How WatchMerit calculates a fair price
A walkthrough of the median engine, the ±5–15% fair band, and why we exclude editorial pages from the price signal.
The single question WatchMerit exists to answer is simple: for this exact reference, in this exact condition, what is a fair price today? Everything else — the WatchMerit Score, the 12-month outlook, the deal filter — is downstream of that number.
This post walks through how the number is built, what goes into it, and what deliberately does not.
1. The median, not the average
For every reference we monitor, we scrape active dealer listings worldwide. We take the median, not the mean, because a single Sotheby's outlier can drag a mean up by 20% and quietly break every downstream signal.
The median is refreshed monthly and stored as a market snapshot so charts and per-watch stats are all reading the same benchmark.
2. The fair-price band
Around the median we draw a ±5–15% band. Steel sports references from Rolex, AP, and Patek get the tighter band because their liquidity makes prices converge quickly. Vintage and independent references get the wider band because condition and provenance dominate.
A listing at or below the median gets flagged as a good buy. A listing above the top of the band gets flagged as overpriced. Everything in between is fairly priced.
3. What we deliberately exclude
Editorial pages, brand press releases, and auction house catalogues are excluded from the median. A Hodinkee review of a Daytona is not a listing you can buy — including it would poison the benchmark.
We also cap the maximum discount at 40%. Anything cheaper than that on the deals page is almost always a mis-parse from a source page and is filtered out or re-validated the next day.
4. Currency and landed cost
International listings are converted to AUD at the current spot rate and shown as a landed price where duty and GST apply. This lets you compare a Sydney dealer, a Zurich dealer, and a Tokyo dealer on the same axis.