Every file on this site carries, somewhere in its strata, the things that went wrong. A bezel material that cracked in the heat. A lacquer that would not stay black. A printing plate with the wrong number on it. The makers treated every one of these as a failure, and the record shows them acting on that belief: recalling parts, swapping dials at service, correcting the print run. This dossier is about what the market did next, which was to invert the ledger completely. The components the makers pulled are now the components the money chases, the chemical damage is now the headline of the listing, and the misprint outprices the corrected version. The register does not have to approve of this. It has to explain the rule underneath it, because there is one, and it is the same rule that runs every other dossier in this file: a flaw is worth nothing until the record can date it.
Exhibit one: the part the maker took back
The purest case is the first GMT-Master, the 6542, and its Bakelite bezel insert. The insert was luminous, which meant radioactive, and it was early plastic, which meant fragile: in hot climates it cracked badly enough to be useless. After atomic-energy regulators tested the watches, Rolex recalled them, and the record of what happened next is exactly as messy as a real recall always is. Inserts were swapped for anodized aluminum. Some service centers reportedly scraped the radium out of bezels and dials by hand. One account puts the North American recall at 605 watches, and the sources this register checked do not even agree on the timing, placing the action anywhere from 1956 to the early 1960s, a split the file records rather than resolves. Here is the inversion: today, the single largest factor in what a 6542 is worth is whether that condemned Bakelite insert is still on the watch. The dealers say so in plain language. The component the maker judged dangerous and defective, recalled and destroyed, is now the thing the market pays for precisely because the recall worked and so few survived. Rolex's remedy created the rarity, and the rarity outlived the reason for the remedy.
Exhibit two: the chemistry that would not behave
The second class of defect was never recalled because it happened slowly, on the wrist. The glossy gilt dials of the 6542 and early 1675 era were lacquered, and the lacquer was chemically unstable in two directions. Light, and by several accounts the radiation of the dial's own radium, cooked some of them from black to brown, the condition the hobby calls tropical. Others crazed into networks of fine cracks, and when that happens on the gloss dials of the 16750 the trade calls it a spider dial, a nickname doing exactly the work The Nickname Problem described: naming a condition the maker never would. For decades both were damage, discounted accordingly, and the correct repair was a service dial. Now read a current listing for any of these watches and the browning is in the title. Tropical is a feature. Spider is a rarity. And the service dial that honestly fixed the defect ranks below the defect itself, because the market's hierarchy flipped from condition to originality somewhere around the moment these watches stopped being tools, and an original flaw is evidence while a correct repair is an absence. The maker's definition of fixed and the market's definition of valuable now point in opposite directions, which is the condition this register's founding dossier, Watches That Disagree With Their Makers, exists to record.
Exhibit three: the number that was simply wrong
The third exhibit is not decay but a plain mistake. For roughly seven weeks in the fall of 1970, Speedmaster 145.022-69 cases left the line wearing a tachymeter bezel printed 225, 220, 190 instead of 225, 200, 190. This is not a variant with a story; it is arithmetic failure, because sixty times sixty divided by eighteen seconds is two hundred, and the marker dots under the numerals sit exactly where the correct bezel's dots sit. The scholarship has pinned the error to a serial band around 30.588.xxx to 31.008.xxx, a window later research has already widened at the top, the record revising itself in public the way it did all through One Reference, Two Watches. Omega caught the mistake and corrected the print. The corrected bezel is everywhere and worth little. The wrong one is now among the rarest Speedmaster bezels there are, and the phrase sellers use for it is the most register-shaped sentence in this whole dossier: a genuine 220 error bezel, correct for the serial. Read that twice. The error has its own correctness criteria. A mistake, properly dated, stops being a mistake and becomes a reference point.
The rule underneath
Put the three exhibits side by side and the rule is visible. The Bakelite insert is valuable because the recall dates it: nothing but a period 6542 carries it. The tropical and spider dials are valuable because the chemistry dates them: only certain lacquers from certain years fail that way, which is why the browning must sit on the right dial type for the right era, verified against the serial record, before it is worth anything. The 220 bezel is valuable because seven weeks and a serial band date it to the month. In every case the defect is doing the one thing this site exists for: it is testifying about when the watch was made, more precisely than the reference number ever could. An undated flaw is damage. A dated flaw is a variant. That is the entire inversion, and it is also the warning, because a premium on damage is an invitation, and the trade answered it: baked dials, faked spiders, reprinted bezels, every one an attempt to counterfeit not a watch but a defect. The defense is the same as everywhere else in this register. The flaw must agree with the serial, the serial must agree with the strata, and when a watch's damage tells a tidier story than its paperwork, believe the paperwork. The makers spent the twentieth century trying to build watches without flaws. The market spent the twenty-first paying extra for the failures, and the register's job is neither to cheer nor to scold but to keep the one thing that separates a precious mistake from an ordinary one: the date.