Properly generated line power keeps better long-term accuracy than a 32.768 crystal oscillator, which is the timebase of any 'quartz' wristwatch. Most of these are good for about a second a day systematic error, which the makers appear to run about that fast; this makes it easier to resynchronize the watch to a trusted time source, simply by pulling out the stem for a few seconds once in a while.
Complications arise with a line-frequcny clock when you have blackouts, of course; they're an inconvenience on the same order as your watch battery running out.
Nowadays, many clocks and some watches have Very Low Frequency receivers for NIST station WWVB, or equivalents in Europe and Asia. These timepieces are the ones touted as 'atomic.'
A few, but not many, clocks and appliances have an internal 32.768 KHz oscillator to back up their line frequency time base. A good combination, but not that common. The 'atomic' timebase is, I think, the more common type by now.
By the way---My most recent watch, a Seiko chronograph, comes amazingly close to zero. I believe it stays within about two seconds per month. Even after falling on it, smashing the crystal, and having it fixed.
Apropos of very little, here's a bit of trivia.
One of the inventors and early marketers of the electric clock was a man by the name of Laurens Hammond. He went on to become better known as the inventor of the electronic organ.
I have an old electric that used to belong to my grandmother. It had a synchronous motor driving a gear train. However, it was a first generation synchronous motor: It had a gnarled knob on the back you needed to give a twist to to get it going. It kept great time (unless you twisted the knob the wrong way, in which case it kept great time in reverse).
The power companies kept the frequency very close to 60 hz. And if it deviated one way, they would run it deviated the other for a while to compensate. That meant the frequency over a long time period averaged extremely close to the nominal value. So, barring power failures and DST changes, you didn't need to touch an electric with a synchronous motor.
Properly generated line power keeps better long-term accuracy than a 32.768 crystal oscillator, which is the timebase of any 'quartz' wristwatch. Most of these are good for about a second a day systematic error, which the makers appear to run about that fast; this makes it easier to resynchronize the watch to a trusted time source, simply by pulling out the stem for a few seconds once in a while.
That's been my observation as well. The average electronic timepiece, be it an expensive computer (not using NTP) or a cheap Radio Shack stick-on, tends to gain between half a second and a second per day. For some reason, they usually seem to run fast, not slow.
The typical WWVB timepiece synchronizes once per day, in the wee hours of the morning when reception is likely to be best. During the day, it relies on its crystal control, which is no more accurate than typical. But at least the error gets corrected each day.
The better WWVB watches also synchronize only early in the morning. However, they note the degree of error accumulated while using the crystal and compensate for it (similar to what a good NTP client on a computer will do). The biggest source of error seems to be temeperature. So, as long as the watch experiences a uniform average ambient temperature, such compensation works well.
The better radio-controlled watches contain hardware / software that allows them to use European and Japanese time signals as well as WWVB.
By the way---My most recent watch, a Seiko chronograph, comes amazingly close to zero. I believe it stays within about two seconds per month. Even after falling on it, smashing the crystal, and having it fixed.
Does it use WWVB? Or does it have one of those premium, temperature-compensated crystal time bases?