How accurate is the Apple Watch's pedometer?

[Saturn007]

I'm on the verge of buying my first Apple Watch. The major thing I'd use it for is fitness tracking, especially miles walked and steps taken.

How accurate —and how reliable — is the Watch in those departments?

I've recently compared two iPhone SEs (both the original model) and an iPod touch 6G and have gotten wildly varying results -- not only do they vary each time I use them -- they differ greatly from each other -- over the exact same path!

For example, I recently tested the two SEs three times on the same walk at the same time. (One SE in one pocket, the other SE in the other. 3 different days.)

One gave me results ranging from .78 mile to 1 mile, while the other varied from .92 to 1.2 miles. Steps on SE #1 ranged from 1,886 to 2,085 steps -- while those on SE #2 varied from 2,460 to 2,704.

As you can see, the phones are both unreliable and inaccurate as measures of distance and steps! The first varied a whopping 28% from low in high in distance; the second, 32%! Step variation was better, at 11% and 6%, but still high.

But they differed from each other substantially. The second one averaged 18% more in distance and a large 29% more in steps.

I'd have to dig up the IPT's measurements, but those differed from the iPhones' and varied greatly from walk to walk, too.

I'm hoping the Apple Watch would do substantially better than my iPhones! Does it?

Note that I'm concerned with two things: consistency and accuracy. In the social sciences, we refer to them as reliability and validity.

 

[wilberforce]

I'm on the verge of buying my first Apple Watch. The major thing I'd use it for is fitness tracking, especially miles walked and steps taken.

How accurate —and how reliable — is the Watch in those departments?

I've recently compared two iPhone SEs (both the original model) and an iPod touch 6G and have gotten wildly varying results -- not only do they vary each time I use them -- they differ greatly from each other -- over the exact same path!

For example, I recently tested the two SEs three times on the same walk at the same time. (One SE in one pocket, the other SE in the other. 3 different days.)

One gave me results ranging from .78 mile to 1 mile, while the other varied from .92 to 1.2 miles. Steps on SE #1 ranged from 1,886 to 2,085 steps -- while those on SE #2 varied from 2,460 to 2,704.

As you can see, the phones are both unreliable and inaccurate as measures of distance and steps! The first varied a whopping 28% from low in high in distance; the second, 32%! Step variation was better, at 11% and 6%, but still high.

But they differed from each other substantially. The second one averaged 18% more in distance and a large 29% more in steps.

I'd have to dig up the IPT's measurements, but those differed from the iPhones' and varied greatly from walk to walk, too.

I'm hoping the Apple Watch would do substantially better than my iPhones! Does it?

Note that I'm concerned with two things: consistency and accuracy. In the social sciences, we refer to them as reliability and validity.

See this:

Calibrate your Apple Watch for improved Workout and Activity accuracy - Apple Support

You can calibrate your Apple Watch to improve the accuracy of your distance, pace, and calorie measurements. Calibrating your watch can also help it learn your fitness level and stride, which improves accuracy when GPS is limited or unavailable.

support.apple.com

I find the watch distance tracking to be fairly accurate, comparing mine, my wife's and iPhone tracking. Maybe within 2%

 

[Night Spring]

I'm not as concerned about step accuracy as the OP -- I don't really check my step counts *that* closely. I do have a pedometer app where I set a goal of 5000 steps a day, and I've met or exceeded that goal every day for the past year. I just checked my steps data for the last year, and it says my average is 5600 steps a day. So that jives with the way I've been meeting my daily step goals. Not sure if this info is helpful or not, but my impression is that the watch step count is consistent over a span of months/years, but it could give inconsistent results over a small sample, like 3 walks.

 

[Mr. Awesome]

I'm on the verge of buying my first Apple Watch. The major thing I'd use it for is fitness tracking, especially miles walked and steps taken.

How accurate —and how reliable — is the Watch in those departments?

I've recently compared two iPhone SEs (both the original model) and an iPod touch 6G and have gotten wildly varying results -- not only do they vary each time I use them -- they differ greatly from each other -- over the exact same path!

For example, I recently tested the two SEs three times on the same walk at the same time. (One SE in one pocket, the other SE in the other. 3 different days.)

One gave me results ranging from .78 mile to 1 mile, while the other varied from .92 to 1.2 miles. Steps on SE #1 ranged from 1,886 to 2,085 steps -- while those on SE #2 varied from 2,460 to 2,704.

As you can see, the phones are both unreliable and inaccurate as measures of distance and steps! The first varied a whopping 28% from low in high in distance; the second, 32%! Step variation was better, at 11% and 6%, but still high.

But they differed from each other substantially. The second one averaged 18% more in distance and a large 29% more in steps.

I'd have to dig up the IPT's measurements, but those differed from the iPhones' and varied greatly from walk to walk, too.

I'm hoping the Apple Watch would do substantially better than my iPhones! Does it?

Note that I'm concerned with two things: consistency and accuracy. In the social sciences, we refer to them as reliability and validity.

The Apple Watch seems to be very reliable, because it combines data from the Watch and the phone.

 

[michaelb5000]

I find the apple watch to be accurate, certainly enough for my purposes. But I think you are talking about two different things, and I am not sure they are directly related: miles and steps. So I think "miles" is determined by the GPS; there are lots of professional reviews of the watch GPS on runs; but any GPS tracking can produce errors in overall distance, particularly if you are rounding corners. When I regularly walk on a pretty straight road in open country, I hit a mile where I expect it. When I bike, I have an old Garmin Vista GPS unit on my bars (it has the high sensitivity GPS receiver) but I also often turn on biking as an activity on the watch. There are always slight variations between the two, but we regularly do a 10 mile ride that consists of riding to a point and then turning around; that 5 mile turn around point on the watch and the garmin is always the same place and usually within 50 feet of each other (I realize that I usually have my phone with me on the bike, but I have done this same ride with just the watch; I should do that again though and compare with the new watch).

