UBUDASHI

Rolex · thirty references

What the guides miss

13,898 watches dated by case number, set against 3,132 facts from 49 guide sites

Collector guides date a change by year and take one reference at a time. The library can do something they cannot: lay thousands of case numbers from thirty references end to end, in the order the cases were numbered, and see what was on each watch. These are the places where that view adds to the guides or disagrees with them.

Each finding gives what the guides say, what the library shows, and how far to trust it. Three kinds of evidence are used, in falling order of strength. Case numbers alone are the firmest. Labels read from photographs come from an automated reading that was right at least 96% of the time on watches held back for testing. Sellers' descriptions are the loosest, and are used only where a pattern is too sharp to be a habit of speech.

9,650 case numbers known to the nearest 20,0004,000 published in full6,502 watches with a dial read from a photograph35 close-ups on this page

Rolex left three stretches of case numbers unused, and gave two more to its strap watches

The guides give serial charts that list the number each year began with. Read as they stand, they suggest the count ran on without a break. Some place a year's first number where the library finds almost nothing: one chart starts 1975 at 4,265,000, 1981 at 6,910,000 and 1983 at 7,860,000; another starts 1983 at 7,862,000.

The library holds 9,650 watches from the thirty references with a case number known to the nearest 20,000. Between 3.9 and 10 million they turn up at about 91 watches for every 100,000 numbers, except in five stretches where the thirty are absent or nearly so: together about 1.0 million numbers holding 4 of their watches, none with a number published in full. Three of the five are all but empty of every Rolex reference in the library: 1 watch of any reference from 4.54M to 4.70M, 3 watches of any reference from 4.77M to 5.00M and 11 watches of any reference from 7.66M to 8.00M. The other two were used, for other watches: 92 from 4.26M to 4.38M and 43 from 6.80M to 6.96M, 119 of those 135 on references beginning 3 or 4, the Cellini and other strap watches. Two thin runs sit between them: 32 watches from 4.38M to 4.54M and 20 from 4.70M to 4.77M. The count also stops short: no watch lies above 9.92M, so the 9,999,999 that some charts give as the last number before the letter R was probably never reached.

0501003.90M to 3.95M: 75 watches3.95M to 4.00M: 70 watches4.00M to 4.05M: 55 watches4.05M to 4.10M: 44 watches4.10M to 4.15M: 84 watches4.15M to 4.20M: 45 watches4.20M to 4.25M: 55 watches4.25M to 4.30M: 1 watch4.30M to 4.35M: 1 watch4.35M to 4.40M: 11 watches4.40M to 4.45M: 5 watches4.45M to 4.50M: 14 watches4.50M to 4.55M: 2 watches4.70M to 4.75M: 12 watches4.75M to 4.80M: 8 watches5.00M to 5.05M: 73 watches5.05M to 5.10M: 98 watches5.10M to 5.15M: 72 watches5.15M to 5.20M: 66 watches5.20M to 5.25M: 73 watches5.25M to 5.30M: 100 watches5.30M to 5.35M: 37 watches5.35M to 5.40M: 71 watches5.40M to 5.45M: 54 watches5.45M to 5.50M: 81 watches5.50M to 5.55M: 84 watches5.55M to 5.60M: 62 watches5.60M to 5.65M: 68 watches5.65M to 5.70M: 49 watches5.70M to 5.75M: 87 watches5.75M to 5.80M: 39 watches5.80M to 5.85M: 65 watches5.85M to 5.90M: 65 watches5.90M to 5.95M: 30 watches5.95M to 6.00M: 56 watches6.00M to 6.05M: 65 watches6.05M to 6.10M: 50 watches6.10M to 6.15M: 53 watches6.15M to 6.20M: 100 watches6.20M to 6.25M: 38 watches6.25M to 6.30M: 55 watches6.30M to 6.35M: 37 watches6.35M to 6.40M: 38 watches6.40M to 6.45M: 26 watches6.45M to 6.50M: 31 watches6.50M to 6.55M: 33 watches6.55M to 6.60M: 36 watches6.60M to 6.65M: 22 watches6.65M to 6.70M: 40 watches6.70M to 6.75M: 38 watches6.75M to 6.80M: 24 watches6.95M to 7.00M: 4 watches7.00M to 7.05M: 8 watches7.05M to 7.10M: 10 watches7.10M to 7.15M: 44 watches7.15M to 7.20M: 21 watches7.20M to 7.25M: 42 watches7.25M to 7.30M: 58 watches7.30M to 7.35M: 35 watches7.35M to 7.40M: 36 watches7.40M to 7.45M: 69 watches7.45M to 7.50M: 41 watches7.50M to 7.55M: 27 watches7.55M to 7.60M: 18 watches7.60M to 7.65M: 34 watches7.65M to 7.70M: 4 watches7.90M to 7.95M: 1 watch8.00M to 8.05M: 28 watches8.05M to 8.10M: 70 watches8.10M to 8.15M: 27 watches8.15M to 8.20M: 28 watches8.20M to 8.25M: 47 watches8.25M to 8.30M: 73 watches8.30M to 8.35M: 48 watches8.35M to 8.40M: 67 watches8.40M to 8.45M: 56 watches8.45M to 8.50M: 43 watches8.50M to 8.55M: 41 watches8.55M to 8.60M: 33 watches8.60M to 8.65M: 43 watches8.65M to 8.70M: 49 watches8.70M to 8.75M: 46 watches8.75M to 8.80M: 72 watches8.80M to 8.85M: 39 watches8.85M to 8.90M: 41 watches8.90M to 8.95M: 37 watches8.95M to 9.00M: 38 watches9.00M to 9.05M: 40 watches9.05M to 9.10M: 31 watches9.10M to 9.15M: 45 watches9.15M to 9.20M: 52 watches9.20M to 9.25M: 15 watches9.25M to 9.30M: 11 watches9.30M to 9.35M: 68 watches9.35M to 9.40M: 60 watches9.40M to 9.45M: 53 watches9.45M to 9.50M: 50 watches9.50M to 9.55M: 19 watches9.55M to 9.60M: 33 watches9.60M to 9.65M: 45 watches9.65M to 9.70M: 37 watches9.70M to 9.75M: 32 watches9.75M to 9.80M: 37 watches9.80M to 9.85M: 35 watches9.85M to 9.90M: 32 watches9.90M to 9.95M: 1 watch4M5M6M7M8M9M10M1975 starts here (one chart)1981 starts here (one chart)1983 starts here (two charts)

Watches of the thirty references per 50,000 case numbersStretch with almost none of themWhere a chart starts a year

StretchLengthWatches of the thirty foundExpected at the usual rateWatches of any Rolex reference
4.26M to 4.38M120,000 numbers1about 10992
4.54M to 4.70M160,000 numbers0about 1461
4.77M to 5.00M230,000 numbers0about 2103
6.80M to 6.96M160,000 numbers2about 14643
7.66M to 8.00M340,000 numbers1about 31011

A case number inside one of the empty stretches is more likely a misreading, a wrong reference or a problem with the watch than a real Rolex number. In the stretches given to the strap watches, a number on an Oyster sports model or a Datejust is unusual and deserves a second look at how it was read, no more than that. It also means a chart cannot be read by drawing a straight line between two years: the line crosses numbers that were never used.

How sure: strong. It rests on case numbers and nothing else, and 4,000 of them are published in full by auction houses. Every Rolex reference in the library was counted as well, 19,718 watches, to tell a stretch that is empty from one that is only empty of these thirty.

References were cased in batches

The guides mention batches for a handful of rare watches. Two of them list the three batches of the Patent Pending Sea-Dweller, for example. One of them also uses batches on the Daytona, to check a watch against its neighbours. None shows batches as the rule across ordinary references.

