Your Fitness Tracker Is Probably Wrong (And Here's the Proof)

Your fitness activity tracker probably isn't fully trustworthy and the science backs it up.

In this episode, we dig into the research on fitness activity tracker (smart watch) accuracy, diving deep into the research especially for the Apple Watch, Fitbit, Garmin and Oura Ring.

From step counts and heart rate monitoring to calorie burn and sleep tracking, we break down what the data actually says, and the results might surprise you.

Whether you're training for a race, trying to hit 10,000 steps, or just wondering if your tracker is giving you the full picture, this episode is for you.
What we cover:

- How accurate are Apple Watch, Fitbit, Garmin and Oura Ring really?
- Which metrics are most (and least) reliable.
- What the latest research says about wearable accuracy.
- Should you trust your tracker's calorie and heart rate data? If not, how big is the margin of error?

Perhaps most importantly, does it matter if smart watchers and fitness activity trackers aren't fully accurate? Maybe that doesn't matter.

P.S., if you enjoy these free podcast episodes and want to support us, there are two main ways of doing so. 

If you would like monthly lifting workouts, you can join the Momentum by Sohee fitness app. http://momentumbysohee.com

If you are interested in fat loss science, you can purchase Ben's best-selling comprehensive fat loss book, Everything Fat Loss. http://geni.us/EverythingFatLoss

Transcript

SPEAKER_00 0:00

Can you trust your activity tracker?

SPEAKER_02 0:02

Probably not. Or at least not blindly. Yeah.

SPEAKER_00 0:06

First things first, what is your current step count?

SPEAKER_02 0:09

Great question. How many steps have I done today?

unknown 0:13

16,062 steps.

SPEAKER_02 0:15

16,000 and 62 steps.

SPEAKER_00 0:17

You've had a very active day. Let's write that down. He is on an Apple Watch. I have a Garmin. Now keep in mind today's been a rest day.

SPEAKER_01 0:24

Yeah.

SPEAKER_00 0:24

Today's a rest day, okay?

SPEAKER_01 0:26

Yeah.

SPEAKER_00 0:26

My steps are at 2555.

SPEAKER_01 0:29

Okay.

SPEAKER_00 0:29

Okay, so we're gonna mark this as baseline data, and we'll come back at the end. Why are we interested in doing an episode on activity tracker?

SPEAKER_02 0:37

So this came from a flippant comment that I said in an earlier episode, which was I was in a long car journey, and when I arrived at my destination, it told me that I had hit my step goal, apart from I clearly wasn't walking anywhere.

SPEAKER_00 0:51

Yes. And also when I was driving, I think up a hill, it said I hit my stairclimbing goal for the day.

SPEAKER_02 0:56

And I I made a joke that you were hitting too many curbs if it thought that you were climbing stairs.

SPEAKER_00 1:00

And then it reminded us that we'd both done posts in the past about activity tracker accuracy overall. And then of course, since then we've gotten some comments from people saying actually an episode would be very helpful.

SPEAKER_02 1:10

That is the important bit. People have asked us for this episode. So Bish Bash Bosh. Alright. You asked for it, we're delivering. Here's your episode.

SPEAKER_00 1:20

Tell us your heart rate data anecdote.

SPEAKER_02 1:22

I actually have a couple of images I can put on the screen. If you're listening on Spotify, I think you can actually switch to the video, perhaps. But I have a couple of race heart rate or run heart rate line graphs. And as you can see from one image, my heart rate slowly climbs, slowly climbs, slowly climbs, as you would expect, and then plummets. And then apparently stays exactly the same for the remaining last mile, which seems unlikely.

SPEAKER_00 1:47

Also, usually if you're actually in race mode, it should stay climbing.

SPEAKER_02 1:51

Uh that one I think was a regular run, but I actually do have a race. I have a half marathon line graph, and for some reason, my heart rate stays exactly at 170 beats per minute for approximately 10, maybe 11 miles. Right. Which is obviously not going to happen. Now, this is quite clearly a heart rate anomaly. Something has happened somewhere where it's stopped picking up the data and whatever. We all know that that is not true. And even though this is a kind of silly example because maybe it was just too sweaty, maybe it wasn't making contact with my wrist, whatever it was, this is actually indicative of where errors can come in because heart rate is also used to then calculate other things that smartwatches do. So if heart rate isn't always accurate, then it also means that other things aren't always accurate as well.