Steps is a different metric, and different devices are going to track and consider a step differently. Yes, you have to calibrate the watch for tracking steps (easy to do). But the watch will count steps cooking in the kitchen and moving around the house, ie many movements during the day count as steps. The refresh rate for seeing steps can vary, and there is no built in display or default app to view steps. I use Pedometer+ to track steps and have for 5 years, so since the S0. Pedometer+ merges step data between the watch and the phone and does a nice job keeping track of the overall total. I average 2000 steps per mile. But I am not going to count to 2000. Back with the S0, when I was starting, I would count off 200 steps or so, and then stand still and wait for Pedometer to update, and it was usually pretty close to correct.

 

[Saturn007]

Definitely appreciate the feedback. It's lead to further thoughts.

Awesome: “The Apple Watch seems to be very reliable, because it combines data from the Watch and the phone.”

Does that mean you have to have the phone with you for it to track distance and steps?

Hope not! I was looking forward to having the watch on and being able to take a walk and get some accurate distance and step measurements without having my phone with me. From what Wilber mentioned, it sounds as if the watch doesn't even report steps on its own! “there is no built in display or default app to view steps.”

Also, it seems that distances come from the GPS and can be affected by corners. My test walk, however, has only one corner in it. Would not have expected GPS measurements to vary by 28-32% from one walk to another! That seems way off.

Night, can you look back at your step data and see how much variation there was? In other words, low to high? Of course, that's not over a set course!

Michael, glad to hear that your phone matches your Garmin. Which phone is it? Maybe the GPS sensors aren't as good in the original iPhone SE. Wonder how well a Series 5 or 6 watch would do in comparison.

Wilber, that's impressive that the watch and your two household phones are within 2%!
Which phones are you using? How variable are the results? In others, how consistent are they?

As I mentioned, accuracy and consistency are two different aspects of this.

Does anyone know of a tech site that might have done formal testing of the watch's accuracy and consistency?

Or, determined how great the discrepancies are between different Apple Watches?

Maybe it's a quixotic quest to use a watch for fitness, if I can get the same data with an iPhone! (Although in my case, it's highly suspect data!)

 

[Night Spring]

Does that mean you have to have the phone with you for it to track distance and steps?

Oh, no, you definitely don't need your phone. if you just have the watch and no phone, it tracks your steps using the watch. If for some reason, you walk around with just your phone and no watch, it counts your steps using the phone. When you have both devices with you, I believe there's a setting somewhere that tells it which device to use. Then all steps from the phone and watch are totaled to give your daily step count.

Night, can you look back at your step data and see how much variation there was? In other words, low to high? Of course, that's not over a set course!

I don't know what I can tell you. It's not like I walk a set distance every day! All I have is a minimum goal of 5000 steps a day. If I go shopping downtown and walk around a lot, I could hit 8000-9000 steps, and if I just spend the day at home, I walk up and down my hallway until I hit 5000. I don't have any way to match up these step counts with independent measurements of what distance I actually walked. I can tell you that walking to the nearest subway stop is about 500-600 steps. I can't get any more granular than that, because it doesn't bother me if one day it's 500 steps and the next day it's 650. I just kind of ignore any digits smaller than the 100s. I'm satisfied if it isn't 500 one day and 800 the next.

I just have the feeling you are overthinking step count a bit. If you are running a marathon or otherwise focusing on running/walking a set distance daily, you use the gps to measure the distance, and step counts aren't that important. If you are using step counts to motivate yourself to move around more, then distance isn't that important. That's how I see it, and that's why I'm not so concerned about the correlation between step count and distance.

 

[wilberforce]

The watch continuously tracks steps whether with the phone or without, and you can check it at any time by opening the activity app on the watch. However there isn't a standard complication that shows the current step count on the face directly.

An app like WorkOutDoors will monitor and display steps (and cadence, stride, track on a map, etc, etc) for a specific workout, like outdoor walking. I highly recommend WorkOutDoors. It is worth the cost ten times over for anyone doing any outdoor activity.

 

[taneff]

The watch is ok regarding step accuracy, but nothing more. I collect steps data for a decade now, almost addicted to steps, 10000 minimum daily, most times 15000 steps and more. Starting with a Fitbit One tracker on my hip etc. So I compared many bands and watches to my Fitbit One over the years. I wear Garmin for many years now, which are definitely the most accurate, taking in consideration that it is still a band, counting steps while moving just the arm too. Also comparing it while hiking etc. to other watches from friends, Garmin and Samsung watches are the most accurate. Apple is ok but not as accurate. In general, I observed the following over many years of wearing and comparing many devices, most of all Fitbit, Garmin and AW, all bands compared to a tracker on the hips. Watches were always worn on my non dominant left hand when compared. When having around 10000 steps on my "Fitbit on the hips tracker" my Garmin had around 11.000, AW 12-13000 and Fitbit bands around 13000-15000. Polar had almost 200%, like 20000 steps, ridiculous. So I would say AW is ok with about 10-20% more steps than really done but there are more accurate steps tracker.
Sometimes I wear my quite accurate Garmin on the right, dominant, hand and my AW on the left, not dominant hand, when working and moving and I still get the same steps, even though I move my right hand way more. So if you want the most accurate step tracker, look for a Garmin. But in general AW is ok, there are way worse step accuracy trackers.

Plus you should wear it on your NON DOMINANT arm/wrist. Otherwise steps are sometimes 10-20% off.

PS: The opposite is true for steps during a workout with moving arms. I do a lot of rebounding, including dumbbells etc. And after a 1h workout my AW always has way less steps than my Garmin. But in everyday step tracking Garmin is way more accurate.