Short answer
3399 is not an Audemars Piguet reference number. Genuine perpetual calendar models from the brand carry five-digit references such as 25820, 26574, and 26674. Sellers who label a watch 'AP perpetual calendar 3399' are using an internal batch code, and that code is shared across several factories and case finishes. Because the label has no official meaning, nothing else in the listing can be taken at face value. You have to verify the calendar mechanics, the dial print, the case measurements, and the movement finish by hand before money changes hands.
Before you start
- A 10x jeweler's loupe
- Digital calipers accurate to 0.01 mm
- A scale accurate to 0.1 g
- A timegrapher, optional but decisive on movement questions
- Photos of the genuine reference pulled from the maker's own model list
- Bright raking light, such as a desk lamp placed at a shallow angle
How to inspect the watch, step by step
- Look up the reference number in Audemars Piguet's official model list and write down the case size, thickness, caliber, and dial layout for that exact reference.
- Read the engraving on the caseback and compare the caliber number with the one you just wrote down.
- Measure the case thickness with calipers. Genuine perpetual calendar pieces in steel sit near 9.5 mm; most counterfeits are 1 mm to 2 mm thicker because they carry a stacked calendar module.
- Weigh the watch head and the bracelet separately, then compare both numbers with a documented genuine example of the same reference.
- Examine the subdials with the loupe and confirm the layout: month and leap year indicator at 12, date at 3, moonphase at 6, day of the week at 9.
- Press each corrector with the supplied pusher and watch the matching disc jump. A genuine corrector sits flush with the caseband and snaps once.
- Turn to the moonphase disc and count the stars under the loupe. A genuine disc shows a soft three-dimensional moon over a varnished blue field; printed moons with hard edges and misaligned star clusters are a giveaway.
- Study the hands and hour markers at an angle. Genuine hands carry a polished bevel along every edge with no burrs and no clouding under the lacquer.
- Read the dial print with the loupe. Lettering on a genuine dial stays crisp at 10x with even spacing; on a counterfeit the print frays and the brand signature sits too close to the logo.
- Run a fingernail along the bracelet edges and around the bezel screws. Genuine bevels are cut by hand, so surface transitions are sharp and flat; counterfeits feel rounded or show polishing drag.
- Look through the display back and check for hand finishing: bevelled bridges, Geneva stripes under the rotor, blued screws with polished slots, perlage on the mainplate.
- Look for a spacer ring or a second plate stacked above the main plate. A calendar module bolted onto a common base caliber is the most frequent construction in a counterfeit perpetual calendar.
- Put the watch on a timegrapher and record amplitude, rate, and beat error in three positions. Wide swings between positions point to an unstable base movement.
- Ask for the original certificate and compare the case number and movement number on the papers with the numbers engraved on the watch.
- Check the asking price against auction results and dealer listings for the same reference. A genuine perpetual calendar starts in the tens of thousands; a four-figure asking price answers the question on its own.
What counterfeit perpetual calendars get wrong
The failure points repeat across batches. The calendar display may run as a simple annual calendar, so the date still needs manual correction at the end of February. The leap year indicator is often printed rather than driven, which means the pointer never moves when the month changes. Moonphase discs drift because the gear train does not divide into a 29.5-day cycle. Case profiles lose the flat bezel facets, bracelet links develop play after weeks of wear, and sapphire crystals show internal reflections that genuine crystals do not produce.
When to walk away
Walk away if the seller will not let the caseback come off, if the reference number has to be explained, if the papers and the engraved numbers do not match, or if the price sits below the range that auction houses report for the same reference. None of those problems can be corrected after you pay.