SPEAKER_00 2:45

So we have our baseline step counts for this episode in the beginning. What I'm also gonna do is if you're on video, you'll see me intermittently throughout the episode swinging my left arm, which is where my garment is attached to.

SPEAKER_02 2:58

You need like a protein shaker or something.

SPEAKER_00 3:01

So if you see me doing that, there is a purpose, but you can go ahead and ignore. And what we're gonna do is revisit the arms, the step count on my Garmin at the end of the episode. Um clearly, I'm staying put. I've actually log zero extra steps. I will have log zero extra steps. We'll see what it actually tells me. Um my step count has updated two by the end.

SPEAKER_01 3:20

Yeah.

SPEAKER_00 3:21

Okay, so let's go over the popularity of activity trackers.

SPEAKER_02 3:26

Yeah, so even though this was requested by I don't know how many people. Some people, it was requested by some people, it is also very important because fitness activity trackers have been described as the biggest fitness trend for 2026 by the American College of Sports Medicine. So this is a very, very popular topic. It makes sense that we're going to talk about a popular fitness topic. It has also been estimated, there are different estimates because it's very hard to know for sure how many people own a fitness tracker. You can kind of survey a couple of thousand people and then like try and predict the whole nation, but that's you know difficult. So a couple of different estimates. One is that 39% of American adults own a wearable activity tracker. And the arm thing's really gonna No, no, I I like it. I was just wondering what's happening. It looks like you had like an itch or something. Uh the market is projected to reach a couple of different estimates, but 186 billion by 2030, and another estimate was 352 billion by 2020 2033.

SPEAKER_00 4:33

So a lot of money being spent on activity trackers.

SPEAKER_02 4:36

Yes.

SPEAKER_00 4:37

Which makes sense. We both have activity trackers of ourselves ourselves. They cost probably most of them the nice ones cost several hundred dollars each.

SPEAKER_02 4:46

Yeah, I I would imagine like the low end you're probably looking, what I would guess closer to a hundred dollars for the low end. Like, like a lot of the fitness trackers started very, very basic where people were using them for step count. Yeah. And then it was like step count and calorie burn and sometimes speed and distance or whatever, and they've gotten more and more advanced.

SPEAKER_00 5:04

So sophisticated, uh sophisticated. Sophisticated. Yeah. My reason for because I actually did have an Apple Watch. I switched to Garmin. I felt like it was giving me more data and better data for my running, um, which is the reason for my switch. But we'll be looking into those brands in a little bit.

SPEAKER_02 5:20

I haven't done that because I in my head I still don't take running seriously enough to switch to a dedicated fitness watch, even though you have told me that yours is amazing. Yeah. I can't do it. Also, I like the walkie-talkie function.

SPEAKER_00 5:32

Walkie-talkie function is nice, and the find my iPhone function I definitely miss from the Apple Watch too. All right. Before we dive into what the research says overall, let's go into what are the pros and cons of activity trackers. Because obviously, this episode is not about poo-pooing on all wearable devices. We obviously are fans of them, but it is important to understand what they're good for and what they're maybe not so great for.

SPEAKER_01 5:57

Yes.

SPEAKER_00 5:58

Okay. I will start with the pros. So obviously they are generally unobtrusive. Um this is from a research paper. It says cost effective. Depends on obviously the brand and the model and stuff. Um, but they're also very comfortable to wear. They don't really get in the way of your day-to-day life.

SPEAKER_02 6:14

So cost-effective for reference, because saying that a device that costs a couple of hundred dollars up is cost-effective sounds a bit tone-deaf. They will often say it's cost-effective relative to a more so, for example, if you if you want heart rate data, sure. Getting an ECG at home, sure, unlikely, or getting like a sophisticated sleep study, very unlikely. That's fair. Okay. So it's but it's supposed to be cost effective relative to something.

SPEAKER_00 6:41

Yeah. Also, it can help generate insight generate insights into your individual health pattern. So it's helpful to see trends, for example, um, which you and I both we like to discuss, like how much sleep did it did your device say you got last night and whatever, those things like that. Um, I do think that, and a lot of people have told us this as well, wearing a device makes them feel more motivated to want to be more active throughout the day. So it's kind of like, okay, you know a device is tracking you. I'm gonna get up and walk more.

