Watch Accuracy Calculator
Enter how many seconds your watch has gained or lost, and over how long, and this tool works out its true accuracy in seconds per day — then tells you exactly where it lands against the real chronometer standards, from an everyday automatic to a COSC, METAS, or Rolex-certified movement.
Seconds-Per-Day Accuracy Calculator
Set your watch against an atomic clock, wait a few days, then enter the drift you measured. Use a positive number if the watch ran fast, negative if it ran slow.
A single-day reading can be misleading because a mechanical watch keeps different rates in different positions. For a rate that reflects real wear, measure over at least 3–5 days of normal use. Learn more in the how to measure section below.
“How accurate is my watch?” is one of the first questions a new owner asks — and the answer is almost always a single number: the average seconds per day the watch gains or loses. Getting that number is simple arithmetic (total drift ÷ days elapsed), but knowing whether it is good depends entirely on what kind of watch you own. A rate that would be embarrassing for a quartz watch is excellent for a vintage automatic. This guide gives you the calculator, the verified benchmarks to judge against, and a repeatable method for measuring accuracy properly so your result actually means something.
The accuracy spectrum
Every mechanical watch lives somewhere on a scale from “needs regulation” to “chronometer.” The tighter the band around zero, the more precise the watch. Here is how the major standards stack up, all in seconds per day.
The benchmarks, explained
These are the figures the calculator compares your watch against. Each is an officially published specification, not a rule of thumb — but remember that certifications describe a guaranteed maximum, and a well-cared-for watch often does better.
| Standard / type | Accuracy | What it tests |
|---|---|---|
| Rolex Superlative Chronometer | −2 / +2 s/day | Fully cased watch, in-house, since 2015 |
| METAS Master Chronometer | 0 / +5 s/day | Cased watch + anti-magnetism (Omega, Tudor) |
| COSC chronometer | −4 / +6 s/day | Uncased movement, 15 days, 5 positions |
| Typical automatic (non-chronometer) | ≈ −10 / +20 s/day | Normal, uncertified mechanical watch |
| Seiko Spring Drive | ≈ ±1 s/day (±15 s/mo) | Mechanical drive, quartz-regulated |
| Standard quartz | ≈ ±0.5 s/day (±15 s/mo) | Everyday battery-powered quartz |
| High-accuracy quartz (GS 9F) | ±10 s/year | Grand Seiko 9F caliber |
| High-accuracy quartz (Chronomaster) | ±5 s/year | Citizen Chronomaster (The Citizen) |
Note: COSC certifies the uncased movement, so a finished chronometer may behave a little differently on the wrist. METAS and Rolex both test the complete, cased watch, which is why their windows are quoted for real-world wear. Curious about the testing itself? See our guide on what COSC certified means.
How small drift adds up over a month
A few seconds a day sounds trivial, but it compounds. This is why the difference between a chronometer and a typical automatic is very visible after a month: the same watch left unset can be off by anything from under a minute to several minutes.
How to measure your watch's accuracy properly
A reading is only as good as the method behind it. Follow these steps to get a rate you can trust — and that this calculator can classify meaningfully.
- 1Sync to an atomic-clock reference. Use a source like time.gov, time.is, or an NTP-synced phone. Set the watch's seconds hand exactly to the reference — hack the seconds (pull the crown to stop them) if your watch allows it, so you can start precisely on the minute.
- 2Wear it (or wind it) consistently. An automatic needs to stay powered. Wear it the way you normally would, or keep it fully wound; a half-wound watch on a nightstand will read differently from one on your wrist all day. Power-reserve state changes the rate.
- 3Wait several days — not one. Measure over at least 3–5 days. This averages out positional variance (a watch left dial-up overnight can gain, crown-down can lose) and gives a rate that reflects real use.
- 4Compare against the same reference. At the end, note how many seconds ahead (fast) or behind (slow) the watch is versus the atomic clock. That's your total drift.
- 5Divide, or let the tool do it. Drift ÷ days = seconds per day. Enter both numbers at the top of this page and read off the verdict.
Want a deeper look at what the numbers say about the movement's health? Our movement accuracy grader interprets amplitude and beat error alongside rate.
Why an honest reading is harder than it looks
Mechanical watches are not metronomes. Three factors move the number around, and ignoring them is how people convince themselves a healthy watch is “broken” — or a struggling one is fine.
Positional variance
The balance wheel runs at slightly different rates depending on gravity's pull, so dial-up, dial-down, crown-up, crown-down, and crown-left all give different figures. This is exactly why COSC measures across five positions. A single overnight resting position can swing a one-day reading by many seconds, which is why you should always average over several days.