The library shows that it was the usual way. The case numbers of a reference bunch into narrow blocks with long empty runs between, while other references fill the gaps. Counting only numbers published in full, since a number with its last digits hidden falls into one block whatever it was, half of a reference's watches typically sit in 4% of the 10,000-number blocks between its first and last. The same number of watches scattered at random over the same stretch would need about 10%. 24 of the 28 references with thirty or more full numbers are bunched by a clear margin, into at least a fifth fewer blocks than random scatter needs and fewer than it gives in nineteen trials out of twenty; the 1002, 16013, 1603 and 6694 are not shown to be. On the 5512, 16 watches fall in the 690,000s alone and 13 in the 660,000s. 12 of the 12 dated 5513s sold with an exclamation-point dial carry numbers in a single block, the 765,000s. 92 of the 156 dated Daytona 6241s sit in just three blocks, at 1.76M, 1.94M and 2.08M. And the Khanjar-signed 1665 and 1680 share one block near 3.57M (8 of 12 and 5 of 8 dated examples), which points to a single order filled from both references together.

0.4M0.5M0.6M0.7M0.8M0.9M1.0M1.1M1.2M1.3M5512 (114)5513 (134)1675 (193)1016 (80)

One watch. Case numbers from 0.4M to 1.3M, roughly 1959 to 1966

The number after each name is the count of watches in its row.

This extends to most references a check that one guide applies to the Daytona. A watch whose number sits in a block where its reference is otherwise never found is not condemned by that, but it has earned a closer look. Across the thirty references there are 1,786 blocks holding three or more watches.

How sure: strong for the pattern. Case numbers alone, again. One caution on the counts: dealers hide the last digits of a number, so the same watch offered twice can be counted as two. A block with only three or four watches may be fewer watches than it looks. Blocks built on auction-house numbers, which are published in full, do not have this problem.

The 5513 went matte before the 5512 and the GMT-Master

The guides date the change from gilt to matte to 1966 or 1967 on all three references. In case numbers, one puts the start of matte dials on the 5513 around 1.6M, and another treats a matte dial in a 1.41M 5512 case as probably not original.

The library shows three different crossing points. On the 5513, matte dials are 14 of 26 (54%) of those read between 1.4M and 1.5M and 18 of 21 (86%) in the next 100,000. On the 5512 and the steel 1675 together they are only 11 of 51 (22%) between 1.4M and 1.5M; the 5512 reaches 6 of 8 (75%) at 1.6M to 1.7M and the 1675 12 of 15 (80%). So the plain Submariner changed about 100,000 case numbers before its chronometer sibling and the GMT-Master. The 1.6M figure fits the 5512 and the 1675. It is too late for the 5513.

Submariner 5513: share of dials that are matte050%100%0.9M to 1.0M: 1 of 8 watches (12%)80.9M1.0M to 1.1M: 1 of 11 watches (9%)111.0M1.1M to 1.2M: 0 of 9 watches (0%)91.1M1.2M to 1.3M: 3 of 12 watches (25%)121.2M1.3M to 1.4M: 3 of 31 watches (10%)311.3M1.4M to 1.5M: 14 of 26 watches (54%)261.4M1.5M to 1.6M: 18 of 21 watches (86%)211.5M1.6M to 1.7M: 20 of 21 watches (95%)211.6M1.7M to 1.8M: 16 of 17 watches (94%)171.7M1.8M to 1.9M: 6 of 6 watches (100%)61.8M1.9M to 2.0M: 12 of 12 watches (100%)121.9M2.0MwatchesOne guide: change begins about hereSubmariner 5512050%100%0.9M to 1.0M: 0 of 1 watches (0%)10.9M1.0M to 1.1M: 0 of 4 watches (0%)41.0M1.1M to 1.2M: 0 of 1 watches (0%)11.1M1.2M to 1.3M: 1 of 10 watches (10%)101.2M1.3M to 1.4M: 0 of 7 watches (0%)71.3M1.4M to 1.5M: 1 of 10 watches (10%)101.4M1.5M to 1.6M: 6 of 13 watches (46%)131.5M1.6M to 1.7M: 6 of 8 watches (75%)81.6M1.7M to 1.8M: 11 of 11 watches (100%)111.7M1.8M to 1.9M: 14 of 14 watches (100%)141.8M1.9M2.0MwatchesOne guide: a 5512 here should still be giltGMT-Master 1675, steel050%100%0.9M to 1.0M: 0 of 2 watches (0%)20.9M1.0M to 1.1M: 1 of 14 watches (7%)141.0M1.1M to 1.2M: 4 of 14 watches (29%)141.1M1.2M to 1.3M: 4 of 23 watches (17%)231.2M1.3M to 1.4M: 4 of 28 watches (14%)281.3M1.4M to 1.5M: 10 of 41 watches (24%)411.4M1.5M to 1.6M: 4 of 10 watches (40%)101.5M1.6M to 1.7M: 12 of 15 watches (80%)151.6M1.7M to 1.8M: 21 of 25 watches (84%)251.7M1.8M to 1.9M: 26 of 28 watches (93%)281.8M1.9M to 2.0M: 19 of 26 watches (73%)261.9M2.0Mwatches

Matte, as a share of dials read as matte or giltA guide's figure

Along the foot are case numbers in millions. The figure under each bar is the number of dated watches in that stretch. A stub on the baseline means watches were found there and none has the feature; a blank means no watch was found.

The numbers behind the chart
Case numbersSubmariner 5513: share of dials that are matteSubmariner 5512GMT-Master 1675, steel
0.9M to 1.0M1 of 80 of 10 of 2
1.0M to 1.1M1 of 110 of 41 of 14
1.1M to 1.2M0 of 90 of 14 of 14
1.2M to 1.3M3 of 121 of 104 of 23
1.3M to 1.4M3 of 310 of 74 of 28
1.4M to 1.5M14 of 261 of 1010 of 41
1.5M to 1.6M18 of 216 of 134 of 10
1.6M to 1.7M20 of 216 of 812 of 15
1.7M to 1.8M16 of 1711 of 1121 of 25
1.8M to 1.9M6 of 614 of 1426 of 28
1.9M to 2.0M12 of 12none seen19 of 26

Matte dials out of all dials read as matte or gilt from photographs, steel watches only.

Gilt print on 5513s numbered 1.40M to 1.50M

5513 · 1.41M Tropical Watch
5513 · 1.46M Bob's Watches
5513 · 1.47M Tropical Watch
5513 · 1.48M Antiquorum

White print on matte, metres first, in the same stretch

5513 · 1.42M Tropical Watch
5513 · 1.42M Tropical Watch
5513 · 1.45M Craft + Tailored
5513 · 1.47M Tropical Watch
5513 · 1.48M Phillips
5513 · 1.49M Tropical Watch

How sure: fair. The dials were read from photographs by an automated reading that was right 96% of the time on watches held back for testing. Each block holds only tens of watches, so the crossing points are good to about 100,000 numbers. A replacement dial reads the same as an original one, which is why the first columns of each chart are not at zero. The 46 dials read on 5513s between 1.36M and 1.56M were also looked over by eye: the matte ones are metres-first dials of the earliest matte type, apart from three later replacements. Two cautions: 32 of the 47 5513 dials read between 1.4M and 1.6M come from one seller's listings, and a watch listed twice with its last digits hidden is counted twice, so the lead is real but rests on fewer watches than the chart suggests.

Gloss dials arrive 100,000 to 200,000 numbers before the guides place them

The guides give case numbers for two of the three references. One has matte dials on the 16750 up to about 8.4M. Another has matte dials on the 16660 up to about 8.5M.

The library finds gloss dials earlier on all three, and at the same point. From 8.0M to 8.2M, 3 of 39 (8%) dials read are gloss. From 8.2M to 8.3M it is 19 of 47 (40%), and from 8.3M to 8.5M 24 of 34 (71%). On the Sea-Dweller alone, 7 of 8 (88%) dials between 8.2M and 8.3M are already gloss. Since almost nothing was numbered between 7.66M and 8.0M, that is early in the first run of cases after the gap. Matte dials do not vanish: 11 of 50 dials from 8.5M to 9.0M are still matte, whether from old stock or later service.

