Inside COSC

Watchbore investigates the only organization in the world that decides whether a watch is a chronometer or not.

There are three questions that dominate the forums at TimeZone: How good is my watch? Is it really worth all the money I paid for it? and, Am I mad?

Watchbore also gets a lot of email asking the same three questions, and his invariable answers are: It’s as good as you think it is; No, and Probably.

For all its flaws, there is one organization that provides some sort of standard by which to judge a watch. In his constant search for choice trivia with which to thrill his readers into a definitive stupor, Watchbore succeeded in infiltrating the secretive COSC organization, and with the help of his Albanian associates, secured some valuable confessions from the high priests of this obscure cult.

But first, some exciting facts based on documents that were recently handed to Watchbore anonymously by an unknown woman in the street.

They reveal that COSC stands for contrôle officiel suisse des chronometres (official Swiss chronometer inspection). Furthermore, in 2001, COSC’s three laboratories in Geneva, Biel and Le Locle, individually tested 1,315,752 horological movements, almost all for compliance with international chronometer standard ISO 3156 for mechanical wristwatches, and issued 1,255,515 chronometer certifications worth at least USD4.5 million. This is a 23.3% rise on the previous year.

Mechanical Movements (ISO 3156)

TESTED	PASSED	FAILED	% FAILED
1,254,248	1,198,043	56,175	4.5%

Quartz Movements (COSC Standard)

TESTED	PASSED	FAILED	% FAILED
61,504	57,442	4,062	6.6%

The Rolex Factor

These documents, secured at great expense by Watchbore, also rip apart the veil of secrecy surrounding the exact annual production of Rolex mechanical watches.

Out of 77 brands and a handful of watch schools submitting movements for chronometer certification, Rolex is by far the biggest contributor to COSC. It sends almost their entire output of mechanical movements to COSC and in 2001, 761,601 of them were given chronometer certificates — a 20% increase over 2000. "All the mechanical watches Rolex sells are officially certified chronometers," intoned a bimbo in charge of misinformation at the Geneva company headquarters.

The Geneva and Biel laboratories are almost entirely devoted to testing Rolex movements. Interestingly, Geneva, where 96% of movements tested are from Rolex, shows the lowest failure rate at 2.2%. It rises to 4.5% in Biel (86% Rolex) and to 5.7% in Le Locle where virtually no Rolex movements are tested.

Watchbore estimates that at least 15,000 Rolex movements failed in 2001. According to Rolex, the rejects are fixed and sent back to COSC until they pass. "We don’t use COSC to tell us how good our movements are," said a source deep inside the Wilsdorf foundation. "We test them ourselves. All we want is the chronometer certification. It’s for marketing."

Top Six COSC Brands in 2001

BRAND	NO. CERTIFICATES	% TOTAL (Mech.)	REMARKS(MEN = >20MM)
Rolex	761,601	64%	All mechanical + 573 quartz mvmts, men's and women's
Omega	207,879	17.4%	All mechanical, men's
Breitling	142,825	11.4% *	40% quartz
Bulgari	36,380	3%	All mechanical, men's. 70% increase over 2000
Panerai	27,275	2.3%	All mechanical, men's
Tag Heuer	20,650	1.7%	All mechanical, men's

* % of total quartz & mechanical.

The World’s Most Accurate and Precise Movement Revealed

It becomes clear that virtually all mechanical movements gaining COSC chronometer certificates are what the industry calls "tracteurs" — the 28,800 v/h workhorses such as the Rolex 3035, ETA’s 2892 and Valjoux 7750.

But the essential question hanging on the lips of the few readers still struggling to keep awake is: "What is the best-performing caliber of them all?" Watchbore asked the person most likely to know, Mr Jean-Pierre Curchod, former dean of the Geneva Watchmaking School, president of the Swiss Society of Chronometry and director of the Geneva laboratory of COSC.

For an official institution to reveal this information would throw the entire watch industry in disarray. To hold up one movement as superior to all the others would deflate the bubble of myth that sustains the tightly linked cartel of brands. Therefore, according to COSC, all chronometers are equal, and if any are more equal than others, it’s a state secret.

However, Watchbore soon discovered that he and Mr Curchod happened to share an interest in the wines produced by a certain Mr Hutin in Dardagny, and it was in the course of a thorough examination of the relative merits of the Merlot and the Pinot Noir that a source close to Mr Curchod made a startling revelation.

Now, among all the assorted WISes, watch enthusiasts and experts that frequent TimeZone to tell the world what watches they wear, it is extremely unlikely that any owns a watch with the world’s best performing movement. It is equally unlikely that the woman who owns a Rolex automatic Oyster Datejust is aware that the caliber 2235 automatic inside is the most consistently precise and accurate movement tested by COSC.

