Do I need a chronometer?

The word chronometer on a dial means one specific thing in Switzerland: the movement inside passed the test defined in ISO 3159 and administered by the Contrôle Officiel Suisse des Chronomètres. Fifteen days, five positions, three temperatures, seven criteria, headed by a mean daily rate of minus 4 to plus 6 seconds a day. Each movement is tested individually and gets its own numbered certificate. Our answer on what COSC certified means works through the procedure.

So the certificate is not marketing. The question is a different one: given what it measures, and what it costs, is it something you should be buying?

What the certificate actually buys

Three things, and they are worth naming precisely.

• Independent measurement of your individual movement. Not a sample, not a batch average. That movement ran in a laboratory for two weeks and the results were recorded. Almost nothing else on a specification sheet has that status.
• Proof of consistency, not just of rate. The headline number gets quoted, but the criteria that constrain variation between positions and across temperatures are the demanding ones. A movement that averages plus 1 second a day by swinging wildly between positions fails.
• Evidence of how the movement was finished. Passing reliably means better regulation, often a better hairspring, tighter tolerances and more time on the bench. The certificate is a proxy for that care.

COSC publishes the scale it works at: about 2.1 million movements certified in 2025, and roughly 57 million since 1973. Certification is a mass industrial process, not a bespoke one, which is why it is affordable enough to appear on watches costing a few hundred dollars.

What it does not buy

Four honest limits, and they matter more than the marketing.

It tests a movement, not a watch. The test happens before the dial, hands and case are fitted. Casing changes things: hands can add friction, the rotor behaves differently inside a case, and the watch then meets magnetic fields, knocks and a wearer whose daily mixture of positions no laboratory reproduces. A certified movement in a finished watch is an excellent starting point and not a measured outcome.

It says nothing about longevity. There is no durability element in the test at all. Water resistance, shock resistance, magnetic resistance, corrosion and service life are covered by separate standards or by nothing. A chronometer is not built to last longer than an equivalent uncertified movement, and it is not exempt from servicing.

It is a snapshot with a date on it. The certificate describes the movement in the weeks before it was cased. Oils thicken, the mainspring settles, magnetism creeps in, and rate drifts. A ten-year-old unserviced chronometer running at plus 15 seconds a day is a normal watch, not a failed certificate.

It is not the only route to the same result. A movement can be adjusted to chronometer performance and never submitted, which is common: certification costs money and takes time, and a brand may prefer to regulate carefully and declare its own figure. That is a self-declared claim rather than an independent one, but it is measured on the finished watch, which the COSC figure is not.

What an uncertified movement achieves

This is the comparison that decides the question, and manufacturers publish enough for you to make it.

Two things follow. First, the gap between a good uncertified movement and a certified one is smaller than the gap between a cheap movement and a good one. Second, published tolerances are worst-case limits, not expected performance: plenty of standard-grade automatics settle within a few seconds a day once run in, which is why our answer on good accuracy for a mechanical watch starts with measuring rather than assuming.

The premium, and what it is for

Where a brand offers the same model in certified and uncertified form, the retail difference is usually modest, often in the low hundreds of dollars before sales tax, and figures move with model and market, so treat any number as indicative and check the price where you are buying. The laboratory fee itself is a small part of that. Most of the difference is the better-regulated movement, sometimes better components, and the price a brand can charge for a word on the dial.

Set that premium against what else it could buy: a service fund, a better bracelet, or simply a different watch. There is no wrong answer, but there is an unexamined one.

When a chronometer is worth buying

There are good reasons to want one.

• You want the engineering, not the timekeeping. Certification is a reasonable proxy for how carefully a movement was assembled and adjusted, and wanting that is legitimate on its own terms.
• The model you want is certified anyway. Several brands certify their entire mechanical output, so the decision never arises: you are choosing a watch, and the certificate comes with it.
• You need the specification for a real reason. Some professional and navigational uses genuinely benefit from a documented rate, and some buyers simply want the tightest achievable figure without going to quartz.
• Magnetism is a daily problem for you. In that case the useful certification is METAS rather than COSC, because it tests magnetic resistance on the cased watch.

When it is not

Skip it if the watch will be one of several in rotation. A watch that sits in a drawer half the week will lose or gain more from being unwound and stationary than from any difference between grades. Skip it if you are buying your first mechanical watch and the premium would push you past your budget: fit, legibility and the right size matter more day to day, as our guide to buying a first mechanical watch argues.

Skip it, too, if you expect the certificate to hold indefinitely. It will not, and the fix is maintenance rather than certification. A watch running consistently fast or slow can usually be brought back by a watchmaker in an hour, which our answer on regulating a watch explains, and which costs a fraction of the certified premium.

The honest answer

For most owners, no. Buy the watch you want and check what it actually does: wear it for a week, compare it against a synced clock each morning, and run the numbers through our accuracy calculator, using the method in timing your own watch. Most people discover their uncertified watch is running within a handful of seconds a day, which is all the certificate would have promised.

Buy certified when the model you want carries it, when you value the evidence of careful regulation for its own sake, or when magnetic resistance on the cased watch solves a real problem, in which case look at METAS rather than COSC. The full landscape of accuracy standards, in-house figures and how to read them sits in our guide to accuracy and chronometer standards. What none of them can tell you is what your watch will do next Tuesday. Only your own measurements do that.