Chronometer

A chronometer is a watch that has passed an official precision test and holds a certificate to prove it. In Swiss practice the testing body is COSC, the Contrôle Officiel Suisse des Chronomètres, which tests movements individually against the criteria of ISO 3159 for wristwatches with a spring balance oscillator. The word is a protected claim in Switzerland, not a marketing adjective: a watch dial that says chronometer should be backed by a certificate with a serial number.

COSC is an independent Swiss association founded in 1973, running laboratories in the watchmaking cantons, and it certifies well over a million movements a year. Certification is per movement, not per model, so two watches of the same reference carry different certificate numbers and different measured rates. The older sense of the word survives alongside this one: for two centuries a chronometer was a marine timekeeper accurate enough to find longitude at sea, and the modern certificate descends from the observatory trials that graded those instruments.

What the test involves

Mechanical movements are tested over 15 days, in 5 positions and at 3 temperatures. The headline criterion is mean daily rate, which must fall between minus 4 and plus 6 seconds per day. Six further criteria apply alongside it, covering consistency across positions, days and temperatures, and they are listed in full below. A movement that averages beautifully but behaves erratically from position to position will still fail.

The positions are dial up, dial down and three vertical crown orientations. The temperatures are 8, 23 and 38 degrees Celsius, which is roughly 46, 73 and 100 Fahrenheit. The movement is wound daily and its rate read every 24 hours, and all seven criteria must pass together.

Those figures apply to movements larger than 20 mm across; smaller ones are allowed wider tolerances. The consistency criteria are what fails movements, and they are why a certificate is worth more day to day than the average alone suggests.

Testing is done on the bare movement, before casing. Some manufacturers go further with their own standards applied to the finished watch, and METAS Master Chronometer certification, used by Omega, tests a cased watch that has already passed COSC.

Standards that go further

METAS, the Swiss federal metrology institute, certifies the finished watch to a mean daily rate of 0 to plus 5 seconds a day and to function in a field of 15,000 gauss, and it also checks water resistance and power reserve. Rolex tests its cased watches after COSC and publishes a Superlative Chronometer figure of minus 2 to plus 2 seconds a day. Grand Seiko and Patek Philippe run their own published standards on assembled watches. The weakness of house standards is structural rather than technical: they are administered by the party with an interest in the result. Our comparison of COSC, METAS and brand standards sets out what each measures.

Common misconceptions

A chronometer is not a chronograph. A chronograph is a stopwatch function; a chronometer is a precision rating. The two words are constantly swapped, including by retailers, and they describe entirely unrelated properties of a watch.

Certification is also not a guarantee of what you will see on the wrist. The certificate describes an uncased movement under controlled laboratory conditions. Once the movement is cased, worn, magnetized or partly wound, real-world results vary. Equally, plenty of uncertified watches keep chronometer-grade time. Certification tells you the maker paid to have the movement measured against a defined standard, which is useful, but it is not the only route to good timekeeping.

Two smaller confusions. Chronometer-grade, chronometer-tested and adjusted to five positions are manufacturer phrases, not certifications, and none comes with a numbered certificate. And the fee COSC charges per movement is small: the real cost is the extra regulating time and the movements that fail, which is why certified references sit a few hundred dollars above uncertified siblings.

If your chronometer drifts

Measure before you worry. One day is a sample, not a rate: compare the watch with an internet time source each morning for a week and average the result, as our guide to timing your own watch describes. If the mean is out of specification and the watch used to be fine, suspect magnetism first, since it costs nothing to rule out and usually makes a watch gain minutes rather than seconds. If the rate is genuinely off, a watch inside warranty should go back to the maker; outside it, regulation on a timing machine runs an indicative US $60 to $150 before sales tax.