SPEAKER_02 7:13

And you have mentioned that on a previous episode where you said it got to 5 p.m., you had only walked 500 steps all day because you had been sitting working on your laptop, etc. You're like, ooh, and then you're like, oh shit, maybe I'll do a little bit more walking. Yeah, like yeah. One pro that I want to add, and I've I've deliberately picked something that's a little bit more niche because people know that they you have step counts, etc. But there is growing technology for medical purposes. Correct. So, for example, um things like atrial fibrillation, uh collecting data on markers for Alzheimer's disease, there is new technology where it is going to pair with blood glucose readings, so people can see their blood glucose readings on their watch. There is also things like detecting falls.

SPEAKER_01 7:56

Yep.

SPEAKER_02 7:56

So, for example, if you put a watch on your grandmother or granddad and it can detect a fall, it could then alert someone. So there is an increased level of sophisticated technology that is going to make these better and better and better. And if you go back 10 years, a lot of these functions just didn't exist. Right. So I think it's actually really exciting to see where they're going.

SPEAKER_00 8:17

Yeah. And I do think that the increased data that we have capture is going to become increasingly helpful, of course, as long as the data is actually accurate.

SPEAKER_02 8:29

That is a big thing. Because obviously, if the data are not, data being plural, if the data are not accurate, then how much you can actually conclude from it is redundant. Sure. And that is one of the big issues. And part of the reason that I mentioned my heart rate at the beginning is because heart rate is used as a basis for other measurements. Right. And if you can't rely on heart rate data, then can you rely on VO2 Max, which comes from heart rate data, amongst other things? For example. Right. So to go over some of the cons, and these are pulled from different research papers.

SPEAKER_00 9:00

And/or limitations.

SPEAKER_02 9:01

One research paper basically said, are these sophisticated or are they sophisticated marketing? Which I thought was a like fairly aggressive title. And here are some things that they stated. So, number one, are calories burned from physical activity additional to what you do, or are they inclusive? So, for example, if I go for a run and my watch says, You have burned 300 calories, is that 300 calorie number additional to what I would have already burned?

SPEAKER_00 9:30

Or if you had been at rest, basically.

SPEAKER_02 9:33

Or is it inclusive of how many calories I burn total? Anyway. Like your body is burning energy all the time, whether you're stationary or not. And even though that seems like a very basic question, it said, is this even made clear to the user? So if someone has burned 300 calories, is this on top of what they are normally doing, or is this inclusive of?

SPEAKER_01 9:51

Yeah.

SPEAKER_02 9:52

A kind of basic thing to consider. Then we get a little bit more complicated. So say energy expenditure. If two people weigh the same, me and a twin both weigh the same, uh the measurements that it will give you for energy expenditure, if they are based on body weight, would be the same. But there are other things that will impact body composition. So, for example, if one of us has more fat mass and less aline body mass or vice versa, the number of calories that we would burn would be expected to be different. Some people burn naturally more energy than others, and it might be a couple of hundred calories per day or whatever.

SPEAKER_01 10:32

Yeah.

SPEAKER_02 10:32

But your watch can't take that into account. It doesn't know how much you weigh unless you put it in. It then doesn't know your body composition. So it's it's kind of coming up with calculations based on assumptions. And because it's based on assumptions, every time there is an assumption, there is room for error.

SPEAKER_00 10:49

Yeah. Another limitation or a con is I'll read it off here. As VO2 max tends to be estimated based on algorithms that factor in your heart rate, as you mentioned. Uh if an activity tracker is not good at measuring heart rate, then it's unlikely to be good at measuring VO2 max. And another example of this would be uh I think training readiness, for example, which I know Garmin is.

SPEAKER_02 11:10

Training readiness is a big one.

SPEAKER_00 11:12

Yeah, training readiness is a score, at least I know that Garmin gives you training readiness score. So if heart rate accuracy is not great, then your trading readiness score is also not going to be great.

SPEAKER_02 11:21

Yeah. Same with like say sleep scores or sometimes stress scores. If things like this use heart rate data, amongst other things, to spit out a number at you or a percentage or whatever, yeah. The initial data has to be accurate, otherwise you get inaccuracies all the way along the chain.

SPEAKER_00 11:39

Okay, so what are some questions or considerations about accuracy of activity trackers?