Even more amazingly, at less than 20mm, the Rolex 2235 falls into the smallest category where the tolerances are at their widest, yet performs well within the tightest allowances reserved for pocket-watches. Almost 200,000 of these movements passed the COSC test in 2001. Does Rolex have a secret? "It’s their immense know-how in construction and manufacturing," says Mr Curchod reverently.

What makes the chronometer? Watchbore put the question to COSC’s managing director, Mr Pierre-Yves Soguel. For the best chance of passing the test, the movement has to be conceived and engineered from the outset as a chronometer, he explains. It’s all in the design, construction and especially in the machining of the parts. The most precise machine tools are only viable in high-volume production, which explains why the mass-produced tracteurs are consistently chronometers.

According to Mr Curchod, good lubrication is also an essential attribute of the COSC chronometer. Free-sprung or indexed balances, Breguet or flat springs, coaxial or lever escapements or tourbillons make little or no difference to performance in the COSC test. "It’s the quality of workmanship throughout the movement rather than any single feature that makes a movement accurate and reliable."

What about those expensive, lovingly handcrafted crafted, pursuit-of-perfection in-house movements? It is possible for such movements to reach chronometer standard, acknowledges Mr Curchod, but at the cost of much expensive and time-consuming tweaking. "It is more difficult and the failure rates are high — as much as 60%."

Other Significant Brands Submitting to COSC

BRAND	NO. CERTIFICATES	REMARKS
Baume & Mercier	10,416	All Men'ss
Chopard	7,776	Incl. 7272 L.U.C calibers, all men's
Zenith	5,556	All men's, incl. 85 pocket & 23 clocks
Vacheron Constantin	3,038	All men's
Ebel	2,837	All men's
Ulysse Nardin	2,253	All men's
Patek Philippe	1,286	All men's
Chronoswiss	198	All men's
Girard-Perregaux	2	Both men's

Audemars Piguet, Jaeger-LeCoultre, Lange, IWC, Breguet and Piaget are among the brands absent from COSC.

What Does COSC Measure and How Good is the Test?

In order to satisfy the insatiable apathy of his readers for the most irrelevant details, Watchbore went in person to the COSC laboratory in Geneva to see for himself how the movements are tested.

"We test the engine and not the car; that is the responsibility of the brand," says Mr Curchod ushering Watchbore into the dust-free, temperature- and moisture-controlled climate of the laboratory.

COSC tests movements at their barest functional level, although brands can enter movements with as many complications as they like. As all the movements are wound by the crown, automatics have to leave their rotors behind because the machine that turns the crown would damage the highly geared winding mechanism. Most of the mechanical watches tested by COSC become automatics.

Each movement is fitted with a COSC standard dial, seconds-hand (sweep or small) and winding crown. Every 24 hours, an electronic camera records the state of the seconds-hand to the nearest tenth of a millisecond compared to the atomic reference clock. The camera shoots twice in succession to check whether the movement has stopped. Then the movement is rewound and returned to the appropriate position and temperature for the next 24-hour period. This goes on for 16 consecutive periods.

For the first 11 periods, the movements spend at least 48 hours in each of five positions at a constant 23°C. The readings indicate both how accurate and how precise the movement is. Using a shooting analogy, accuracy is how close you are to the target. Precision is a tight grouping of shots, which may be off target. Thus, a watch that gains 15 seconds a day might not be accurate, but if it gains (or loses) exactly the same amount every day, it is extremely precise. High precision can be adjusted to accuracy, but low precision indicates inherent faults such as an inconsistent power supply, probably due to defects in the going train.

By analyzing the rate variation between different positions, the COSC test can diagnose a badly poised balance, too much oil, or a need to review the profile and roundness of the pivots.

The next three test periods determine how much the rate varies between three different temperatures — 8°, 23° and 38°C. Excessive changes in the rate could show that the balance-spring alloy is not up to standard.

For the last two days, the movement resumes its original position and temperature. Comparing the readings here with the first two days’ results shows to what extent the test itself has affected the performance of the movement.

Rolex has a special machine to test its vast quantities of movements. These are loaded into magazines like bullets. The machine extracts the movement, reads it, winds it and returns it to the magazine. Non-Rolex movements are placed in recesses on trays. In an adjoining room, large cupboards hold batches of watches in various positions at different temperatures.

The COSC testing program is divided into 16, 24-hour periods. The watch is rated in five positions and at three temperatures.

