Apple watch GPS issues during workouts

[Snowkarver]

I'm getting the same consistent 7-10% deficit as the OP when comparing run distances from Watch/Workout vs. Nike+ Running, which I know is correct.

I'm using both Workout and Nike+ at the same time, carrying my iPhone as usual in a Flipbelt, and running the same exact route.

#1 - Nike+/GPS 8.70km, AW 7.93km
#2 - Nike+/GPS 8.19km, AW 7.66km
#3 - Nike+/GPS 8.20km, AW 7.40km

It's kind of annoying. My route admittedly has a lot of elevation change - maybe I should take the iPhone and AW to the track to improve the learning? Is the optimization of the AW accelerometer via the iPhone GPS automatic?

I had the running distance source set in Health as phone, watch, health and just changed it to health, watch, phone. Not sure if that will do anything...

 

[BD1]

I'm getting the same consistent 7-10% deficit as the OP when comparing run distances from Watch/Workout vs. Nike+ Running, which I know is correct.

I'm using both Workout and Nike+ at the same time, carrying my iPhone as usual in a Flipbelt, and running the same exact route.

#1 - Nike+/GPS 8.70km, AW 7.93km
#2 - Nike+/GPS 8.19km, AW 7.66km
#3 - Nike+/GPS 8.20km, AW 7.40km

It's kind of annoying. My route admittedly has a lot of elevation change - maybe I should take the iPhone and AW to the track to improve the learning? Is the optimization of the AW accelerometer via the iPhone GPS automatic?

I had the running distance source set in Health as phone, watch, health and just changed it to health, watch, phone. Not sure if that will do anything...

First I assume you calibrated your watch.....

Have you tried the route with just AW workout app running?

I just did a 2.63 mile run today. This is a regular route that with prior run meter/runkeeper runs measure 2.64 to 2.70 so I am really close. I have some elevation change here but not great. Based on a few other routes I have run, for me the AW is very close but maybe slightly less like today's run.

I hope everyone is submitting feedback to: https://www.apple.com/feedback/watch.html

 

[eyespii]

Did another run today - seems like the tracking on the watch is getting better!

Garmin Forerunner: 13.11 miles, time 1:59:02, 9:05/mi
Runmeter: 13.26 miles, time 1:56:06, 8:45/mi
Apple Watch: 12.95 miles, time 2:04:32, 9:37/mi

I stopped my garmin every time i hit a stoplight, and runmeter has an auto-stop feature, but i let the apple watch keep running.

Also nice, is that the even with a 2-hr running workout w/ the HR monitor going, battery only dropped about 30ish%.

 

[Andy in Embsay]

I'm baffled by this suggestion AW doesn't use GPS for running - given that it does for cycling (it obviously can't use the accelerometer for cycling) - especially as cycling workouts tend to be longer than running ones.

DC Rainmaker has hinted that his tests have been "ok for running, less so for cycling", which is the opposite to what I've found.

 

[Cuyler]

I'm baffled by this suggestion AW doesn't use GPS for running - given that it does for cycling (it obviously can't use the accelerometer for cycling) - especially as cycling workouts tend to be longer than running ones.

DC Rainmaker has hinted that his tests have been "ok for running, less so for cycling", which is the opposite to what I've found.

I don't know what DC Rainmaker might've said, but there is no GPS chip inside the Apple Watch. However, the watch communicates with the iPhone which does have a GPS. I suppose the question is what modes/apps use the phone's GPS and when. I would think the watch would use the GPS at least somewhat during running just as it does in the outdoor walk workouts.

 

[Andy in Embsay]

I don't know what DC Rainmaker might've said, but there is no GPS chip inside the Apple Watch. However, the watch communicates with the iPhone which does have a GPS. I suppose the question is what modes/apps use the phone's GPS and when. I would think the watch would use the GPS at least somewhat during running just as it does in the outdoor walk workouts.

I (and Ray) know the AW doesn't have a GPS receiver in it - but many people assumed that when the phone is there the watch will use the GPS from the phone. But upthread someone says it doesn't use the GPS for running apart from calibrating the accelerometer - if true, I think that's a bit odd as it must use it for cycling, so why not use it for running where the workouts are typically shorter (so not a battery saving thing).