GMT-Master 16750: share of dials that are gloss050%100%8.0M to 8.1M: 0 of 7 watches (0%)78.0M8.1M to 8.2M: 1 of 7 watches (14%)78.1M8.2M to 8.3M: 7 of 19 watches (37%)198.2M8.3M to 8.4M: 6 of 8 watches (75%)88.3M8.4M to 8.5M: 2 of 4 watches (50%)48.4M8.5M to 8.6M: 3 of 5 watches (60%)58.5M8.6M to 8.7M: 12 of 13 watches (92%)138.6M8.7M to 8.8M: 2 of 2 watches (100%)28.7M8.8M8.9M to 9.0M: 2 of 2 watches (100%)28.9M9.0MwatchesOne guide: matte up to about hereSubmariner Date 16800050%100%8.0M to 8.1M: 0 of 9 watches (0%)98.0M8.1M to 8.2M: 0 of 4 watches (0%)48.1M8.2M to 8.3M: 5 of 20 watches (25%)208.2M8.3M to 8.4M: 3 of 4 watches (75%)48.3M8.4M to 8.5M: 8 of 13 watches (62%)138.4M8.5M to 8.6M: 2 of 3 watches (67%)38.5M8.6M to 8.7M: 1 of 3 watches (33%)38.6M8.7M to 8.8M: 4 of 5 watches (80%)58.7M8.8M to 8.9M: 5 of 7 watches (71%)78.8M8.9M to 9.0M: 2 of 4 watches (50%)48.9M9.0MwatchesSea-Dweller 16660050%100%8.0M to 8.1M: 1 of 6 watches (17%)68.0M8.1M to 8.2M: 1 of 6 watches (17%)68.1M8.2M to 8.3M: 7 of 8 watches (88%)88.2M8.3M to 8.4M: 3 of 3 watches (100%)38.3M8.4M to 8.5M: 2 of 2 watches (100%)28.4M8.5M to 8.6M: 2 of 2 watches (100%)28.5M8.6M8.7M to 8.8M: 2 of 2 watches (100%)28.7M8.8M8.9M to 9.0M: 2 of 2 watches (100%)28.9M9.0MwatchesOne guide: matte up to about here

Gloss, as a share of dials read as gloss or matteFewer than five watchesA guide's figure

Along the foot are case numbers in millions. The figure under each bar is the number of dated watches in that stretch. An outlined bar rests on fewer than five watches, so its height means little. A stub on the baseline means watches were found there and none has the feature; a blank means no watch was found.

The numbers behind the chart
Case numbersGMT-Master 16750: share of dials that are glossSubmariner Date 16800Sea-Dweller 16660
8.0M to 8.1M0 of 70 of 91 of 6
8.1M to 8.2M1 of 70 of 41 of 6
8.2M to 8.3M7 of 195 of 207 of 8
8.3M to 8.4M6 of 83 of 43 of 3
8.4M to 8.5M2 of 48 of 132 of 2
8.5M to 8.6M3 of 52 of 32 of 2
8.6M to 8.7M12 of 131 of 3none seen
8.7M to 8.8M2 of 24 of 52 of 2
8.8M to 8.9Mnone seen5 of 7none seen
8.9M to 9.0M2 of 22 of 42 of 2

Gloss dials out of all dials read as gloss or matte from photographs.

Metal-ringed markers of gloss dials on cases numbered 8.20M to 8.30M

16750 · 8.25M Craft + Tailored
16750 · 8.25M Craft + Tailored
16800 · 8.26M SwissWatchExpo
16800 · 8.27M Bob's Watches
16660 · 8.30M Tropical Watch
16660 · 8.30M Tropical Watch
16660 · 8.30M Antiquorum

Painted markers of matte dials in the same stretch

16750 · 8.23M Tropical Watch
16800 · 8.24M SwissWatchExpo
16750 · 8.25M Craft + Tailored
16800 · 8.25M Craft + Tailored
16800 · 8.29M Bob's Watches
16750 · 8.29M Tropical Watch

How sure: fair. Read from photographs, with the same 96% test result. The markers of all 44 photographed dials between 8.2M and 8.3M were also looked at one by one: every dial read as gloss has metal-ringed markers and every dial read as matte has painted ones. The counts for any one reference are small, so the claim is about the three together: gloss is common from 8.2M and in the majority from 8.3M. A gloss dial is also the usual service replacement for a matte one, so some of the early gloss dials will be later fittings: before the gap, where every dial began matte, 6 of 109 (6%) are read as gloss.

Launch years and the first watches do not always match

The guides give one year for when a reference began and one for when it ended, usually the year it was announced or dropped from the catalogue.

The library can ask a different question: what are the lowest and highest case numbers actually found? For most of the thirty references the answer sits within a year or two of the guides. For four it does not.

ReferenceThe guidesThe library
Sea-Dweller 16660Introduced 1978 (6 sources)120 dated watches. The lowest number is 6.55M, which the charts place in 1981. Only 3 lie below 7.0M; the rest begin at 7.12M (1982). Two guides already put the first 16660 dials at about 5.9M; nothing that low has been found here.
Milgauss 1019Introduced 1960 (11 sources)117 dated watches. 1 lies below 1.6M, at 1.07M (1964); every other one starts from the 1.62M batch (1967).
Explorer 1016Introduced 1960 (4 sources), 1963 (7 sources)341 dated watches. 16 with numbers published in full lie below 0.80M, which is 1962 or earlier on the charts used for dating (36 counting dealers' half-hidden numbers, some of which will be the same watches again). The sources giving 1963 cannot be describing the start. 2 of the full numbers lie below 0.45M, earlier than any launch year the guides give; those want checking before they are believed.
Oyster Perpetual 1002Ended 1969 (2 sources), 1979 (2 sources), 1980 (2 sources), 1984 (1 source)147 dated watches. 5 carry numbers above 7.5M, from 1983 on, and a further 9 carry the letter prefixes of 1987 to 1990 (4 with an L prefix and 5 with an R prefix). It ran to the end of the four-digit era like the 5500 and 6694.

How sure: strong as far as it goes. Each row counts watches with published case numbers, and every listing behind the surprising ones was read to confirm the reference. Finding late watches proves a late end. Not finding early ones is weaker: a hundred or so watches with none from the first years makes those years rare at best, and does not show that nothing was made.

How each old reference handed over to the new one

The guides give years, and the years overlap or leave gaps depending on the site. For the Sea-Dweller they agree the two references ran side by side.

The library can put the handovers on one scale, and they fall into three groups. The dress watches went first and all at one point: the last 1803s, 1601s and 1603s are near 5.25M, 5.27M and 5.35M, and the first 18038s, 16014s and 16030s near 5.23M, 5.26M and 5.56M. A handful of 18038s and 16013s are earlier still, in the thin run at 4.7M among the unused stretches. The GMT-Master and the Submariner Date followed about a million numbers later: the last 1675s are near 5.99M and the first 16750s near 6.22M; the last 1680s near 6.19M and the first 16800s near 6.46M. The Explorer II changed last, 8.31M to 8.40M. The Sea-Dweller is the exception the guides describe, and the library measures it: 16660s are found from 6.55M (3 watches below 7.0M) and regularly from 7.12M, while 1665s run on to 8.06M: about 0.9 million numbers of overlap, or 1.5 million counting the first few.

4M5M6M7M8M9M10MGMT-Master1675: seen up to 5.99M16750: seen from 6.22M1675 to 5.99M16750 from 6.22MSubmariner Date1680: seen up to 6.19M16800: seen from 6.46M1680 to 6.19M16800 from 6.46MSea-Dweller1665: seen up to 8.06M16660: seen from 6.55M1665 to 8.06M16660 from 6.55MExplorer II1655: seen up to 8.31M16550: seen from 8.40M1655 to 8.31M16550 from 8.40MDay-Date1803: seen up to 5.25M18038: seen from 5.23M1803 to 5.25M18038 from 5.23MDatejust, white gold bezel1601: seen up to 5.27M16014: seen from 5.26M1601 to 5.27M16014 from 5.26MDatejust, steel1603: seen up to 5.35M16030: seen from 5.56M1603 to 5.35M16030 from 5.56M

Old referenceNew referenceStretches the thirty references skip

How sure: strong. Case numbers only. To keep a stray misreading from setting an end point, each figure is the number that 99 in 100 watches of the reference lie below, or above, not the extreme.