Period (day)	Temperature	Position
0	23°C	6 o'clock up/ 12 o'clock up (pocket-watches)
1	23°C	6 o'clock up/ 12 o'clock up
2	23°C	6 o'clock up/ 12 o'clock up
3	23°C	3 o'clock up
4	23°C	3 o'clock up
5	23°C	9 o'clock up
6	23°C	9 o'clock up
7	23°C	Dial down
8	23°C	Dial down
9	23°C	Dial up
10	23°C	Dial up. Chronographs run 24 hours
11	8°C	Dial up
12	23°C	Dial up
13	38°C	Dial up
14	23°C	6 o'clock up/ 12 o'clock up
15	23°C	6 o'clock up/ 12 o'clock up

CRITERIA	>20mm DIAMETER	<=20mm DIAMETER	POCKET WATCHES
Mean daily rate: during the first 10 days.	-4+6 secs/day	-5+8 secs/day	-2+5 secs/day
Mean rate variation: average of the 5 absolute variations in 5 positions during the first 10 days of test.	2 secs/day	3.4 secs/day	1.5 secs/day
Maximum rate variation: in five positions during the first 10 days of test.	5 secs/day	7 secs/day	2.5 secs/day
Maximum difference in rate between vertical and horizontal positions: mean rate of days 1 and 2 minus mean rate of days 9 and 10.	-6+8 secs/day	-8+10 secs/day	±4 secs/day
Greatest rate difference: between one of the first 10 daily rates and the average daily rate for the test.	10 secs/day	15 secs/day	7 secs/day
Rate variation according to temperature: the rate at 38°c minus the rate at 8°c divided by the temperature difference.	±0.6 secs/day °c	±0.7 secs/day °c	±0.35 secs/day°c
Secondary error: the difference between the average rate at 38°c and at 8°c (days 11 and 13) and the average rate with the dial up at 23° (days 9 and 10).	not applicable	not applicable	±4.5 secs/day
Rate resumption: the final rate minus the average rate of the first two days	±5 secs/day	±6 secs/day	±2.5 secs/day

Why COSC Can’t Raise the Standards

Just about every working watchmaker has a Witschi machine that listens to the watch and gives the performance of the movement in real time. It is on these real-time readings that most mechanical watches are adjusted. But COSC still rates wristwatches according to the static and temperature tests established for pocket-watches in the late 19^th century. The ISO 3159 chronometer standard, developed entirely by the Swiss watch industry, is more than a quarter of a century old; 95% of movements submitted pass the COSC test. Isn’t it time to raise the standard and update the test?

Alas, says Mr Soguel, COSC’s hands are tied, and to change the status quo would be extremely dangerous.

The reason is that Britain, France and Germany have dropped out of the ISO committee on chronograph standards, leaving Switzerland in a minority against India, Japan and China. If Switzerland tried to change ISO 3159, these three countries could sabotage the Swiss watch industry by imposing a standard that mechanical movements would find impossible to meet. That would mean the end of mechanical chronometers, of COSC and of one of the Swiss watch industry’s main marketing planks.

Breitling Dominates Quartz

Nevertheless, COSC has managed to modernize the test for quartz chronometers. The old test gave chronometer certification to watches that came free with the breakfast flakes, and COSC wanted to give the latest thermo-compensated quartz movements the same status as mechanical chronometers. The new 20-day environmental test for the certification of quartz-crystal chronometers subjects the cased-up watches to a total of 7800 100G shocks from six directions, continuously changing positions, magnetic fields, extreme temperature and humidity, and expects them to keep time within ⁷/_100thsof a second a day.

Quartz Movements at COSC, 2001

BRAND	NO. MOVEMENTS
Breitling	54,140
Krieger (Miami Beach)	1,956
Rolex	573
Radiant (Barcelona)	526
Ventura	238
Sector	9

Why COSC Doesn’t Grade Watches According to Performance

The objective assessment and testing of civilian watches started in the railway age when confidence in the timekeeping qualities of your watch became paramount. Observatories and laboratories in major cities rated timepieces. Manufacturers competed for prizes. Customers paid premiums for high-rated watches.

COSC differs in one important respect from all previous watch testing institutions and observatories. It is strictly non-competitive. There are no points awarded or any prizes. There are no degrees of success or honorable mentions. The watches either pass or fail.

This was the one condition demanded by the Swiss watch industry when COSC was founded in 1973. Until that time, there were two institutions in Switzerland that issued rating certificates to watches. The observatories rated prepared timepieces, held competitions and awarded prizes. Local testing laboratories in seven watch making towns issued rating certificates to time-of-day watches. These were grouped into an association called ABDO. ABDO rating certificates gave commendations such as "especially good" to deserving movements. Ninety percent of the watches submitted to ABDO laboratories were from three brands — Rolex, Omega and Mido.

