single dial photo, 75% on three good photos, 92% on the standardised five-angle set, 95%+ when paired with a movement inspection. The jump from one photo to five is the steep part of the curve. The jump from five to a watchmaker inspection is more incremental and very expensive.
What each of the five angles captures
- Dial (head-on). Typography, applied logo, hour markers, hands, date window, dial pattern, signed text. The most-faked region; also the easiest to verify if photographed flat under even light.
- Bezel (head-on, slight tilt). Bezel insert alignment, numerals, pearl height, bezel screws (Royal Oak / Nautilus), tachymeter (Speedmaster), rehaut engraving (modern Rolex). Many counterfeit watches get the bezel pearl wrong because it is a small detail buyers do not always think to check.
- Crown (profile). Crown logo, threading, dot pattern (Rolex Triplock), cabochon stone (Cartier), hexagonal flats (Royal Oak). Counterfeit crowns are cheap to manufacture but hard to manufacture correctly, which is why they fail often.
- Case (profile). Lug shape, case finishing, chamfer, crown guards, case-back fit, lug-to-bracelet transition. The case-finishing tells are some of the most reliable across brands because they require manufacturing tolerances counterfeits rarely hold.
- Finishing (3/4 view). Polish/brush boundaries, bracelet H-link transitions (Royal Oak), end-link gaps, clasp engraving, bracelet stretch (vintage). Often the deciding angle on counterfeits that pass the head-on checks.
Why these five and not seven, or three
Three angles miss too much. Seven angles loses people — buyers stop sending photos after the fifth, sellers stop responding to lists. Five is the empirical sweet spot where capture compliance stays above 90% and inspection accuracy clears the threshold where the verdict becomes useful for buyer-decisioning. WatchRadar uses these five because they are the smallest set that captures the highest-confidence authentication signal.
Why the order matters
The five-angle scan is sequential by design. The Live Loupe walks you through them in the order: dial, bezel, crown, case, finishing. The order matters for two reasons. First, the verdict starts forming after the first two photos — by the time you photograph the crown the model has already narrowed the candidate reference, which lets it weight subsequent angles by what it expects to see. Second, the order forces consistent lighting and pose across users. A reseller running 30 watches in a session benefits from the standardisation; an individual collector benefits from never having to guess what to photograph next.