Analysis of Apple Watch running data • Hacker News Board

Re: Analysis of Apple Watch running data

#11

Post by oezi » Sun, Sep 25, 2022, 2:23 PM UTC

cianmm wrote: post #5

I’m part of a running group with a mix of Garmin and Apple Watches and the Apple Watches always have pretty different distant readings according to Strava, maybe 2%-5%. The Garmin devices are generally much closer together. I wonder if the dual-band GPS on the Apple Watch Ultra is an attempt to fix these problems? I would guess that it’s software, with the author, if for no other reason than I’d be surprised if Garmi…

I wonder if author really got a GPS fix before starting to run. I have clocked 100 runs on watch os 8 exactely along one identical path and usually it is within 3-5 meters where I get the 1km announcement.

Re: Analysis of Apple Watch running data

#12

Post by Clent » Sun, Sep 25, 2022, 2:23 PM UTC

kadoban wrote: post #6

cianmm wrote: post #5

I’m part of a running group with a mix of Garmin and Apple Watches and the Apple Watches always have pretty different distant readings according to Strava, maybe 2%-5%. The Garmin devices are generally much closer together. I wonder if the dual-band GPS on the Apple Watch Ultra is an attempt to fix these problems? I would guess that it’s software, with the author, if for no other reason than I’d be surprised if Garmi…

> I would guess that it’s software, with the author, if for no other reason than I’d be surprised if Garmin were all that much better at putting GPS in a tiny housing than Apple. After "you're holding it wrong", anything seems possible. But yeah that does seem more likely to be software, it's a surprisingly difficult and fuzzily-defined problem.

Weird hill to die on with Apple. That's a 10 year old issue with an iPhone 4, said by a CEO that is long dead.

They gave everyone free bumpers, because sweaty fingers would close the antenna gaps. The statement was true, the gaps were placed where they were expected to least affect the grip.

What exactly would you expect in this case?

The messaging worked. If your iPhone 4 was having signal issue, consider readjusting your hand.

Apple sent out free cases to compensate for the issue.

This is not the nightmare scenario that non-iPhone users made it out to be. Apple haters, like any group of haters are a silly bunch.

Re: Analysis of Apple Watch running data

#13

Post by wslh » Sun, Sep 25, 2022, 2:27 PM UTC

mtts wrote: post #9

cianmm wrote: post #5

I’m part of a running group with a mix of Garmin and Apple Watches and the Apple Watches always have pretty different distant readings according to Strava, maybe 2%-5%. The Garmin devices are generally much closer together. I wonder if the dual-band GPS on the Apple Watch Ultra is an attempt to fix these problems? I would guess that it’s software, with the author, if for no other reason than I’d be surprised if Garmi…

> I’d be surprised if Garmin were all that much better at putting GPS in a tiny housing than Apple. Given Garmin’s long, long history as a manufacturer of (often very small) GPS devices, I personally wouldn’t be. (I also know for a fact they do mapping better. This summer I was in the South of Italy and only Garmin accurately distinguished between small public roads and long private driveways while both Google and Ap…

I used to use a FitBit Ionic (GPS enabled watch) and now use a Garmin watch and fortunately it shows a more accurate and better exercise results. In terms of running I saw a difference of ~10-20% in the distance measured in some terrains. I have not precisely compared everything.

BTW the Fitbit Ionic GPS stopped working and now (if it didn't expire) I will go for the health (battery) recall: https://help.fitbit.com/en_US/ionic.htm and give them to another person. This will be the second replacement since this model stopped working after 1 yr of use and Fitbit sent me a new one.

Re: Analysis of Apple Watch running data

#14

Post by jquery » Sun, Sep 25, 2022, 2:29 PM UTC

“Pausing always fixes erratic mode” makes me suspect they don’t have a GPS lock for some reason.

Re: Analysis of Apple Watch running data

#15

Post by fuzzy2 » Sun, Sep 25, 2022, 2:29 PM UTC

I wonder if cycling is also affected. Theoretically, the watch should be in an almost ideal position on flat handlebars, if less so on drop bars.

Re: Analysis of Apple Watch running data

#16

Post by PragmaticPulp » Sun, Sep 25, 2022, 2:34 PM UTC

> After about a kilometer I have to cross a busy street, I need to pause there regularly. Pausing always fixes erratic mode.

Given that the “erratic mode” only impacts the beginning of the run and disappears suddenly at known locations, this sounds like an issue of delayed GPS lock.

And given that the problem didn’t exist in the past, it could be a software issue due to upgrades. Or it could be a hardware issue that developed over time, such as something impacting GPS receive sensitivity. Or it could even be a new source of RF interference in the GPS range near the author’s start point, which impacts GPS lock until they get far enough away from it.

Interesting issue, but note that this issue appears to be specific to this one specific person, not a general issue with all Apple Watches as some in this thread are speculating. I certainly have not noticed this behavior on my Watch even with the latest updates.

Re: Analysis of Apple Watch running data

#17

Post by PragmaticPulp » Sun, Sep 25, 2022, 2:37 PM UTC

cianmm wrote: post #5

I’m part of a running group with a mix of Garmin and Apple Watches and the Apple Watches always have pretty different distant readings according to Strava, maybe 2%-5%. The Garmin devices are generally much closer together. I wonder if the dual-band GPS on the Apple Watch Ultra is an attempt to fix these problems? I would guess that it’s software, with the author, if for no other reason than I’d be surprised if Garmi…

> I wonder if the dual-band GPS on the Apple Watch Ultra is an attempt to fix these problems?

The author said their watch worked fine in the past. It also works fine after a warm up period, which suggests it’s not getting a GPS lock at the start of the workout.

It’s a new issue of either a software regression, hardware degradation, or RF interference near their start point.

Re: Analysis of Apple Watch running data

#18

Post by rvba » Sun, Sep 25, 2022, 2:37 PM UTC

Apple showing more than people actually run to make them feel better?

Seems like a lawsuit incoming due to "ruined health" by overstating measured activities.

Re: Analysis of Apple Watch running data

#19

Post by lifeinthevoid » Sun, Sep 25, 2022, 2:38 PM UTC

aside: My girlfriend tracks her runs with Strava through her Iphone 12 mini. Upon zooming in, the "straight" parts never look straight and look a lot more like a triangle wave. She consistently tracks around 10% larger distance than my Garmin watch. I've checked on the map, and the Garmin distance is the accurate one of the 2.

Re: Analysis of Apple Watch running data

#20

Post by rajman187 » Sun, Sep 25, 2022, 2:42 PM UTC

mwidell wrote: post #8

Only way to get accurate real time data when running is to use a good foot pod like Stryd. GPS just isn't as accurate as people think, no matter if you use Apple Watch or Garmin.

That's right, GPS error can be on the order of 10s of meters, worse in urban environments. Google (and presumably Apple) use WPS to enhance location data, meaning a triangulation (simplification here) or location based on known WiFi router and bluetooth beacon positions. They also have a team dedicated to more complex geometric calculations of how the GPS signal may be reflected given the known positions of buildings.

But if you have a clear line of sight that's a different issue. I went cycling on a path next to a river and the GPS trace was buttery smooth.