Rate deviation
The difference between a watch's actual rate and a reference time, measured in seconds per day. The headline figure in chronometer testing and the key indicator of movement condition at service and purchase.
At a glance
- Unit
- seconds per day (s/d); for high-precision quartz seconds per month or year
- Sign
- + = fast, − = slow
- Instrument
- Witschi or Greiner timing machine with microphone
- Standard positions
- dial up, dial down, crown left, crown up, crown down
- Healthy amplitude
- 270° to 310° horizontal, ~220° to 270° vertical
- Ideal beat error
- < 0.5 ms
- Uncertified mechanical tolerance
- typically ±20 to ±40 seconds per day
- COSC tolerance
- -4 to +6 seconds per day
- METAS tolerance
- 0 to +5 seconds per day (cased)
- Degradation across service cycle
- 5 to 15 s/d added deviation over 7 to 10 years
Rate deviation is the difference between a movement's actual rate and a reference time, typically measured in seconds per day. It is the headline figure in every chronometer test and, in day-to-day service, the key indicator of movement condition. In the premium segment, allowed deviation sits in single-digit seconds per day — for a mechanism whose components engage several million times in those twenty-four hours.
How it is measured
The classical tool is the timing machine — today Witschi or Greiner with microphone and sensor, formerly the optical timegrapher. The movement is set in a defined position, the microphone picks up the escapement, and software computes the rate from frequency and stroke pattern. A multi-position run produces the complete rate profile.
Three parameters matter:
- Rate — seconds per day, positive (watch runs fast) or negative (slow).
- Amplitude — swing angle of the balance in degrees. Healthy values are 270 to 310 degrees horizontal.
- Beat error — asymmetry of the escapement strokes in milliseconds. Ideally near zero.
A full rate test at our atelier in Munich captures all three values across multiple positions — usually dial up, dial down and crown down.
Tolerances by movement type and standard
| Movement / standard | Rate tolerance |
|---|---|
| Standard quartz | ±15 seconds per month |
| High-precision quartz (e.g. Grand Seiko 9F) | ±10 seconds per year |
| Mechanical, uncertified | typically -20 to +40 seconds per day |
| COSC chronometer (mechanical) | -4 to +6 seconds per day |
| Rolex Superlative (cased) | -2 to +2 seconds per day |
| Master Chronometer (METAS) | 0 to +5 seconds per day |
| Patek Philippe Seal (cased) | -3 to +2 seconds per day |
The values are not directly comparable: COSC measures the bare movement, Rolex Superlative and METAS the cased watch — an in-case measurement is methodologically more demanding because case influences enter the result.
What causes rate deviation
Three principal groups:
- Mechanical condition. Contaminated or gummed escapement, fatigued hairspring, damaged gear train. Classical wear patterns addressed at service.
- Magnetisation. A magnetised hairspring sticks to itself; the watch typically runs significantly fast. A single demagnetisation session at our workshop normally returns the movement to tolerance.
- Wearing conditions. Wrist position, temperature and wind state all affect rate. Workshop figures under standardised conditions therefore commonly differ from "wrist-worn" rate.
Practical relevance
- Service trigger. A previously well-running movement that suddenly drifts by 20 or 30 seconds per day points most often to magnetisation or oil ageing — no catastrophic damage, but cause for a workshop check.
- Purchase and valuation. We measure every watch at intake at our atelier. Values within the specified tolerance indicate a healthy movement; significant drift can point to undeclared shocks, water ingress or tampering.
- Regulation. A watchmaker's regulation of the hairspring corrects moderate drift within minutes. Deeper issues — escapement wear, gear-train play — require a full overhaul.
What rate deviation does not say
- Nothing about long-term stability. A workshop snapshot does not exclude later wear.
- Nothing about movement quality alone. A well-regulated ETA can run at -1 second per day; that does not make it manufacture construction.
- Nothing direct about water resistance, function reliability or complication precision. These are tested separately.
Frequently asked
- The simplest method: synchronise the watch with a reference (atomic clock app, radio clock) in the morning; compare 24 hours later. The difference in seconds equals the daily rate. A workshop measurement with a timing machine is more precise because it also reports amplitude and beat error.