Accuracy is an interesting thing in the watch enthusiast community. For some, so long as their watches stay relatively close to the actual time, they're happy. For others accuracy is paramount. At times going as far as investing in a home timegrapher to check the beat rate and amplitude of their mechanical watches, it never ceases to amaze me how far down the rabbit hole some venture. For those more interested in the accuracy side of things, there's a lot to digest. Different brands uphold different standards, and only some adhere to said standards through validation by a third party. By now you've all heard terms like COSC certified, Chronometer certified, METAS certified, etc, but how many of you have gone as far as to look at what that testing truly entails? The end product is a watch that carries a guarantee of running accuracy, which is definitely a perk. Running out of spec? Warranty will cover it (unless they try to prove that you dropped/banged/otherwise damaged your watch prior). This goes for internally certified watches just as much as it does for those marketed under 3rd party standards, so as much as some will say that a 3rd party certification makes a difference, that's not entirely true. Rolex, Grand Seiko, and others have the duty to stand behind their accuracy claims just as much as any COSC or METAS certified watch in the market does.

Continuing with our ongoing series, this 4th installment of the Watchuseek Bench Sessions is once again supported by the folks at eBay, though the editorial direction of the piece was once again left in the hands of the Watchuseek editorial team.

Browse eBay Luxury Watches Here

What is COSC?

It makes the most sense to start with the biggest player in the accuracy game-the Contrôle Officiel Suisse des Chronomètres. The organization itself has been around since 1973, and was launched as a non-profit association to provide an independent testing standard for Swiss made watches.

Any watch that claims to be "Chronometer Certified" has earned the stamp of approval from COSC, and that includes a LOT of watches. To give you a sense of scale, from the annual reporting of the organization, we know that 1,641,087 mechanical movements, and 69,710 quartz movements were tested by the organization in 2020. We also know that these numbers are a decrease of 23.3% over the number of movements tested in 2019. Included in this total volume of movements is Rolex's entire catalog (which passes through COSC before internal testing) as well as 470 references from Omega's current offering (all Master Chronometer watches are also COSC certified). Now, neither of those brands publicly publish annual production numbers, but it's pretty safe to say that Rolex, Omega make up at least a significant portion of COSC's workload, but there's also a hefty chunk of the rest of the Swatch Group's portfolio in there, alongside other brands like Breitling, Baume et Mercier, Norqain, and a pretty reasonable slew of other smaller brands as well.

This certification at a surface level is simple-the movement in a mechanical watch with COSC certification cannot deviate in accuracy beyond -4 seconds to +6 seconds per day for movements that are larger than 20mm in diameter, and -5 to +8 for those smaller than 20mm (which are quite rare in the industry). There are different standards for quartz watches as well, but for the sake of this article we're staying focused on mechanical watchmaking. There are also ample other criteria that relate to deviation during testing, which we will touch on a little further down the line.

The biggest consideration when looking at COSC is how this testing is performed. To pass COSC testing, only the completed and running movements are sent to the testing facilities. These are not complete watches/watch heads, only movements in movement holder/cases with a crown fitted for adjustment, a dummy dial, and a running seconds hand to allow for clear measurement. COSC will certify complete watch heads upon the request of the client, but because this is not part of the ISO 3159:2009 certification standards for chronometers, this is seldom ever done. The movements also need to be serialized/be marked with clear serial numbers in order to ensure that the watch movement tested is the watch movement delivered as COSC certified. Now, the gut response of many will be that this testing isn't all it's cracked up to be, as there's so much that can happen to a watch movement after return transport from COSC, as well as during the final casing process, and shipping/transit into the retail network. Those arguments might seem fair at a surface level, they're also rapidly negated by the fact that the whole point of having said certification is to have a promised guarantee of accuracy. If your watch is running out of spec after delivery, it is the retailer and manufacturer's duty to restore it to proper running order without expense to the owner. We've heard stories of pushback from brands on this topic, but any brand worth its salt should honor its promise.

The Testing Process

After being verified and registered in the system, round one involved a full wind of the movements in the dial up position. Those with automatic winding either have their rotors removed or self-winding disengaged in order to be wound by the automated trays in the COSC test center. A 12-hour temperature stabilization comes next, with the movements left running in a room with a fixed temperature of 23.5 degrees Celsius. From there, the 15-day onslaught of tests begins, only interrupted by the meticulous winding of the movements on a daily basis.

Accuracy measurements are taken in a total of 5 positions over the course of a 15-day period. These include dial up, dial down, crown up, crown down, and upright (as if wall mounted like a clock, 12 o'clock up). There are also tests for temperature deviation, testing the movement's accuracy at eight degrees Celsius as well as at 38 degrees Celsius. Because we aren't talking about complete watches, there is no opportunity for testing of water resistance or anti-magnetic properties, but given the sheer volume that COSC handles we aren't especially surprised. Once all of the testing has been completed, it is evaluated against the deviation constraints as laid out by the ISO standards. This includes a maximum thermal variation of -/+ 0.6 seconds, and a maximum difference between rates in horizontal and vertical positions of -6 to +8 seconds. You can read a full breakdown of these testing parameters at the link below, with added descriptions that break down what it all means.

Read The COSC Accuracy Standards Here

It's also worth noting that watch brands are able to turn to COSC for "non-standard" testing procedures as well. Among other options, this can include high and low temperature testing outside of COSC's standard range, should a brand want to clear a set of movements for a particular environment or expedition.

