[ferongr]

That's terrible. Hacking prevention would make sense if they were ±2 seconds a month.

[Raidion]

1 minute a month (upper bound) is hardly a deal breaker. If you were adjusting the watch regularly without hacking (as a pilot) you'd have an error rate that's probably close to 10x higher. 10+ minutes a month WOULD make a difference.

[throw0101a]

> 1 minute a month (upper bound) is hardly a deal breaker.

For a relative measure: if you're off by one minute when using a sextant, then you could be off by (IIRC) about 15 nautical miles.

So one should at least know the average drift of the time piece if it's being used for navigation, and it should be sync'd up with UTC/GMT regularly.

For regular day-to-day stuff, one minute is probably less dire.

[NhanH]

+- 2 seconds per day is their worst case specifications (watches being stored in weird orientation, with high temperature etc.). It is common for standard usage watch to be within 10 secs monthly.