SPEAKER_02 11:46

So there are various things that impact the accuracy. So some common examples, uh, say for example, the temperature or humidity. So environments of environmental factors kind of make sense. Like if someone is sweating more, maybe that will uh uh affect conduction.

SPEAKER_01 12:04

Yeah.

SPEAKER_02 12:05

If someone has a darker skin tone, that can also affect the measurement. And a couple of research papers actually acknowledge that there's a research paper on this that said there isn't enough research to say definitively how much skin tones will impact numbers, but it does look like darker skin tones will have different numbers than lighter skin tones.

SPEAKER_00 12:26

Yeah. But device placement is also a factor, difference between placing it on your wrist versus maybe your ankle, something like that.

SPEAKER_02 12:34

Yes. Another one that we will probably circle back to later, but one of the things with smartwatches is let's say, for example, there is a research study on an Apple Watch and they use the current version of the Apple Watch. If Apple release a software update and they change the algorithm behind the scenes, that would then impact the validity and the accuracy of the information that it's giving you, but you don't necessarily know that, and researchers don't necessarily know that. So, for example, if we let's say Ben and Sohi, or Sohi and Ben, I like giving my wife's name first, if we launch our own activity tracker today and we do a study on it, and we say, Oh, this is actually really, really accurate, you should buy it. But if we change the software behind the scenes, we can influence the act the accuracy of that.

SPEAKER_01 13:23

Sure.

SPEAKER_02 13:23

But every time you change something behind the scenes, in theory, you would need more research to validate that it's actually accurate. Right. That is an issue.

SPEAKER_00 13:31

So when you consider the need for the funding and the ethics approval and the pup the whole publication process, which can take a long time, oftentimes it it may be the case that a certain uh brand and and and model of activity tracker is like several iterations ahead by the time a certain study comes out for a what was at the time a current uh update.

SPEAKER_02 13:56

This is this is going to be a very, very big theme, but there's a quote from here, it's not particularly long, but I just want to kind of drive this point home. So this is straight from a research paper. The pace of academic research struggles to keep up with the more agile commercial ecosystem. Primary research studies are slowed by the need to secure funding, develop validation protocols, recruit and test participants, and navigate the peer review process. While meta uh while systematic reviews and meta-analyses are often out of date by the time they are published. Right. So what that means is if Apple launch a brand new watch right now and researchers want to test it, they have to get funding, they have to recruit people, they have to develop the study, then they have to do the testing, then they publish the data. By the time the data is published, chances are a new Apple Watch is out, which actually means that they are constantly kind of playing whacking. You're always behind.

SPEAKER_00 14:48

Yeah, the research is always going to be behind.

SPEAKER_02 14:50

And I've this has happened before when I've talked about research before. If I say they study these devices, people will always always say, but that's an old device. Right. That is literally how research works.

SPEAKER_00 15:00

That's all we have so far.

SPEAKER_02 15:01

And in some ways it's kind of interesting because it means that manufacturers don't have to publish or invalidation research. No, they don't have to say, we've launched a new watch, we have tested it to show how accurate it is. It's for researchers to then independently test it to see whether it's accurate.

SPEAKER_00 15:16

To me, I think obviously I'm sure there's a lot of logistical barriers to why they don't do it the other way around. But in in my head, it would make more sense to put the onus on the brand, the companies, to provide validation. Of course, that would pu that would provide a huge bottleneck to actually getting things these things out on the market because obviously their priority is making pro you know, selling devices and making profit. So I get it, but obviously for us when it comes to accuracy, it can make things a bit trickier.

SPEAKER_01 15:44

Yes.

SPEAKER_00 15:45

Question for you how are activity trackers validated?

SPEAKER_02 15:49

So I have a list of things that activity trackers measure. This is a non-exhaustive list, but I just want to throw it out there because most people aren't using the number of functions that are available. So for example, heart rate and heart rate variability, most people will know about heart rate. Step count, again, very, very common. Distance, again, very, very common. It might say you have walked X number of steps per day. Energy expenditure, or the number of calories you burn per day. Most people know what that is. Blood oxygen is not across devices, but that is something that mine tells me, I think. Uh sleep duration and sleep stages. So not just how many hours you've been to sleep. We can both look on our watch or look on our phone and it will say, like, you had this many minutes of REM sleep or deep sleep or whatever. Uh skin temperature, respiratory rate, uh sound exposure. Mine, if I go to a concert, will be like stop the bus, you are deafening yourself. Yeah. Uh but also interestingly, it sometimes does that if I wear it in the shower. Just the sound of the water hitting it is like the shower? Uh I think it's waterproof or water resistant. I don't, but some sometimes I do sometimes I do, and it will tell me that it's too loud. And I'm like, that's picking up on noise here, not noise here. No one is deaf from having a shower.

