Timegrapher repeatability calculator
Enter two mechanical-watch rate readings and their fit uncertainty. The calculator tells you whether the measurements support each other under the same setup.
These readings agree.
0.2 s/d apart · 2.0 s/d agreement limit
The difference is inside the agreement limit. Keep the pair as repeatable evidence under this setup.
What the verdict means: agreement demonstrates repeatability under these conditions. It does not prove calibration against a trusted reference, and it does not diagnose the watch.
The comparison rule
Each rate fit has its own standard error. The calculator combines both errors in quadrature, doubles that combined uncertainty, and compares it with a practical 2.0 seconds-per-day floor:
combined uncertainty = √(first error² + second error²)
agreement limit = max(2 × combined uncertainty, 2.0 s/d)
The readings agree when their difference is less than or equal to the agreement limit. The 2.0 s/d floor covers ordinary between-session changes that a fit's within-run standard error cannot see, including small differences in placement, wind state, temperature, and the watch itself.
| First | Second | Difference | Verdict |
|---|---|---|---|
| +2.1 ± 0.4 s/d | +2.3 ± 0.4 s/d | 0.2 s/d | Agree |
| +2.1 ± 0.4 s/d | +6.8 ± 0.5 s/d | 4.7 s/d | Check again |
How to get useful inputs
- Use two fresh measurements rather than continuing the first fit.
- Keep the watch in the same position and state of wind.
- Do not move the watch or phone while either run is settling.
- Enter each rate and the fit uncertainty reported beside that rate.
- If the pair disagrees, re-seat the watch and repeat before trusting either number.
The longer guide explains why timegrapher readings change between tests, how repeatability differs from accuracy, and which setup differences matter most.
Measure and save the evidence with Hairspring
Hairspring is an iPhone timegrapher for mechanical watches. It measures rate, beat error, live amplitude, and common beat rates; states statistical fit uncertainty; withholds results the signal cannot defend; and saves a two-run agreement receipt in History and PDF reports.
One $9.99 purchase. No subscription, account, ads, analytics, or automatic uploads.
Frequently asked questions
How do I compare two timegrapher readings?
Enter both rate readings and each run's fit uncertainty. The calculator compares their absolute difference with the larger of twice their combined uncertainty and a 2.0 s/d floor.
Do agreeing readings prove that a timegrapher is accurate?
No. Agreement demonstrates repeatability under the recorded conditions, not calibration against a known reference. Two readings can agree while sharing the same systematic bias.
Why is there a 2 seconds-per-day floor?
Fit uncertainty describes scatter inside each run. It cannot capture every small change between runs, so the floor prevents unrealistically tight comparisons when both internal error bars are tiny.
Does this calculator upload my readings?
No. The calculation runs in your browser. There is no account, analytics, or submission endpoint. The tested source is public.