What Is METAS?

Onto the next phase of independent watch testing, the Federal Institute of Metrology (METAS) defines itself as the federal centre of competence for all issues related to measurement and for measuring equipment and measuring procedures. Perhaps the most dry and mundane place on earth, this is also the home of a rather elevated level of watch accuracy testing when compared to COSC. To pass METAS certification, testing is done both on the COSC certified movement, as well as on the complete and assembled watch head.

Aside from a couple of outliers, the majority of the watches to pass through METAS come from Omega. At time of publishing, 470 references in Omega's catalog (which we believe to be almost all references) are METAS certified, meaning they've passed through accuracy tests with a maximum deviation of -0/+5 seconds per day for movements larger than 26mm, -0/+6 seconds per day for movements greater than 20mm up to 26mm, and -0/+7 for movements up to 20mm in diameter. For the sake of clarity, many Omega movements fall into the -0/+5 category, though not all. The 8800 seen above has to hit -0/+6. This movement is fitted in the current generation Seamaster 300m, the 38mm Aqua Terra, and others.

These movements also have to pass magnetic resistance tests to a rating of 15,000 gauss, as well as METAS pressure tests for water resistance. As with COSC, this isn't a "batch testing" process, but rather each watch to get METAS certification has to pass through its own individual litany of testing. Though METAS is an independent body, the testing for Omega is done in the brand's own facilities, which have been built out to suit the testing process. Within the Omega campus, a separate office for METAS staff who sample and monitor the testing procedures. The reason for this setup will be made clear as we talk through the testing process.

The Testing Process

Due to the fact that METAS deals with testing all sorts of instrumentation and measuring devices, navigating their own website to find details of the testing process can prove to be challenging. If you're looking for the really nitty gritty stuff, like how each individual test breaks down, and the mathematical equations involved, you can find the PDF of the Requirements for Certification here. When we're talking math we aren't kidding, as is seen in the excerpt below from the PDF that speaks to measurement uncertainty:

5.4 Measurement uncertainty
The measurement tools and procedures must ensure extended measurement uncertainties for each criterion while satisfying the following equation:

?? ≤ ??/??
?? : Extended measurement uncertainty for the criterion (? = 2)
?? : Tolerance on the criterion
??: Measurement capability index. ?? must be ≥ 5.

Thankfully the Omega website breaks things down into a much more digestible format, showcasing the specifics of testing through all eight of its test protocols. The first of the tests for METAS certification immediately makes it clear why it's more efficient to have a METAS lab in Omega's HQ. You see, the first test is of the movement on its own, whereas the balance of the tests are on a complete/assembled watch head. To run these tests through METAS would involve a fair bit of wasted time and shipping, and given the volume that Omega pushes through testing annually, that would add up quickly.

To the actual test sequence, phase one is magnetic resistance testing of the movement, followed by the same test of a completed watch head. 15,000 gauss is the benchmark to pass, and thanks to the use of silicon within the co-axial escapement, this is easy as pie. These first two tests look and listen for mechanical deviation within the movement, using highly sensitive microphones. From there, the watches are fully wound, and go through an additional two-day test that starts under the same magnetic field, looking at the accuracy of the watch across six positions. On day one, accuracy data is registered under magnetism, and the second day, the watches are demagnetized. These values are compared in order to ensure the movement doesn't experience any deviation caused by the magnetic fields.

The next two rounds of testing are effectively in the same vein, first testing in 6 positions in magnetic fields through varying temperatures (testing at 33 and 23 degrees Celsius), and then tested for accuracy again in a machine that alters the position of the watch on a 60-second cycle over an extended period to further replicate a "daily wear" scenario. The last of the requisite accuracy tests is one that's equally interesting and practical-a test of isochronism. For this test, measurements are taken of the running accuracy of the watch when fully wound, and again when at 33% remaining to its power reserve. As with other test groups, these measurements are done in all six positions. They are then compared in order to record a maximum deviation rate.

Before bringing testing to a close, two more checks must be completed. First, the duration of the movement's power reserve is verified, and then the watch goes off for water resistance testing. During the latter, water resistance tests are performed at 25% greater pressure than the watch is rated for, which is a fairly standard practice for water resistance checks. Unlike COSC certified watches where access to the testing results is at the discretion of the brand in question (some will supply a certificate, and others will not), Omega makes it easy for owners to access the test data for each of their watches through a portal on their website, as well as via their proprietary app by scanning the METAS card supplied with your watch.

Which is Better and Why Should You Care?

At an objective level, it's easy to see that METAS certification is superior here, between its more stringent constraints, additional measured position, and of course magnetic resistance. There's also the fact that the testing takes place at a more complete level, allowing less room for problems to occur during handling and casing. That said, do you need any of this? Are you working in research labs or power generation facilities where you're regularly exposed to significant magnetic fields? Will you notice a difference between a watch running 4 seconds slow or 5 seconds fast per day?

This really all boils down to my opening statement. Some people will revel in the technical specificity involved in these accuracy checks, whereas others will simply write them off as brands trying to flex their marketing superiority. Neither camp is wrong, and no one is being hurt by someone being obsessed about how accurate their watch is. Even if you aren't one to pay attention to these sorts of things, having that added security of knowing that a brand has to live up to their promises of accuracy certainly isn't a bad thing.