In 1972, an important delegation of Swiss watch manufacturers went to see Mr René Meylan, then industry minister in the Neuchâtel cantonal government. They demanded the end of the observatory competitions. The reason: the Japanese had swept the board in the last two events. Mr Meylan replied that he thought that the whole point of the competitions was for the best to win. The brands then threatened to boycott the contests. Meylan gave in. The observatory competitions were suspended and never revived.

At the same time Rolex, Omega and Mido started to dismember ABDO. By selectively boycotting one or other of the seven testing laboratories they caused each to grant increasing discounts and favors until the organization collapsed.

Mr Soguel says COSC does not compile or publish comparative results because there is no demand for it from the brands.

He compares the COSC certificate to a university degree. “It certifies that you have reached a certain standard, but it does not guarantee that you can still pass the test 20 years hence. And when you frame your diploma on your office wall, you don’t mention the marks you got.”

Were COSC to introduce any sort of ranking by test results, Swiss watchmakers would be forced to compete on the intrinsic qualities of their watches and the whole value hierarchy of Swiss watches would be overturned.

Is COSC Really Independent?

COSC rose from the ashes of ABDO in 1973 as an association of the five watch making cantons of Bern, Geneva, Neuchâtel, Solothurn and Vaud. This government membership was intended to give COSC official independence, but the association is controlled by its general assembly of government and industry representatives. Although the governments have a majority of one, the quorum rules enable a majority of the brands if any government delegates fail to attend.

Mr Soguel declares that the main aim of COSC is to defend its chronometer certificate as a label of excellence, and that maintaining COSC’s total independence from the watch industry is key to the defense of the chronometer. His strategy is uncompromising integrity in the tests. Since he took over as managing director in 1997, COSC has invested heavily in developing its measuring systems and in complying with standards governing testing procedures and environment. The Swiss Federal Office of Metrology has also accredited the COSC laboratories. "I am aware that COSC is a monopoly and of the danger that implies," says Mr Soguel. "But I cannot endanger the credibility of COSC with any lapse from absolute rigor."

An increasing number of manufacturers are submitting their movements for chronometer certificates, mostly minor brands with a handful of pieces. Rolex, which accounted for 80% to 90% of the COSC chronometers, has now seen its share drop to 64%. "Rolex has shown a very strong desire that COSC remains totally independent," observes Mr Soguel.

After much consideration, Watchbore must reach the conclusion that COSC is either an independent institution or a marketing tool for Swiss brands, but it cannot be both. Even though COSC is a monopoly, it is unable to raise chronometer standards and thus the standards of Swiss watch making. Even though it is a government association, COSC cannot or will not publish the results of its tests. Three brands provide 90% of its turnover. COSC has to be dependent on their goodwill.

And Finally, What is a Chronometer?

With the bottom of his bottle of Mr Hutin’s best Gamay in sight, Watchbore feels that it is time to bring this tiresome stream of drivel to its welcome end by getting to the point. Namely, what is a chronometer?

Watchbore’s Oxford dictionary says it’s an instrument for measuring time specifically applied to time-keepers having a special escapement and a compensation balance, used for determining longitude at sea, and for other exact observation. His Webster’s concurs with: "an instrument measuring the passage of time with great accuracy, esp. one used in navigation for determining position."

However, COSC’s very survival depends on the definition enshrined in the international standard — a "precision watch, rated in different temperatures and positions and which has obtained an official rating certificate." Since COSC is the world’s only body that provides "official" certificates (as a government organization), it vehemently defends this definition. Any manufacturer claiming its watches are chronometers without having a COSC certificate is open to prosecution for "unfair competition."

The only watchmaker Watchbore knows who flouts this rule is F. P. Journe. "They can take me to court if they like, but they don’t own the definition of a common word."

As Watchbore put away his notebook and prepared to leave the Geneva COSC laboratory, Mr Curchod took him to one side. "I want to show you a real chronometer," he whispered, taking a metal box from a drawer. The movement inside had Russian markings and all the attributes of a chronometer — compensation balance free-sprung on a helical spring, pivoting-detent escapement, fusée and chain. Russia is the only country that still makes marine chronometers for finding longitude at sea. (Girard-Perregaux markets them under the John Harrison name.)

Does COSC issue chronometer certificates to genuine marine chronometers? Yes, but only in Le Locle where they are rated as "table clocks".