Almost everyone's I would have guessed. You don't own it, do you?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple's Kevin Lynch Explains Method Behind Apple Watch's Precise Timekeeping
- Thread starter MacRumors
- Start date
- Sort by reaction score
Almost everyone's I would have guessed. You don't own it, do you?
If you make sure you are always 10 mins earlier than the appointment time, it would not matter what watches you wear, and it would also add points towards your politeness.
I couldn't agree more in this instance. A MacBook Pro will lasts user 4-5 years. An HP top end laptop will give you 2, maybe 3 years reliable use.
I switched to Mac 6 years ago and I've never looked back.
Great old aphorism, but unfortunately it doesn't apply to the Apple Watch.
Yes, they're consumer level watches. What in the world does that have to do with NTP synchronization? A hammer is also a consumer level product. It also does not sync time using NTP. I quite clearly said I was happy that there was a consumer level watch available that syncs time via NTP. The Casio Waveceptor syncs time several times a day by listening for a radio signal. It often does not receive the signal, and just... keeps running with its best guess of the time. Go read the FAQs for the watches, they say it's best to not have them in a building, or to put them in a window (with the watch oriented towards the direction of the transmitter tower), to get the syncing to work. It's an iffy thing. I have a Casio G-Shock Solar that works the same way. Sometimes it syncs. Often it doesn't. If the reason for syncing it to be assured you have the correct time, there's no point in having a watch that maybe syncs sometimes - because you can't be assured it has the right time. And there is no feedback loop between client and server to actively account for the transmission/reception delay, something which the NTP protocol properly handles.
In order to sync, the Apple Watch needs a good Bluetooth connection to the phone, and the phone needs a decent network connection (wifi or cell). I'd wager most Apple Watch users have both of these almost all the time. There's no "well, I should put it in the window pointing north east at midnight to maybe get the time". It just happens.
To be clear:
To be assured of the correct time with the Apple Watch (yes, assuming it's charged - I haven't found that to be a problem):
1. Look at the watch's display and read the time.
To be assured of the correct time using a Casio Waveceptor:
1. Look at the watch's display and read the time.
2. Wonder when the last time was that it successfully received a time signal.
3. Push some buttons to check the last time it synced correctly (mine will show the date last synced), and decide if it was recent enough for the watch's notion of time to be reasonably accurate.
To be assured of the correct time using most other watches:
1. Look at the watch's display and read the time.
2. Wonder when the last time was that you set the time on the watch (and how accurately you set it).
3. Decide if it was recent enough, or if you should look around for some other more trustworthy time source.
I strongly prefer the one-step Apple Watch / NTP way. I like that I can glance at the watch and "know" the time with a high degree of reliability (since the prerequisite network connections are basically always there).
That's fine, but let's not paint it as sone sort of "innovation" from Apple. As already stated and confirmed in the article itself, this concept is nothing new and basically how smartwatches operate since years before the Apple Watch.
How accurate is it?
If you remove the Apple watch from being able to "copy the time" from outside sources, like the vast majority of time devices in the world, just how accurate is it?
Do we know?
As I said jokingly earlier, I don't think anyone would regard YOU personally as having the amazing ability to always know the correct time, if I rushed up to you every xx minutes and told you the right time.
I was playing with the display model Apple Watch in Currys on Thursday so I checked it's accuracy after reading this thread and it was running 12 seconds fast.
Compared to the Atomic Time Synced time on the Watchville app on my iPhone.
First, why characterize my industry time as a claim? I didn't question yours. I only mentioned it because you put yours on the table to attempt to add authority to your words. Second, I don't think "the public" spends much time reading the comments on MacRumors. Third, show me where I said Apple's use was technically challenging. I love that they've done it, as I've long wanted a watch that got its time from NTP, but I never characterized doing so as especially challenging. Yes, NTP is in many devices - but server and desktop uses of NTP have NO relevance in a conversation about WATCHES, unless you regularly carry a server around with you. Why keep bringing them up?
As far as other wearables are concerned, try googling around about the accuracy of Pebbles, you'll find their internal timekeeping accuracy is not good, and they don't sync with the phone's time very often, and when they do they get the phone's time which often isn't that accurate. (An iOS app called "Emerald Time" will display the time "live" off of NTP servers, and will also show the offset between the iPhone's clock and the NTP source - currently mine is showing 16ms off, which is well within Apple's 50ms claim; in previous iOS versions, before they made changes for the Apple Watch, this number was often somewhat higher.)
Considering the number of apps in the Google Play store for using NTP to sync the phone's system clock, I think you'd be hard pressed to argue that most Android phones use NTP as a time source by default. And it looks like Android Wear watches generally get their time from the associated phone, and googling shows plenty of problems with Android Wear watches showing the wrong time.
So please tell me which other "wearables" make use of NTP to set their time, out of the box?
Does it surprise you to learn that I don't really care what you want?
I brought up the things I did specifically in response to other comments.
Stating something repeatedly does not make it true, and as for the confirmation in the article it comes down to this note they got from a watch blogger: "Apple's claims about timing accuracy is impressive, but it isn't what I would consider really innovative given that it is logical to have your connected watch feed from your mobile phone's clock which itself is based on GPS signals or the network (which probably relies on an atomic clock somewhere along the line)"
Note the highlighted parts - he is merely supposing about how smart watches work. And he is talking about accuracy, not about NTP. NTP use is what I've been discussing. Please tell me another smartwatch that syncs to NTP. I said that using NTP in a watch was novel, and several folks keep saying, "no, NTP has been used for lots of things like desktop computers and servers." Which is neat, but is NOT a refutation of my original point that NTP for a (consumer level) watch is novel.
