Watch Accuracy by Brand: Rolex, Omega, Seiko & More

A modern Rolex is specified to run within ±2 seconds per day. An Omega Master Chronometer must stay between 0 and +5 seconds per day. A COSC-certified chronometer is allowed −4/+6, and a mainline Seiko automatic with a 4R movement is within spec anywhere from −35 to +45 seconds per day. In other words, the "right" accuracy for your watch is whatever its manufacturer publishes for that movement — and those numbers vary by a factor of more than twenty across brands.

Why accuracy specs differ so much between brands

A mechanical watch does not have one fixed rate. It runs slightly differently dial-up than crown-down, faster or slower depending on how much the mainspring is wound, and it shifts with temperature. An accuracy spec is a promise about the average daily rate under defined test conditions — not a guarantee that every single day lands inside the range.

What separates a ±2 s/day watch from a +45/−35 s/day watch is mostly labor and materials. Tighter specs require regulating the movement in more positions, better-poised balance wheels, more stable hairspring alloys (silicon or other antimagnetic materials at the high end), and a certification process that verifies all of it. Each step costs money, which is why accuracy promises track price tiers so closely. For a grounding in what counts as normal for mechanical watches in general, see how accurate a mechanical watch should be.

Published accuracy specs by brand and movement

The table below sticks to figures the brands and movement makers actually publish. Where a published number varies by source or by model, the range is noted.

Two details matter when reading it. First, these are average daily rates over a multi-day test — a certified watch can have individual days outside the range and still pass. Second, the standards do not all test the same thing: COSC certifies the bare movement before it goes into a case, while Rolex's Superlative test and METAS certification test the fully assembled watch. A cased test is the stricter promise, because casing, dial and hands can shift a movement's rate.

Rolex: Superlative Chronometer

Since 2015, every Rolex has carried the brand's in-house Superlative Chronometer certification: −2/+2 seconds per day, tested by Rolex after the movement is cased up. The movement inside has already passed COSC (−4/+6), so the Superlative standard is a second, stricter pass. Rolex models from before 2015 are held to the COSC figure, not ±2.

Omega: Master Chronometer via METAS

Omega's Master Chronometer watches are certified by METAS, the Swiss Federal Institute of Metrology, to a rate of 0/+5 seconds per day — notable because the watch is never allowed to lose time. The METAS process also tests the cased watch in multiple positions and winding states and verifies resistance to magnetic fields of 15,000 gauss. Omega models without Master Chronometer certification are generally COSC chronometers at −4/+6.

COSC-only brands: Tudor, Breitling and many others

COSC (Contrôle Officiel Suisse des Chronomètres) is the independent Swiss body behind the word "chronometer" on most dials. Its mechanical standard allows an average daily rate of −4/+6 seconds per day, measured over roughly two weeks in five positions and at three temperatures — on the uncased movement. Tudor and Breitling certify essentially their whole mechanical lineups with COSC; many other brands certify selected models.

Grand Seiko vs mainline Seiko

These are different worlds under one parent company. Grand Seiko's mechanical standard is +5/−3 seconds per day, verified in-house over a 17-day static test in six positions and three temperatures; Grand Seiko also quotes a looser "normal usage" figure (around +10/−1 s/day) because wrist conditions are harsher than a test bench. Mainline Seiko is far more conservative: per Seiko's own published ranges, 4R and NH-class movements (the ones in most Seiko 5 Sports and entry divers, and in countless microbrands) are commonly spec'd around +45/−35 seconds per day, and mid-tier 6R movements around +25/−15. The older 7S26 is wider still. These are deliberately cautious worst-case ranges — most examples run meaningfully better.

Microbrands: Sellita, ETA and Miyota

Most microbrands buy movements, so the spec belongs to the movement maker. A standard-grade Sellita SW200-1 or ETA 2824-2 is typically quoted around ±12 seconds per day average rate, with elaborated and top grades regulated tighter (roughly ±7 and ±4 s/day; exact figures vary slightly by source and production year). Miyota publishes −10/+30 s/day for its 9-series movements and −20/+40 s/day for the 8 series. If a microbrand advertises its own tighter number, that reflects extra regulation on top of the base movement.

Spec vs reality: what owners actually measure

Published specs are floor-level promises, not predictions. In practice, most watches beat their spec — an NH35 rated +45/−35 frequently settles within ±15 seconds per day, and plenty of COSC movements run within a second or two. The spread cuts the other way too: your wrist is not a test bench. Your mix of positions, how much you move, how long the watch sits overnight, and ambient temperature all pull the real-world rate away from the certificate number.

Certification is also a snapshot, not a lifetime guarantee. The test happened once, at the factory, when the lubricants were fresh. As oils age over the years, rates drift — usually slowly, sometimes in one direction — so a watch that left the factory at +1 s/day can be running +8 a decade later without anything being "broken."

That means the only number that matters for your watch is the one you measure yourself, averaged over days rather than hours. This is exactly the problem an app like ChronoLog solves: it timestamps your watch against atomic time, computes the daily rate in seconds per day, and tracks the history per watch so you can see whether yours sits inside its brand's spec — and whether that rate is drifting over time.

What to do if your watch is outside spec

First, confirm it properly. Measure the average rate over at least three to five days of normal wear with the watch reasonably wound. A single bad day proves nothing; a consistent average outside the published range is real data. If you want to isolate the cause, add one controlled variable: rest the watch overnight in one position (dial-up, then crown-down on another night) and compare the rates. A large gap between positions points to a regulation or poise issue rather than anything serious. Then match the overall pattern to the cause:

Sudden jump to running fast. A watch that was fine last month and now gains 30+ seconds a day is very likely magnetized — common, harmless, and fixable in minutes with a demagnetizer. See why a watch runs fast or slow for the symptoms and the fix.

Consistently out of spec, but steady. A healthy movement running a steady +12 when its spec says +6 usually just needs regulation — a small adjustment to the balance, quick and inexpensive for any watchmaker. Under warranty, the brand will typically handle it; Rolex, Omega and others regulate in-spec as part of warranty service.

Erratic and drifting. If the rate swings wildly from day to day or has slowly worsened over years, the movement may be due for a service — dried lubricants and worn parts show up first as inconsistency, then as a slow decline in timekeeping.