SPEAKER_00 17:05

Has it done it with my voice around you?

SPEAKER_02 17:07

Uh no. If it was near you, it definitely would. It definitely would. You you speak louder than me. By a lot.

SPEAKER_00 17:13

Yeah.

SPEAKER_02 17:14

VO2 max, uh AFib detection, and ovulation detection.

SPEAKER_00 17:18

Yeah, very interesting.

SPEAKER_02 17:20

Which we are going to circle back to later because I like a nod to our female audience and if there's any specific research that is dedicated to females.

SPEAKER_00 17:28

And then for some of these metrics, how are they validated?

SPEAKER_02 17:31

Okay, so what we need people to understand is let's say we launch a device and it measures step count. To know whether step count is accurate, you have to validate it against a more perfect measure. Yeah. So for example, we could conduct a research study, or actually not us, we're the manufacturer. We don't bother doing our own research study. That's a small dig. Someone else has to study our device, and they might recruit 100 participants or more likely 30 participants, and you say, You're going to walk for five minutes wearing this device, and then while you're doing it, we're filming you, and we will watch the footage back. So you might have a couple of researchers that watch the footage independently, and they will say, She has walked this many steps. Yeah. The watch said she walked this many steps. Yeah. There is an error of five percent between those two. So basically, you're validating it against a more accurate test. Yeah. So, for example, step count and also wheelchair push, you would film it, watch it back, validate it against that. Heart rate you would compare to something like ECG, uh, which most people know. If you went to the hospital and they were monitoring your heart rate, they wouldn't put an activity tracker on you, they would put electrodes on your chest. Uh, also, sometimes they will use a chest strap, which is a more commercially available device people can buy relatively cheap because those are known to be more accurate, generally speaking, than wrist trackers, which is why you will often see some athletes wearing a wristwatch and a chest strap. So they would validate your heart rate from the watch against an ECG device or a chest strap. Energy expenditure, we won't go into this too much because this is probably too nerdy for most of our audience, but they would compare your calories burned to direct or indirect calorimetry or doubly label labelled water, which is like the gold standard. If you want to measure people's calorie burn in research studies, you would use one of those methods. You would never ever ever just use an activity tracker. Sleep, you would compare to polysomnography. So, like if you went and did a sleep study, they would have sophisticated technology. Uh, and then blood oxygen, you would use pulse oximetry. Again, if you went to a hospital, you would have a different device. So you would measure it against a more gold standard and then compare the differences between the two.

SPEAKER_00 19:50

Cool. Now, with that covered, let's move on to the next section, which is what does the research say on device accuracy across some of the biggest brands? Now let's go over the research that gives us an overview of everything, and then we can break it down by brand.

SPEAKER_02 20:08

Okay, so I have kind of cherry-picked, and I've cherry-picked for a slightly obscure reason. I've talked about this study before, and the reason I've talked about this study before is because back in the infographic days, I liked I literally liked the image. Yeah. But it's because it's easier to teach some people things if they have a visual. Sure. So there's a research study from 2017. I love the image. A lot of the research studies, the like the validation studies, started a few years earlier than this. Yeah. But in 2017, there was a study that tested various devices. Brands including Apple, Fitbit, Microsoft, and Samsung, which already like Microsoft You're like, oh, Microsoft had an activity tracker?

SPEAKER_00 20:50

Times have evolved, obviously.

SPEAKER_02 20:51

And the reason that I like the image is because if you look at this, you will see that they are testing different activities.

SPEAKER_01 20:58

Yeah.

SPEAKER_02 20:58

And the reason why this is important to set the tone for the episode is because if you're measuring, say, heart rate, if we have 20 participants in our research study, we don't just say go for a walk and we'll measure your heart rate. You would also say go for a cycle and measure your heart rate. Right. Go for a run or a jog or whatever and you do different activities. And the reason you need to understand this is because sometimes the activity changes the accuracy. Correct. So if we say that this device has a 5% margin of error on heart rate, it's actually not that you're like, for what?