Power-reserve state
A mainspring delivers more even torque when it's well wound. As it unwinds toward the end of the power reserve, amplitude drops and the rate can drift. An automatic that spends the night off the wrist and low on power will often read differently from one kept fully wound. Keep the winding state consistent between measurements.
Magnetism and temperature
A magnetized hairspring is the single most common cause of a watch suddenly running fast — sometimes minutes per day. Phone speakers, laptop lids, magnetic clasps, and tablet covers are frequent culprits, and demagnetizing is a quick fix. Large temperature swings also nudge the rate. If your number is wildly outside the expected band, suspect magnetism before you suspect a fault.
Verify your watch is genuine — first scan free
Accuracy is one clue; a full authentication is the real answer. Upload a few photos and our AI checks the dial, caseback, movement, and finishing against thousands of references.
Scan my watch freeverdict in ~30s · no app to install
Reading your verdict
The calculator sorts your seconds-per-day rate into plain-language verdicts. Here's what each one means and what to do next.
“Within COSC / chronometer”
Excellent for a mechanical watch. Your rate falls inside the certified windows. Nothing to do — enjoy it.
“Typical for an automatic”
Normal and healthy for an uncertified movement. If you'd like it tighter, a watchmaker can regulate it closer to zero.
“Needs regulation”
Outside the normal band. First rule out magnetism and low power; if it persists, it's a candidate for regulation or service.
“Quartz-grade”
Sub-second-per-day precision points to a quartz or quartz-regulated movement like Spring Drive — no mechanical watch measures this cleanly.
Not sure whether your watch is mechanical or quartz in the first place? Our guide on automatic vs quartz explains how to tell at a glance.
Frequently asked questions
How do I calculate my watch's accuracy in seconds per day?
Set your watch to the exact time from an atomic-clock reference, wait a fixed number of days without resetting it, then compare it to the same reference again. Divide the total drift in seconds by the number of days elapsed. For example, if your watch is 12 seconds fast after 4 days, that is +3 seconds per day. Gaining time is a positive rate; losing time is negative. This calculator does the division for you and tells you whether the result meets chronometer standards.
What is a good accuracy for a mechanical watch?
A typical non-chronometer automatic runs roughly −10 to +20 seconds per day and is considered normal. A COSC-certified chronometer must average −4 to +6 seconds per day on the uncased movement. METAS Master Chronometer watches target 0 to +5 seconds per day on the fully cased watch, and Rolex guarantees its Superlative Chronometers to −2 to +2 seconds per day. Anything within a few seconds a day is excellent for a mechanical watch.
Is −4/+6 seconds per day the COSC standard?
Yes. COSC (the Swiss Official Chronometer Testing Institute) certifies uncased mechanical movements against an average daily rate of −4 to +6 seconds per day, measured over 15 days in five positions at three temperatures. Because it tests the movement rather than the finished watch, a cased chronometer can perform slightly differently on the wrist.
How accurate is a quartz watch compared to a mechanical one?
A standard quartz watch is accurate to about ±15 seconds per month, which is roughly ±0.5 seconds per day — far tighter than most mechanical watches. High-accuracy quartz movements such as the Grand Seiko 9F (±10 seconds per year) and the Citizen Chronomaster (±5 seconds per year) are dramatically more precise still. Seiko's Spring Drive sits between the two worlds at about ±1 second per day.
Why does my watch's accuracy change day to day?
Mechanical watches keep slightly different rates in different positions (dial up, crown down, and so on) and as the mainspring winds down. A watch measured over a single day can look wildly off simply because it spent the night in a fast or slow position. Averaging over several days of normal wear cancels out this positional variance and gives a rate that reflects how the watch actually performs on your wrist.
Can I regulate my watch to make it more accurate?
Yes. A watchmaker can adjust the regulator or the balance to bring a mechanical movement closer to zero, and many watches can be improved to within a few seconds a day. Consistency matters more than a perfect average: a watch that reliably gains a steady amount each day is healthy and easy to regulate, whereas one whose rate jumps around unpredictably may need service.
Does a large daily gain or loss mean my watch is fake?
Not on its own — genuine mechanical watches can run far outside chronometer specs if they are demagnetized, low on power, or overdue for service. But poor and erratic timekeeping is common on counterfeits, which often use cheap movements. Accuracy is one clue among many; pairing it with a check of the dial, caseback, movement, and serial number gives a far more reliable read on authenticity.
Important Note
Timekeeping accuracy reflects a movement's regulation, not by itself its authenticity or overall condition. Positional variance and power-reserve state genuinely affect readings, so treat a single number as a guide rather than a diagnosis. For a valuation, an authenticity question, or a suspected fault, an in-person inspection by a certified watchmaker is always the gold standard.