Where the tropical dials are

The guides know which dials turn brown. One says a very small number of early 1680 dials did, usually Mark II or III, which it places at 2.2M to 2.45M. Another ties the brown Mark II Sea-Dweller dial to 1.7M cases. None gives a rate.

The library gives the rate, and among watches that come to market it is not small. Between 2.2M and 2.5M, 60 of 115 (52%) of the 1680s are sold as tropical. Across the rest of the reference up to 4.3M it is 25 of 535 (5%). The same holds elsewhere: each reference has one stretch, sometimes two, where a brown dial is common, and outside it brown dials are scarce. The Sea-Dweller 1665 has a second: 11 of 13 (85%) from 2.5M to 3.0M.

Submariner Date 1680: share of watches sold as tropical050%100%2.0M to 2.1M: 1 of 2 watches (50%)22.0M2.1M to 2.2M: 4 of 26 watches (15%)262.1M2.2M to 2.3M: 36 of 67 watches (54%)672.2M2.3M to 2.4M: 17 of 26 watches (65%)262.3M2.4M to 2.5M: 7 of 22 watches (32%)222.4M2.5M to 2.6M: 1 of 31 watches (3%)312.5M2.6M to 2.7M: 0 of 22 watches (0%)222.6M2.7M to 2.8M: 0 of 20 watches (0%)202.7M2.8M to 2.9M: 5 of 45 watches (11%)452.8M2.9M to 3.0M: 1 of 41 watches (2%)412.9M3.0M to 3.1M: 1 of 62 watches (2%)623.0M3.1M to 3.2M: 1 of 19 watches (5%)193.1M3.2M to 3.3M: 0 of 43 watches (0%)433.2M3.3M to 3.4M: 3 of 52 watches (6%)523.3M3.4M to 3.5M: 1 of 25 watches (4%)253.4M3.5M to 3.6M: 1 of 8 watches (12%)83.5M3.6M to 3.7M: 0 of 9 watches (0%)93.6M3.7M to 3.8M: 0 of 37 watches (0%)373.7M3.8M to 3.9M: 1 of 26 watches (4%)263.8M3.9M to 4.0M: 0 of 31 watches (0%)313.9M4.0M to 4.1M: 2 of 16 watches (12%)164.0M4.1M to 4.2M: 1 of 4 watches (25%)44.1M4.2M to 4.3M: 2 of 16 watches (12%)164.2M4.3MwatchesOne guide: Mark II and III dials, 2.2M to 2.45M

Sold as tropical, as a share of all dated 1680sA guide's figureFewer than five watches

Along the foot are case numbers in millions. The figure under each bar is the number of dated watches in that stretch. An outlined bar rests on fewer than five watches, so its height means little. A stub on the baseline means watches were found there and none has the feature; a blank means no watch was found.

ReferenceStretchSold as tropical thereEverywhere else
Sea-Dweller 16651.50M to 2.00M17 of 32 (53%)20 of 398 (5%)
Explorer 10161.00M to 1.50M27 of 65 (42%)13 of 246 (5%)
Submariner 55131.00M to 1.50M35 of 183 (19%)15 of 885 (2%)
GMT-Master 1675, gilt era0.50M to 0.75M9 of 52 (17%)71 of 1,280 (6%)
GMT-Master 1675, early matte1.75M to 2.25M25 of 190 (13%)55 of 1,142 (5%)
Daytona 6263 and 62652.75M to 3.50M39 of 143 (27%)37 of 692 (5%)

1680 dials sold as tropical, cases numbered 2.2M to 2.5M

1680 · 2.23M Tropical Watch
1680 · 2.24M Phillips
1680 · 2.25M Hodinkee
1680 · 2.29M Phillips
1680 · 2.36M Tropical Watch
1680 · 2.36M Tropical Watch
1680 · 2.36M Tropical Watch
1680 · 2.36M Phillips

How sure: fair. Tropical is a seller's word and it raises a price, so it is used freely, and brown dials are more likely to be put up for sale than black ones. Both would lift the rate everywhere. Neither would pile the brown dials into one stretch of case numbers and leave the neighbouring stretches almost clear. A listing that only names a dealer whose own name holds the word is not counted.

Cracked gloss dials have an address

The guides agree that the first gloss dials of the mid-1980s were prone to fine cracks. They give years, not case numbers, and no rate.

The library puts numbers on it. Between 8.2M and 8.8M, 53 of 400 (13%) of the watches from these seven references are sold as spider dials; before the gap, on matte dials, it is 1 of 281 (under 1%). After 9.0M the two families part. On the sapphire-crystal references the fault is over: 5 of 200 (2%) from 9.0M up, and none of 91 with an R prefix. On the 5513 it carries on: 5 of 31 (16%) from 9.0M up and 4 of 40 (10%) with an R prefix, then none of 45 with an L. That would mean the 5513 went on cracking for two or three years after its siblings had stopped, but it rests on 9 watches, 7 of them from one seller, so it is a lead and no more.

Gloss-era sports references: share of watches sold with a cracked dial025%25%8.0M to 8.1M: 0 of 45 watches (0%)458.0M8.1M to 8.2M: 1 of 36 watches (3%)368.1M8.2M to 8.3M: 6 of 100 watches (6%)1008.2M8.3M to 8.4M: 18 of 79 watches (23%)798.3M8.4M to 8.5M: 3 of 89 watches (3%)898.4M8.5M to 8.6M: 3 of 33 watches (9%)338.5M8.6M to 8.7M: 11 of 57 watches (19%)578.6M8.7M to 8.8M: 12 of 42 watches (29%)428.7M8.8M to 8.9M: 2 of 17 watches (12%)178.8M8.9M to 9.0M: 2 of 31 watches (6%)318.9M9.0M to 9.1M: 1 of 30 watches (3%)309.0M9.1M to 9.2M: 1 of 32 watches (3%)329.1M9.2M to 9.3M: 0 of 2 watches (0%)29.2M9.3M to 9.4M: 0 of 42 watches (0%)429.3M9.4M to 9.5M: 2 of 55 watches (4%)559.4M9.5M to 9.6M: 0 of 14 watches (0%)149.5M9.6M to 9.7M: 0 of 33 watches (0%)339.6M9.7M to 9.8M: 6 of 21 watches (29%)219.7M9.8M to 9.9M: 0 of 2 watches (0%)29.8M9.9Mwatches

Sold as a spider dial, as a share of all dated watches (5513, 16750, 16753, 16800, 16660, 16550, 16760)

Along the foot are case numbers in millions. The figure under each bar is the number of dated watches in that stretch. A stub on the baseline means watches were found there and none has the feature; a blank means no watch was found. The scale runs from 0 to 50%, not to 100%.

How sure: fair. From sellers' descriptions. A cracked dial lowers a price, so it is more likely to go unmentioned than to be invented, and the true rates are probably higher.

Sigma dials: a window in case numbers, and not on the five-digit references

The guides call the sigma dial a thing of the 1970s, or of about 1970 to the end of the decade. For the Daytona, one places the first at about 3.0M.

The library finds one window for the dress watches and the Daytona alike, and it opens at about 3.0M, which the charts place in 1971 or 1972. Below 2.75M the four-digit dress references are sold as sigma 2% of the time, which is noise, though the last stretch before it is already 12 of 310 (4%). At 3.0M to 3.25M it is 30 of 285 (11%), peaking at 75 of 285 (26%) between 3.25M and 3.5M and fading to 21 of 340 (6%) by 5.0M to 5.5M. The Daytona runs at 101 of 258 (39%) from 3.0M to 4.25M. No dated five-digit watch carries it: none of 1,987, including 102 watches numbered 4.5M to 5.5M, where the last four-digit watches still do. (2 five-digit watches with no dated case number are sold with one.)