SPEAKER_00 21:37

Right.

SPEAKER_02 21:38

So sometimes it's like you have a 5% error when you're resting, but you might have a 10% error when you're walking or a 15% error when you're running. So this kind of breaks it down by different activities rather than just heart rate, step count, etc.

SPEAKER_00 21:55

Not to mention, which we'll get into more later as well, it's not only about the activity you're doing, but Also things like intensity, speed, things like that, which will then infl incline for walking things like that.

SPEAKER_02 22:07

If you are if you're looking on video, the picture will kind of paint a thousand words, but I will describe this for people who are listening to audio only. So the heart rate tended to be much, much more accurate than energy expenditure. So generally speaking, heart rate data from wrist devices is a lot more accurate than how many calories you've burned. So heart rate uh heart rate error rates range from 1.8% cycling, which is very, very low, to 5.5% walking. Uh, six of the seven devices had a median error rate at or below five percent for cycling, which is five percent is like a tight margin of error. Very tight margin of error. Someone's like 60 beats per minute, that's like the difference between 57 and 63, like very, very tight. Energy expenditure, on the other hand, this is where things get messy. Uh, it said that no device achieved an error in energy expenditure below 20%. And if normally they would say, for example, let's say a 10% margin of error is acceptable, if nothing is under 10%, that's literally like saying we cannot trust any of these devices. Just full stop. So for heart rate, things tended to be pretty good, but for energy expenditure, things were an absolute shit show.

SPEAKER_00 23:22

Yeah. So again, now keep in mind this study is from 2017, right? Which is almost 10 years ago now, which hopefully that means, and we'll find out soon, that the accuracy has gotten better since then.

SPEAKER_02 23:35

Yes. We will we'll move on to other data. Now, a couple of things that I just thought were interesting to help kind of reiterate to the audience. One, device error rate increased for males, which I don't know if that's a running theme. Like that surprised me when I read it. I didn't know that that would be a thing. People with higher body mass and darker skin tone, which is something that we had mentioned earlier. And then a quote just because I think it's a very damning quote, even though I've said it already, and I quote None of the devices provided estimates of energy expenditure that were within an acceptable range in any setting. You can't trust them. Full stop was the the conclusion.

SPEAKER_00 24:13

Yeah. Okay, let's talk about this really cool umbrella review that came out in 2024. Yes. Which is a live study. Now, I had never even heard of the term live study until we were prepping for this episode.

SPEAKER_02 24:26

I'm it's pretty cool. I'm excited by this. So obviously you've got to this point in the episode, you understand the limitations with research, you understand that by the time devices are studied, chances are those devices are already old. Right. If we study a device now, the device might not even be sold by the time the research study is published. Right. And I read a research paper where I can't remember the date, so I'll kind of make it up. But if it was published in 2021, they bought devices that were available in 2019. Because they bought devices, then they conducted the study, then they did the number crunching or the publishing or the peer review, blah, blah, blah, blah, blah. So by the time it was published, it's literally out of date. Yeah. So the thing that I liked about this is I found an umbrella review and I actually emailed the lead researcher. I might be able to update this later if I ever speak to the lead researcher. So think of it like this: you get 50 participants to do one study to test five devices. Then above that, you have like a systematic review or a meta-analysis, which is when you look at, say, 10 or 15 studies or whatever the number is, you put them together and you do an analysis of multiple different studies, and then you publish like a systematic review. So systematic reviews tend to be more trustworthy because you look at multiple studies at once. Then above that, you have an umbrella review, and an umbrella review will look at all of those systematic studies. And this umbrella review was described as a live review, meaning that due to the nature of how quickly these research studies are outdated, they will essentially update this every six months. So they will say if there is new research, it will be updated. If there's new research, it will be updated. So even though this is from 2024, the fact that it is updated periodically means that this is likely to be the best research study that we can rely on, which is part of the reason I emailed the lead researcher because I think it'd be cool to get some inner workings of how it went.

SPEAKER_00 26:25

What are some of the important findings from this umbrella review?

SPEAKER_02 26:31

Okay, let's do these like alternate. Okay. Uh so just to give you an idea of the size, they looked at 24 systematic reviews that totaled 430,000 participants. That's massive. This isn't just we looked at 15 people who are walking on a treadmill. This is as big as you're going to get.