Oh, just for fun, I looked up Hoptroff watches (the other brand's website requires Flash) mentioned as already having NTP. The cheapest one comes in at over $17,000 (with a 3 month lead time for them to build it for you). Not exactly consumer level. And for that, you get mechanical hour, minute and second hands. Very elegant to be sure, but not a smartwatch, and not consumer level.
Thanks.
Appreciate you posting that.
Whilst it's great that the watch does sync to external sources to keep correcting itself.
I will admit, I would not dare BOAST about how accurate it is, unless it's pretty accurate in it's own right without constant outside correction.
You see THIS would be something for Apple to shout about if their watch was as accurate without cheating:
http://www.seiko.co.uk/collections/men/grand-seiko/sbgx061#.VofFpo_XLmE
Accuracy: +/- 10 seconds per year
UK Price: http://www.citywatches.co.uk/seiko/grand-seiko-mens-watch-quartz-sbgx061.html
http://www.seiko.co.uk/collections/men/grand-seiko/sbgx061#.VofFpo_XLmE
Accuracy: +/- 10 seconds per year
UK Price: http://www.citywatches.co.uk/seiko/grand-seiko-mens-watch-quartz-sbgx061.html
To play devil's advocate, how long ago was it set, and was it set accurately to begin with? You can't really evaluate the drift without knowing duration and original setting. And since the whole point of the watch is to work in concert with an iPhone, knocking it for how accurate it may or may not be alone... hmm, not sure how valid it is to get upset over that. I don't know.
I think they can get away with it because it is constantly synching to the iPhone in normal use. The functionality of the Watch is severely reduced when it is used away from the iPhone and of course it will need recharging daily which will usually bring it back in contact with the iPhone.
I can't see many people buying the Apple Watch and not using it as an accessory to their iPhone.
I would not mind if it were not for the way Apple market it.
If I said to you, this watch I made will always show the exact time as it's linked to atomic clocks which keep the watch accurate so you know it's always the exact time.
That's fair and honest, all good
Apple say:
At its heart, an incredibly precise timepiece.
High-quality watches have long been defined by their ability to keep unfailingly accurate time, and Apple Watch is no exception. In conjunction with your iPhone, it keeps time to within 50 milliseconds of the definitive global time standard.
Which means the time, but they are HEAVILY inferring that, like other previous watches on the market it is an "incredibly precise timepiece"
You want the Apple Watch, disconnected from the phone it was designed to work in concert with (I assume that's what you mean by cheating), to run more accurately than a watch that costs five and a half times as much? Operating both as designed, and without user intervention, after three years the Seiko will be +/- 30 seconds, and the Apple Watch will still be +/- 50 milliseconds.
FWIW, my Marathon Divers Medium Quartz, also under one-fifth the price of the Seiko, loses half a second per week, or about 26 seconds a year. Better bang for the buck. (If I didn't have an Apple Watch, I'd be wearing the Marathon all the time, it's rugged, simple, and lit by tritium tubes so always readable.)
This is correct. Using NTP in an inexpensive stock consumer watch is fairly novel.
And it still is, because of course the Apple Watch does not use NTP itself.
(Sure, like any smartwatch with built-in WiFi, it could connect directly to an NTP server, but it apparently does not.)
--
Instead, the host iPhone is what uses NTP. Then the watch syncs with the host device, likely using a proprietary protocol.
So the real novelty here is Apple using NTP on the iPhone (*). That's because normally (and this was true for iPhones as well until recently), cell phones default to getting their time via the local cell network. The advantages of this are that the phone automatically gets the local time zone, is mostly in sync with other local phones, and doesn't have to use data allowances for timekeeping. On the downside, the passed cell network time is often off by anywhere from less than a second to a dozen seconds or more.
(*) Although this is also true of 1) WiFi-only devices hosting an associated smartwatch, 2) rooted Android smartphones that have an NTP sync app, which helps their linked Android watches, 3) Android smartphones used on a network with no local NITZ time service, at which time they fall back on NTP.
Last edited:
You are still missing the fact that the Apple Watch cannot use NTP directly either: This, again, is stated in the article:
This means that the Apple Watch ultimately requires its iPhone companion to be able to keep its time synchronized. This is nothing new: Pebble smartwatches were using the same concept back in 2013. What might be new and interesting is the way Apple claims to handle the delay between iPhone and Apple Watch, but in the article doesn't go into details of this aspect.
The Seiko was just a quick example
Of course I would expect a LOT of normal watches to have good accuracy like yours does.
I just feel it's bad if (and we don't know for sure) The Apple was is being sold on HOW ACCURATE it is as a "Timepiece" if it was say pretty inaccurate unless it keeps getting put right by outside sources.
Why, don't you believe me?
Since you're making the extraordinary claim that Apple is lying about the 50ms accuracy of the Apple Watch, and that you witnessed one running 240x faster, then I consider it your responsibility to support your contention with evidence.
Since you're making the extraordinary claim that Apple is lying about the 50ms accuracy of the Apple Watch, and that you witnessed one running 240x faster, then I consider it your responsibility to support your contention with evidence.
As we are trying to say/find out.
It's one thing to state you have created a super 50ms accurate timepiece.
It's a different thing to say it needs correcting perhaps once every few minutes or a hour to keep it accurate.
Personally if the Apple watch was less accurate than say a ten dollar watch when it was away from places that would correct it, I'd not personally shout about it's accuracy as a device.
As I say, put the accurate time on a TV screen, have you watch it, and when asked you can tell anyone the exact time as you are being told the time from the screen you are looking at. Does not make YOU a super accurate timepiece does it?
Turn the TV off, leave you to count the seconds by yourself (you are on your own) and you won't be very accurate.
Perhaps neither will the Apple watch? We don't know?