Four-digit Datejust, Day-Date, Date and Oyster Perpetual: share sold as sigma dials050%100%2.00M to 2.25M: 5 of 275 watches (2%)2752.00M2.25M to 2.50M: 5 of 239 watches (2%)2392.50M to 2.75M: 12 of 310 watches (4%)3102.50M2.75M to 3.00M: 6 of 275 watches (2%)2753.00M to 3.25M: 30 of 285 watches (11%)2853.00M3.25M to 3.50M: 75 of 285 watches (26%)2853.50M to 3.75M: 51 of 230 watches (22%)2303.50M3.75M to 4.00M: 36 of 228 watches (16%)2284.00M to 4.25M: 10 of 160 watches (6%)1604.00M4.25M to 4.50M: 0 of 16 watches (0%)164.50M to 4.75M: 1 of 8 watches (12%)84.50M4.75M to 5.00M: 0 of 3 watches (0%)35.00M to 5.25M: 15 of 252 watches (6%)2525.00M5.25M to 5.50M: 6 of 88 watches (7%)885.50M to 5.75M: 3 of 45 watches (7%)455.50M5.75M to 6.00M: 0 of 23 watches (0%)236.00MwatchesDaytona 6263 and 6265050%100%2.00M to 2.25M: 0 of 27 watches (0%)272.00M2.25M to 2.50M: 0 of 10 watches (0%)102.50M to 2.75M: 0 of 16 watches (0%)162.50M2.75M to 3.00M: 3 of 76 watches (4%)763.00M to 3.25M: 8 of 37 watches (22%)373.00M3.25M to 3.50M: 11 of 30 watches (37%)303.50M to 3.75M: 23 of 59 watches (39%)593.50M3.75M to 4.00M: 25 of 61 watches (41%)614.00M to 4.25M: 34 of 71 watches (48%)714.00M4.25M to 4.50M: 0 of 2 watches (0%)24.50M4.75M to 5.00M: 0 of 1 watches (0%)15.00M to 5.25M: 5 of 25 watches (20%)255.00M5.25M to 5.50M: 3 of 19 watches (16%)195.50M to 5.75M: 4 of 85 watches (5%)855.50M5.75M to 6.00M: 0 of 20 watches (0%)206.00MwatchesOne guide: first sigma Daytona dials

Sold as a sigma dial, as a share of all dated watchesA guide's figure

Along the foot are case numbers in millions. The figure under each bar is the number of dated watches in that stretch. A stub on the baseline means watches were found there and none has the feature; a blank means no watch was found.

How sure: fair. From sellers' descriptions. The sigma marks are small and easy to leave out of a listing, so the rates are floors. The edges of the window and the empty five-digit column do not depend on the rate. Auction houses hardly use the word, so this is a finding about dealers' listings.

The second-generation movements arrived at the same point in every reference

The guides date the move from the first 1500-series calibres to the faster-beating second ones differently for each reference, and often differently from each other: anywhere from 1963 to 1971.

The library shows it happening in one place. Across the six, half the sightings of the older calibre lie below 0.85M to 1.17M, and nine in ten sightings of the newer one lie above 1.35M to 1.65M. On the three sports references the change falls between about 1.1M and 1.4M; the Datejust and Day-Date are noisier and centre on the same stretch. On the charts 1.2M is 1965 or 1966. The 1963 or 1964 given for the 5513 is too early: between 1.0M and 1.2M sellers name the 1530 on 29 watches and the 1520 on 1.

ReferenceCalibresOlder calibre named onNewer calibre named onThe guides' year
Submariner 55131530 to 152061 watches; half below 1.01M233 watches; nine-tenths above 1.41M1963 or 1964 by one guide; two to three years after launch by another
Air-King 55001530 to 152031 watches; half below 0.85M79 watches; nine-tenths above 1.35Mno year given
Submariner 55121560 to 157018 watches; half below 1.17M29 watches; nine-tenths above 1.36Mno year given
Day-Date 18031555 to 155643 watches; half below 1.01M255 watches; nine-tenths above 1.65M1965 by one guide; 1967 by another
Datejust 1601 and 16031560/1565 to 1570/157558 watches; half below 1.07M382 watches; nine-tenths above 1.55M1965 by one guide; 1971 by another
GMT-Master 16751560/1565 to 1570/157590 watches; half below 0.88M330 watches; nine-tenths above 1.38Mabout 1965

How sure: fair. From sellers' descriptions. Auction houses open the case and report what is inside; dealers sometimes copy a calibre from a specification sheet, which is why a few old calibres are named on late cases. The middle of each range is more trustworthy than its ends.

Stella dials turn up on Day-Date cases of the later 1960s

The guides mostly date the lacquered Stella dials to the 1970s; one has them from the very end of the 1960s.

The library has 8 of 56 (14%) of the 1803s numbered 1.50M to 1.75M titled as Stella dials, counting a watch offered twice once, 8 of them with the number published in full. They sit at 1.52M, 1.61M, 1.62M, 1.64M, 1.73M, 1.74M and 1.75M, which the charts place from 1966 to 1968. Below 1.5M it is 1 of 151 (1%) and from 1.75M to 2.5M 5 of 159 (3%); the steady supply the guides describe begins after that and reaches 41 of 241 (17%) from 3.5M on.

Day-Date 1803: share of watches sold with a Stella dial025%25%0.50M to 0.75M: 0 of 26 watches (0%)260.50M0.75M to 1.00M: 0 of 39 watches (0%)391.00M to 1.25M: 1 of 39 watches (3%)391.00M1.25M to 1.50M: 0 of 47 watches (0%)471.50M to 1.75M: 8 of 56 watches (14%)561.50M1.75M to 2.00M: 0 of 64 watches (0%)642.00M to 2.25M: 3 of 68 watches (4%)682.00M2.25M to 2.50M: 2 of 27 watches (7%)272.50M to 2.75M: 2 of 58 watches (3%)582.50M2.75M to 3.00M: 7 of 77 watches (9%)773.00M to 3.25M: 5 of 71 watches (7%)713.00M3.25M to 3.50M: 6 of 63 watches (10%)633.50M to 3.75M: 7 of 58 watches (12%)583.50M3.75M to 4.00M: 5 of 41 watches (12%)414.00M to 4.25M: 11 of 50 watches (22%)504.00M4.25M to 4.50M: 0 of 2 watches (0%)24.50M5.00M to 5.25M: 16 of 85 watches (19%)855.00M5.25M to 5.50M: 2 of 5 watches (40%)55.50Mwatches

Titled as a Stella dial, as a share of all dated 1803s

Along the foot are case numbers in millions. The figure under each bar is the number of dated watches in that stretch. A stub on the baseline means watches were found there and none has the feature; a blank means no watch was found. The scale runs from 0 to 50%, not to 100%.

Day-Date 1803s sold with Stella dials, cases numbered 1.50M to 1.75M

1803 · 1.52M Phillips
1803 · 1.61M Phillips
1803 · 1.73M Sotheby's

How sure: tentative. A lacquered dial can be fitted to an older watch, and a Stella dial adds a great deal to a price, so this is a lead and not a finding. What speaks for it is the shape: dials added later would be spread across old cases of every age, not gathered in one stretch with almost none on either side. What would settle it is a dated guarantee or an early photograph, which the library does not have.

About one gilt-era sports watch in twelve no longer has a gilt dial

The guides warn that service dials are common and explain how to spot one. None says how common.

The library can count, with care. Taking only cases numbered before the matte dials came in, 27 of 227 (12%) dials are read from photographs as matte: 7 of 53 (13%) on the 5513, 2 of 58 (3%) on the 5512 and 18 of 116 (16%) on the steel 1675. Looked at one by one, 7 of those are gloss dials with gilt print that the automated reading misread, 1 is a gold watch filed as steel. That leaves 19 of 226; 2 of those cannot be told from the photograph, and 2 dials read as gilt are sold as service dials, so the count is good to about two either way. A matte dial in one of these cases is a replacement, and about one such watch in twelve on the market has one.