SPEAKER_01 26:50

Yeah.

SPEAKER_02 26:51

And what we are going to do after this is we're going to talk about devices and brands one by one, the most popular ones, because chances are, if you're watching this, you're thinking, or if I wear an Apple Watch, I wear a Garmin, I wear whatever. I want to know how that fed.

SPEAKER_01 27:05

Yeah.

SPEAKER_02 27:05

To begin with, we're just going to give you an overall summary from this umbrella review. So 430,000 participants, they have 310 devices. I'm not going to list all the brands, but it included like Apple, uh, Coros, Fitbit, Garmin, Google, Huawei, Jawbone, Samsung, blah, blah, blah, blah, blah.

SPEAKER_00 27:24

Yeah. Uh, and then of the devices that they looked at, they found that only 11% of the 310 consumer wearables have been validated for at least one biometric outcome.

SPEAKER_02 27:39

Basically meaning that the majority of biometrics that your smartwatch is telling you about have not been validated. Like the vast, vast, vast majority have not been validated in independent research. And even when a device has been validated, it's often not for all of the functions that it's telling you. So you might say, people watching this, if you're watching, you're listening, you're smart because you're listening to us and you care about research. Other people who are less informed will say, I think mine is accurate. That's not how that works. If you s if someone says that, they're already wrong because nothing is accurate across the board. And if it was, you would need to validate it. Yeah. And that just doesn't exist.

SPEAKER_00 28:18

Yeah. Do you want to read the next one?

SPEAKER_02 28:20

Let's go into individual biometrics and things that they measure. So one of the quotes was consumer wearables appear moderately proficient in capturing various health outcomes such as heart rate, heart rate variability, aerobic capacity, and others. So generally speaking, moderately proficient. But if we break this down by individual biometrics, now bear in mind that this is looking at all of the research devices and pooling them together. So something might be really, really inaccurate, but if something is inaccurate in the other direction, technically they're kind of averaging out. So this is the picture. Okay. So heart rate, the mean absolute percentage error was around 3%, which is similar to what we said earlier, under 5%, very, very, very small. That's pretty good. Yeah. Very small.

SPEAKER_00 29:07

Step count, mean average percentage error range from minus 9% to plus 12%. So quite a bit more variability there.

SPEAKER_02 29:17

But you've got to keep in mind that that is actually testing it where you're only walking. If you are sitting and moving around and fidgeting, sometimes it can erroneously pick up on steps. So in like a free living situation, it is very, very common for devices to then get errors. Hence sometimes it thinking that we have walked when we've been in the car, for example. Right. So I think it's not surprising that step count is less precise than heart rate.

SPEAKER_00 29:42

Uh VO2 max overestimates by 15.15% during resting tests and 10% during exercise tests.

SPEAKER_01 29:51

Yes.

SPEAKER_00 29:52

Which I believe because my Garmin, when I was at the peak of my marathon training and everything, told me my VO2 max was like 52. Um, but when I compare it to friends of mine who have actually gotten a proper VO2 max test in a lab setting, and I know they're faster than me, their VO2 max readings have been lower than mine. And I'm like, okay, so I think it's overestimating.

SPEAKER_02 30:15

Uh I'm gonna do the next couple because they're related. Total sleep time overestimated by more than 10%. If it says that you slept for eight hours, there might be a 10% discrepancy there. However, markers of sleep like sleep onset latency and wakefulness after sleep, that is where much, much bigger inconsistencies were. And errors range from 12% to 180%. Yeah. 180. So what you need to know there is a device might say you have slept for eight hours, and it might have a say 10% margin of error if you want. Yeah. However, devices tend to be worse at differentiating between sleep stages rather than just total duration.

SPEAKER_00 30:59

Also, anecdotally, I want to add on that I fully believe this because especially with my sleep quality as bad as it is lately, being in the third trimester of my pregnancy, I uh almost every night I have at least an hour when I'm awake during the night. Um it does not oftentimes it when I wake up, let's say like 5 a.m. Um and I happen to fall back asleep at like six six a.m., it will not capture the sleep from 6 a.m. onwards.