 

[lchlch]

I (and Ray) know the AW doesn't have a GPS receiver in it - but many people assumed that when the phone is there the watch will use the GPS from the phone. But upthread someone says it doesn't use the GPS for running apart from calibrating the accelerometer - if true, I think that's a bit odd as it must use it for cycling, so why not use it for running where the workouts are typically shorter (so not a battery saving thing).

Maybe because it's not as convenient to bring to phone with you when you're running. Whereas it's tolerable while cycling.

 

[dadams1]

For what it's worth, I've done two outdoor runs with the watch/phone on an outdoor track. It nailed the distances both times. So far, so good. I'll keep testing.

 

[Cuyler]

I (and Ray) know the AW doesn't have a GPS receiver in it - but many people assumed that when the phone is there the watch will use the GPS from the phone. But upthread someone says it doesn't use the GPS for running apart from calibrating the accelerometer - if true, I think that's a bit odd as it must use it for cycling, so why not use it for running where the workouts are typically shorter (so not a battery saving thing).

Okay.. sorry I misunderstood. It's also not immediately obvious to me why the GPS is essential for cycling — can't the accelerometer (and time) be used to calculate speeds and total distance? I can see how GPS data might help accuracy — but it'd be nice to do workouts with just the watch. In fact, there's specifically an indoor cycling workout which I assume doesn't use the GPS at all since it's indoors. Forgive me — I haven't done any cycling — I just want to understand the issues.

 

[Julien]

Okay.. sorry I misunderstood. It's also not immediately obvious to me why the GPS is essential for cycling — can't the accelerometer (and time) be used to calculate speeds and total distance?....

Accelerometers work by measuring acceleration. It can measure when you speed up, slow down or turn but at a constant speed in biking there is no measurement and the accelerometer is 'lost' as to what is going on.

In running (and walking) your arm swings back and forth exactly one time for every 2 steps. If the approximate stride length is known then it is easy to calculate distance in running.

 

[Andy in Embsay]

I can only assume that the indoor cycling workout just uses a "average" calorie consumption value and multiplies it by time (presumably uses HR as well).

 

[Cuyler]

Accelerometers work by measuring acceleration. It can measure when you speed up, slow down or turn but at a constant speed in biking there is no measurement and the accelerometer is 'lost' as to what is going on.
...

But once the acceleration is over, wouldn't you know the final speed? Perhaps it's not accurate enough to calculate reliably (or I'm forgetting some fundamental physics).

 

[Julien]

But once the acceleration is over, wouldn't you know the final speed? Perhaps it's not accurate enough to calculate reliably (or I'm forgetting some fundamental physics).

Algorithms need lots of data points to smooth the data and average the points out. For instance in biking if you are still and speed up to 15MPH in 3 seconds that is easy to calculate. However it is only 1 data point. If you stay at that speed and then wipe your forehead then that is another data point. Since it only has 2 data points it could then calculate your speed as 30MPH even though you are still going 15MPH.

Of course the above is oversimplified (any experts feel free to correct me) but you probably get the gist.

 

[Cuyler]

Algorithms need lots of data points to smooth the data and average the points out. For instance in biking if you are still and speed up to 15MPH in 3 seconds that is easy to calculate. However it is only 1 data point. If you stay at that speed and then wipe your forehead then that is another data point. Since it only has 2 data points it could then calculate your speed as 30MPH even though you are still going 15MPH.

Of course the above is oversimplified (any experts feel free to correct me) but you probably get the gist.

Thanks. I asked the question like it was simple to answer. Calculating velocity is obviously complicated even with the addition of a GPS. I'm not sure I could truly understand the problems unless I spent a month working on the math. I'm just grateful the CoreLocation API takes care of all the internal details or a developer could go insane.

I did find something in the Apple Watch User Guide that hints on how the GPS is used during workouts:

Outdoor and Indoor Walk/Run/Cycle are distinct workouts because Apple Watch calculates the calorie burn differently for each. For indoor workouts, Apple Watch relies mainly on your heart rate readings for calorie estimates, but for outdoor workouts, Apple Watch works in conjunction with iPhone (which has GPS) to calculate speed and distance. Those values, along with your heart rate, are used to estimate the number of calories burned.

Also Matthew Miller has an article on ZDNet on running with the Apple Watch where he claims the watch is accurate enough to run without the phone (after initial calibration) — for whatever that's worth.