How sure: fair, after a check by eye. Read from photographs, then every dial read as matte was looked at and its listing read, because the automated reading's own error, about four in a hundred, is a large part of a figure this small. Watches offered for sale are also not a fair sample of watches that exist: dealers prefer the ones with their first dial. Sellers themselves state a service dial on 17 of 497 (3%) dated steel watches of these years, which is a floor. For the GMT-Master the premise is less safe than for the Submariners: one account speaks of early matte dials on cases numbered below 1.4M.

Exotic dials stop by 3.5 million on the screw-pusher Daytonas

The guides date the first exotic dials to 1963 or 1965 and say they sold poorly when new. One puts them on the 6263 from about 2.08M to the early 3M numbers, roughly 1969 to 1971. None gives a rate along the way.

The library agrees on where they are common and shows where they stop. Among 6263s and 6265s sold by auction houses, 59 of 111 (53%) numbered from 2.0M to 3.0M are titled Paul Newman. Just above 3.0M, up to 3.06M, it is 8 of 27 (30%), which is the early 3M the guide gives. From 3.06M to 3.5M it is 3 of 22 (14%), the highest at 3.49M, about 1973 on the charts. From 3.5M on it is 1 of 425 (under 1%) (a gold 6265 at 5.04M). On the pump-pusher 6239 and 6241 the name is on 5 of 93 (5%) below 1.5M, 25 of 66 (38%) from 1.5M to 1.75M and 202 of 246 (82%) from 1.75M. So an exotic dial in a 6263 or 6265 case numbered up to the early 3M is in keeping, a few are found up to 3.5M, and one in a later case has to be explained.

Daytona 6263 and 6265 sold at auction: share titled Paul Newman050%100%2.00M to 2.25M: 24 of 25 watches (96%)252.00M2.25M to 2.50M: 2 of 10 watches (20%)102.50M to 2.75M: 8 of 14 watches (57%)142.50M2.75M to 3.00M: 25 of 62 watches (40%)623.00M to 3.25M: 8 of 29 watches (28%)293.00M3.25M to 3.50M: 3 of 20 watches (15%)203.50M to 3.75M: 0 of 40 watches (0%)403.50M3.75M to 4.00M: 0 of 44 watches (0%)444.00M to 4.25M: 0 of 48 watches (0%)484.00M4.25M to 4.50M: 0 of 2 watches (0%)24.50M5.00M to 5.25M: 1 of 25 watches (4%)255.00M5.25M to 5.50M: 0 of 11 watches (0%)115.50M to 5.75M: 0 of 48 watches (0%)485.50M5.75M to 6.00M: 0 of 14 watches (0%)146.00M to 6.25M: 0 of 52 watches (0%)526.00M6.25M to 6.50M: 0 of 22 watches (0%)226.50M to 6.75M: 0 of 7 watches (0%)76.50M7.00M7.25M to 7.50M: 0 of 23 watches (0%)237.50M to 7.75M: 0 of 13 watches (0%)137.50M8.00M to 8.25M: 0 of 1 watches (0%)18.00M8.50M to 8.75M: 0 of 20 watches (0%)208.50M8.75M to 9.00M: 0 of 20 watches (0%)209.00M to 9.25M: 0 of 17 watches (0%)179.00M9.25M to 9.50M: 0 of 18 watches (0%)189.50MwatchesOne guide: exotic dials to about here

Titled Paul Newman, as a share of the dated watches auction houses have soldA guide's figure

Along the foot are case numbers in millions. The figure under each bar is the number of dated watches in that stretch. A stub on the baseline means watches were found there and none has the feature; a blank means no watch was found.

The numbers behind the chart
Case numbersDaytona 6263 and 6265 sold at auction: share titled Paul Newman
2.00M to 2.25M24 of 25
2.25M to 2.50M2 of 10
2.50M to 2.75M8 of 14
2.75M to 3.00M25 of 62
3.00M to 3.25M8 of 29
3.25M to 3.50M3 of 20
3.50M to 3.75M0 of 40
3.75M to 4.00M0 of 44
4.00M to 4.25M0 of 48
4.25M to 4.50M0 of 2
4.50M to 4.75Mnone seen
4.75M to 5.00Mnone seen
5.00M to 5.25M1 of 25
5.25M to 5.50M0 of 11
5.50M to 5.75M0 of 48
5.75M to 6.00M0 of 14
6.00M to 6.25M0 of 52
6.25M to 6.50M0 of 22
6.50M to 6.75M0 of 7
6.75M to 7.00Mnone seen
7.00M to 7.25Mnone seen
7.25M to 7.50M0 of 23
7.50M to 7.75M0 of 13
7.75M to 8.00Mnone seen
8.00M to 8.25M0 of 1
8.25M to 8.50Mnone seen
8.50M to 8.75M0 of 20
8.75M to 9.00M0 of 20
9.00M to 9.25M0 of 17
9.25M to 9.50M0 of 18

Daytona 6263 and 6265 with a case number, sold by an auction house; the name read from the lot title.

How sure: strong for where they end, weak for the share. Sellers' words, but only the titles of auction houses, which name an exotic dial when there is one. The shares say nothing about how many were made: an exotic Daytona is far more likely to be sent to auction than a plain one. Where they end does not depend on that. Hundreds of later 6263s and 6265s have been sold by the same houses, and almost none is titled Paul Newman. A dial can also be moved into another case, which is the usual explanation offered for a late one. That a few run later than the early 3M numbers rests on a handful of watches.

The Daytona, the Milgauss and the Explorer II waited for their buyers

The guides tell this in anecdotes: the exotic dials sold poorly, and one steel 6263 is said to have sat with a retailer for nine years. None puts a figure on it.

The library can, for the watches sold with a dated guarantee. Of 68 manual-wind Daytonas whose case number and papers are both dated, 27 (40%) were sold three or more years after the case was numbered. The Milgauss 1019 waited longest: all 7 of 7, 5 of them by ten years or more. The Explorer II 1655 is next, at 5 of 13. For the Submariners, GMT-Masters, Datejusts, Day-Dates and the rest it is 14 of 144 (10%), and the usual gap is a year or less. A guarantee dated several years after the case is therefore ordinary on these three and worth a second look on a Submariner.

ReferenceWatches with dated papersSold within a year of the case numberTwo years afterThree or more
Daytona 6239, 6241, 6263 and 626568281327 (40%)
Milgauss 10197007 (100%)
Explorer II 165513445 (38%)
Submariners and Sea-Dwellers 5512, 5513, 1680, 1665, 16800 and 166605033116 (12%)
GMT-Masters 1675, 16750, 16753 and 16760403037 (18%)
All other references544491 (2%)

How sure: fair. It rests on two dates: the year the charts give a case number, and the year sellers read off the guarantee. The charts disagree by a year, so only a gap of three years or more is counted as a wait. The numbers are small, papers can belong to another watch, and a gap of ten years or more is as likely to be a later set of papers as a slow sale. A watch that sat for years in a shop is also more likely to have kept its papers. The difference between these three and the rest is too large to be any of these.

Military and COMEX orders came in batches; the Tiffany 1680s mark a period

The guides date the Royal Navy Submariners to 1972 to 1976, to the 1970s, or to the mid-1970s and early 1980s, and put them at about 1,200 pieces. Of the Tiffany-signed 1680 they say only that it exists. None gives a run of case numbers for either.

The library finds the military and COMEX orders bunched in their own numbers. 45 of the 49 dated 5513s sold as military issue are numbered 3.75M to 4.0M, where they are 45 of the 75 dated 5513s in that stretch, and they sit in three groups there: 10 at 3.76M, 8 at 3.82M and 27 at 3.92M, about 1974 and 1975. 3 more lie near 2.9M, about 1972, which is when the guides start them. 17 of the 23 dated 1680s sold with a Tiffany signature lie between 5.0M and 5.5M, about 1978, though they are found from 3.27M; those 17 are spread over 10 different 10,000-number blocks, so they mark the years in which the retailer took the reference, not one delivery. 7 of the 10 5513s that auction houses and dealers title as COMEX watches share one block at 2.83M. For the military and COMEX watches this is the same habit as the Khanjar order above: a customer's watches were cased together, a delivery at a time.

2.5M2.6M2.7M2.8M2.9M3.0M3.1M3.2M3.3M3.4M3.5M3.6M3.7M3.8M3.9M4.0M4.1M4.2M4.3M4.4M4.5M4.6M4.7M4.8M4.9M5.0M5.1M5.2M5.3M5.4M5.5M5.6M5.7M5.8M5.9M6.0M5513, military issue (49)1680, Tiffany signature (23)5513, COMEX in the title (10)

One watch. Case numbers from 2.5M to 6.0M, roughly 1970 to 1979

The number after each name is the count of watches in its row.

How sure: fair. Sellers' words, and a military or retailer-signed dial is exactly what is worth faking, so some of these labels will be wrong. Case numbers are as sellers give them, often with the last digits hidden, so a dealer's listing can repeat a watch an auction house has already sold.

Dial colours have dates: grey and blue from 1.5 million, and silver leaves the Day-Date

The guides list the colours a reference came in and call some of them uncommon. None gives a share, and none says when a colour began or ended.

The library can. On the four-digit Datejust and Date, none of 127 dials from 0.5M to 1.5M are grey and 4 of 132 (3%) below 1.5M are blue (all 4 of those from 1.0M on); from 1.5M to 4.0M grey and blue together are 158 of 642 (25%). The only grey dials read below 1.5M are 2 in cases numbered below 0.5M, and at least one of those is a dial of the 1970s in an older case. On the gold Day-Date the silver dial goes the other way: 26 of 64 (41%) of dials from 0.5M to 1.5M, 17 of 104 (16%) from 3.0M to 5.0M and 15 of 483 (3%) from 5.0M, nearly all of those 18038s. On the two-tone 16013 the champagne dial is 33 of 69 (48%) from 5.0M to 7.0M and 101 of 147 (69%) from 9.0M; the steel 16014 and 16030 almost never have one (1 of 314 (under 1%)).

Datejust 1601 and 1603, Date 1500: share of dials that are grey or blue050%100%0.5M to 1.0M: 0 of 60 watches (0%)600.5M1.0M to 1.5M: 4 of 67 watches (6%)671.5M to 2.0M: 19 of 105 watches (18%)1051.5M2.0M to 2.5M: 29 of 124 watches (23%)1242.5M to 3.0M: 46 of 150 watches (31%)1502.5M3.0M to 3.5M: 38 of 162 watches (23%)1623.5M to 4.0M: 26 of 101 watches (26%)1013.5M4.0M to 4.5M: 10 of 34 watches (29%)344.5M to 5.0M: 1 of 3 watches (33%)34.5M5.0M to 5.5M: 11 of 61 watches (18%)615.5M to 6.0M: 0 of 10 watches (0%)105.5M6.5M7.5M8.5M9.5M10.0MwatchesDay-Date 1803 and 18038: share of dials that are silver050%100%0.5M to 1.0M: 14 of 29 watches (48%)290.5M1.0M to 1.5M: 12 of 35 watches (34%)351.5M to 2.0M: 18 of 59 watches (31%)591.5M2.0M to 2.5M: 10 of 37 watches (27%)372.5M to 3.0M: 13 of 56 watches (23%)562.5M3.0M to 3.5M: 6 of 49 watches (12%)493.5M to 4.0M: 9 of 38 watches (24%)383.5M4.0M to 4.5M: 2 of 16 watches (12%)164.5M to 5.0M: 0 of 1 watches (0%)14.5M5.0M to 5.5M: 1 of 62 watches (2%)625.5M to 6.0M: 1 of 47 watches (2%)475.5M6.0M to 6.5M: 1 of 67 watches (1%)676.5M to 7.0M: 0 of 30 watches (0%)306.5M7.0M to 7.5M: 2 of 35 watches (6%)357.5M to 8.0M: 0 of 8 watches (0%)87.5M8.0M to 8.5M: 3 of 36 watches (8%)368.5M to 9.0M: 0 of 45 watches (0%)458.5M9.0M to 9.5M: 3 of 88 watches (3%)889.5M to 10.0M: 4 of 65 watches (6%)659.5M10.0Mwatches

Share of the dials whose colour was read from a photographFewer than five watches

Along the foot are case numbers in millions. The figure under each bar is the number of dated watches in that stretch. An outlined bar rests on fewer than five watches, so its height means little. A stub on the baseline means watches were found there and none has the feature; a blank means no watch was found.

The numbers behind the chart
Case numbersDatejust 1601 and 1603, Date 1500: share of dials that are grey or blueDay-Date 1803 and 18038: share of dials that are silver
0.5M to 1.0M0 of 6014 of 29
1.0M to 1.5M4 of 6712 of 35
1.5M to 2.0M19 of 10518 of 59
2.0M to 2.5M29 of 12410 of 37
2.5M to 3.0M46 of 15013 of 56
3.0M to 3.5M38 of 1626 of 49
3.5M to 4.0M26 of 1019 of 38
4.0M to 4.5M10 of 342 of 16
4.5M to 5.0M1 of 30 of 1
5.0M to 5.5M11 of 611 of 62
5.5M to 6.0M0 of 101 of 47
6.0M to 6.5Mnone seen1 of 67
6.5M to 7.0Mnone seen0 of 30
7.0M to 7.5Mnone seen2 of 35
7.5M to 8.0Mnone seen0 of 8
8.0M to 8.5Mnone seen3 of 36
8.5M to 9.0Mnone seen0 of 45
9.0M to 9.5Mnone seen3 of 88
9.5M to 10.0Mnone seen4 of 65

Dial colour read from photographs of dated watches; each block is half a million case numbers.

How sure: good. Read from photographs by an automated reading of colour, which was right at least 96 times in 100 on held-back watches. Its mistakes would add grey and blue dials where there were none, so the empty early blocks are firmer than the shares. A dial can be changed or refinished in another colour, which works the same way.

Wide Boy, linen and tapestry dials: when each is actually found

The guides name these dials and show dated examples. One gives linen as about 1962 to 1992 and tapestry as the late 1970s to about 2000. None gives a first or a last for the Wide Boy, and none gives a rate.

The library finds three different shapes. The Wide Boy is none of 607 from 0.5M to 1.5M, 96 of 2,377 (4%) from 1.5M to 4.0M and 10 of 582 (2%) from 4.0M to 5.5M, and no dated five-digit watch is sold with one (none of 1,989): a dial of roughly 1967 to the end of the four-digit references, about 1978. Linen is 12 of 1,007 (1%) from 0.5M to 2.0M, rises to 132 of 1,560 (8%) from 2.0M to 3.5M and stays at 137 of 2,456 (6%) from 5.0M to the end, so it becomes common about 1968, some years after the guide's start, and stays. Tapestry belongs to the five-digit references: on them it is 4 of 666 (1%) below 7.0M, 4 of 209 (2%) from 7.0M to 8.0M and 56 of 1,114 (5%) from 8.0M, so it is common from the mid-1980s, a few years after the guide's start.

Datejust, Date and Day-Date: share sold as Wide Boy dials010%20%0.5M to 1.0M: 0 of 262 watches (0%)2620.5M1.0M to 1.5M: 0 of 345 watches (0%)3451.5M to 2.0M: 8 of 400 watches (2%)4001.5M2.0M to 2.5M: 21 of 473 watches (4%)4732.5M to 3.0M: 27 of 549 watches (5%)5492.5M3.0M to 3.5M: 25 of 538 watches (5%)5383.5M to 4.0M: 15 of 417 watches (4%)4173.5M4.0M to 4.5M: 4 of 163 watches (2%)1634.5M to 5.0M: 0 of 17 watches (0%)174.5M5.0M to 5.5M: 6 of 402 watches (1%)4025.5M to 6.0M: 0 of 267 watches (0%)2675.5M6.0M to 6.5M: 0 of 253 watches (0%)2536.5M to 7.0M: 0 of 132 watches (0%)1326.5M7.0M to 7.5M: 0 of 180 watches (0%)1807.5M to 8.0M: 0 of 52 watches (0%)527.5M8.0M to 8.5M: 0 of 266 watches (0%)2668.5M to 9.0M: 0 of 323 watches (0%)3238.5M9.0M to 9.5M: 0 of 338 watches (0%)3389.5M to 10.0M: 0 of 243 watches (0%)2439.5M10.0MwatchesSold as linen dials010%20%0.5M to 1.0M: 1 of 262 watches (0%)2620.5M1.0M to 1.5M: 4 of 345 watches (1%)3451.5M to 2.0M: 7 of 400 watches (2%)4001.5M2.0M to 2.5M: 32 of 473 watches (7%)4732.5M to 3.0M: 35 of 549 watches (6%)5492.5M3.0M to 3.5M: 65 of 538 watches (12%)5383.5M to 4.0M: 18 of 417 watches (4%)4173.5M4.0M to 4.5M: 4 of 163 watches (2%)1634.5M to 5.0M: 1 of 17 watches (6%)174.5M5.0M to 5.5M: 23 of 402 watches (6%)4025.5M to 6.0M: 14 of 267 watches (5%)2675.5M6.0M to 6.5M: 22 of 253 watches (9%)2536.5M to 7.0M: 6 of 132 watches (5%)1326.5M7.0M to 7.5M: 9 of 180 watches (5%)1807.5M to 8.0M: 3 of 52 watches (6%)527.5M8.0M to 8.5M: 22 of 266 watches (8%)2668.5M to 9.0M: 20 of 323 watches (6%)3238.5M9.0M to 9.5M: 10 of 338 watches (3%)3389.5M to 10.0M: 8 of 243 watches (3%)2439.5M10.0MwatchesSold as tapestry dials010%20%0.5M to 1.0M: 1 of 262 watches (0%)2620.5M1.0M to 1.5M: 0 of 345 watches (0%)3451.5M to 2.0M: 0 of 400 watches (0%)4001.5M2.0M to 2.5M: 0 of 473 watches (0%)4732.5M to 3.0M: 0 of 549 watches (0%)5492.5M3.0M to 3.5M: 0 of 538 watches (0%)5383.5M to 4.0M: 0 of 417 watches (0%)4173.5M4.0M to 4.5M: 0 of 163 watches (0%)1634.5M to 5.0M: 1 of 17 watches (6%)174.5M5.0M to 5.5M: 0 of 402 watches (0%)4025.5M to 6.0M: 1 of 267 watches (0%)2675.5M6.0M to 6.5M: 2 of 253 watches (1%)2536.5M to 7.0M: 0 of 132 watches (0%)1326.5M7.0M to 7.5M: 3 of 180 watches (2%)1807.5M to 8.0M: 1 of 52 watches (2%)527.5M8.0M to 8.5M: 10 of 266 watches (4%)2668.5M to 9.0M: 20 of 323 watches (6%)3238.5M9.0M to 9.5M: 14 of 338 watches (4%)3389.5M to 10.0M: 12 of 243 watches (5%)2439.5M10.0Mwatches

Sold under the name, as a share of all dated watches of the nine references

Along the foot are case numbers in millions. The figure under each bar is the number of dated watches in that stretch. A stub on the baseline means watches were found there and none has the feature; a blank means no watch was found. The scale runs from 0 to 20%, not to 100%.

The numbers behind the chart
Case numbersDatejust, Date and Day-Date: share sold as Wide Boy dialsSold as linen dialsSold as tapestry dials
0.5M to 1.0M0 of 2621 of 2621 of 262
1.0M to 1.5M0 of 3454 of 3450 of 345
1.5M to 2.0M8 of 4007 of 4000 of 400
2.0M to 2.5M21 of 47332 of 4730 of 473
2.5M to 3.0M27 of 54935 of 5490 of 549
3.0M to 3.5M25 of 53865 of 5380 of 538
3.5M to 4.0M15 of 41718 of 4170 of 417
4.0M to 4.5M4 of 1634 of 1630 of 163
4.5M to 5.0M0 of 171 of 171 of 17
5.0M to 5.5M6 of 40223 of 4020 of 402
5.5M to 6.0M0 of 26714 of 2671 of 267
6.0M to 6.5M0 of 25322 of 2532 of 253
6.5M to 7.0M0 of 1326 of 1320 of 132
7.0M to 7.5M0 of 1809 of 1803 of 180
7.5M to 8.0M0 of 523 of 521 of 52
8.0M to 8.5M0 of 26622 of 26610 of 266
8.5M to 9.0M0 of 32320 of 32320 of 323
9.0M to 9.5M0 of 33810 of 33814 of 338
9.5M to 10.0M0 of 2438 of 24312 of 243

Sellers' descriptions of dated watches; each block is half a million case numbers.

How sure: fair. From sellers' descriptions, across nine Datejust, Date and Day-Date references. A seller who does not use the name leaves the watch uncounted, so every rate is a floor. Where a dial becomes common and where it stops does not depend on the rate.

Where the library agrees with the guides

The same method confirms the guides far more often than it corrects them, which is some reason to trust it where it does not.

  • Double Red to Great White. One guide places the last Double Red batch in the 5.2M numbers. Read from photographs, Great White dials are 3 of 112 (3%) below 5.25M, 13 of 22 (59%) from 5.25M to 5.5M and 71 of 75 (95%) after.
  • Red Submariner to white. One guide has the 1680 turn white at about 4.0M. Sellers title 284 of 533 (53%) of 1680s as red up to 3.75M, 31 of 103 (30%) from 3.75M to 4.25M and 2 of 337 (1%) from 5.0M on.
  • The first sigma Daytonas. The 3.0M one guide gives is exactly where they begin in the library.
  • Which 1680 dials turn brown. The stretch where half the 1680s are sold as tropical is the stretch one guide gives for the Mark II and Mark III dials.
  • Pointed crown guards. Guides end them between 1963 and 1966, and the most detailed has them on the GMT-Master to at least 1.14M. Of the watches titled with pointed guards, 44 of 44 5512s, 37 of 37 5513s and 80 of 80 1675s are numbered below 1.3M.
  • The small 24-hour hand. One guide sees it to about 1.8M. Nineteen in twenty of the 1675s sold with it are numbered below 1.69M.
  • The Maxi dials of the 5513. Two guides number them. One puts Maxi I on cases from the early 5.0M, the other on mid-4M to low-5M cases; both have Maxi IV from the late 6M and Maxi V from 7M to 8M. Sold under those names, their middle watches sit at 5.4M, 7.3M and 7.7M, and no Maxi I is numbered below 5.02M: on that point the library sides with the first.
  • Production spans. For most of the thirty references the lowest and highest case numbers found sit within a year or two of the dates most guides give.

How this was done, and its limits

Each watch in the library is matched to a reference, and where its seller publishes a case number, placed in case-number order. Dealers usually hide the last three or four digits, which is close enough for everything here. For every label the library holds, whether read from a photograph or stated by a seller, the share of watches carrying it was counted along that order, reference by reference. Anything that looked sharp was then set against what the guides say on the same point, and the listings or dials behind the surprising cases were read or looked at one by one.

Working in case numbers and not years is deliberate. The published charts disagree with each other by a year or more, and several of the findings above would vanish into that disagreement if stated in years. Where a year is given, it is the year most charts assign to the number.

The limits are real. These are thirty references, not all of Rolex. The watches are those that came up for sale, which favours the rare, the early and the well kept. Sellers' words are loose, and a watch offered by two dealers with its number half hidden can be counted twice. A dial or movement can be changed, and the library sees what is in a case now, not what left the factory. And the guides' side was gathered by machine reading of 49 sites, checked by sampling and not line by line, so a guide may hold a figure the comparison missed.

Ubudashi: a reference library of collectable watches. Photographs remain the property of the holders credited with them. The text is the library's own.

Findings · About · Photographs · Corrections · Terms · Privacy