SPEAKER_02 31:26

I have also noticed this myself where if I have woken up and I've gone for a pee, and if I have then lay back in bed and I can't get to sleep for ages, when I look at my sleep in the morning, it will often think that I've gone to sleep. So it it's if you lie very, very still and you're resting, sometimes it will accidentally assume that you're asleep when you're not. Right. And that's when it can be slightly harder to differentiate like mid-stage rather than when did you fall asleep? When did you wake up? Yeah. When you wake up is fairly self-explanatory on a watch because you start moving. Right. But if you wake up and you stay perfectly still and you're like half dozing, watches are more likely to be inaccurate.

SPEAKER_00 32:07

Also, when I nap, it doesn't always capture that unless I manually prompt like prompt track my nap. But the other problem with that too is it's not like I hit I'm napping now and I'm instantly asleep either. You can still take me 15, 30 minutes to fall asleep.

SPEAKER_02 32:21

So, you know, if you if you were staying perfectly still watching a film or staying perfectly still lying in bed, like which one you're napping in one case and not the other, but harder for watches to pick up. Yeah.

SPEAKER_00 32:31

And then the last metric was energy expenditure, which was underestimated by around 3%, with error ranges from between negative 21% to plus 15%.

SPEAKER_02 32:43

Now the reason that we've included ranges, and this is very, very important, is because as an average, if they were underestimated by 3%, again, like I explained earlier, if something overestimated but something else underestimated, you look at the two and the average actually looks fairly accurate. So it isn't necessarily about whether something is accurate as an average. You want to look at the worst data in both directions.

SPEAKER_00 33:06

Yeah. So that was, I feel like a pretty good overview of everything.

SPEAKER_02 33:12

I think now would be best to go on to individual brands. Yes. So this was an umbrella review from 2024. What we're going to do is give you some example review papers that were dedicated towards certain brands and give you just like a quick summary. Now keep in mind that a lot of these will need to be updated. You will find other research papers that have different percentages because that's how research papers work. You test them under different conditions, blah, blah, blah, blah, blah. Don't take these as gospel. We will put the images on the screen. But let's do some of the most popular brands so people who are listening are like, that is what I can test.

SPEAKER_00 33:48

That's what I'm wearing now. Okay. We're going to start with the aura ring.

SPEAKER_02 33:51

Yes.

SPEAKER_00 33:52

Okay.

SPEAKER_02 33:52

I put this in on purpose.

SPEAKER_00 33:54

Yeah.

SPEAKER_02 33:55

Because ring technology is new or newer than wrist technology. I personally find ring technology quite interesting because in some ways it's even less invasive. If someone wants to wear a different watch, they can wear a ring, which is very, very stealthy.

SPEAKER_00 34:09

So this one was really cool. As an ovulation tracker, for those of you menstruating, uh, it actually detected 96.4% of ovulation with an average error of 1.26 days, okay, which is pretty spot on, if you ask me. Lower than the calendar method of 3.44 days. So you know, you can say it's more than twice as accurate. That's pretty cool.

SPEAKER_02 34:38

Yeah. Uh, we will look at sleep next. So one meta-analysis, I'll do the first one, you can do the second one, where meta-analysis actually said that sleep was accurate. It said there were no big inconsistencies compared to gold standard testing. Uh, that is comparable to medical grade sleep studies, including polysomnography and actigraphy, uh, no significant differences in total sleep time, sleep efficiency, wake after sleep onset, sleep, blah blah blah blah blah blah blah. Good for sleep was the conclusion of one review. However, over to you.

SPEAKER_00 35:09

However, there is another meta-analysis of looking at different uh different aura ring devices, and they said that sleep weight distinction was 85% accurate, and then distinction between the different sleep stages was worse. And then across all the sleep stages, the aura was 5'3, 53% accurate, which is not a very high number.

SPEAKER_02 35:33

Yeah. Uh the quote that I think is important, which is one of the things that we've explained already, which is critically, while some devices may demonstrate reasonable agreement with polysomnography on average, this agreement masks substantial individual level inaccuracies, prohibiting their use in clinical sleep medicine. So even if a device looks accurate when you pull all of the research studies together, if it if you still get anomalies within that, it means that you still can't test it. So for example, my heart rate data from the race that I showed earlier, the heart rate data for me could be accurate most of the time, but if there are times where it completely fucks up, that is a problem.

SPEAKER_00 36:15

You're like, we can't use that.

SPEAKER_02 36:17

Yeah.

SPEAKER_00 36:17

Yeah. Uh and then there was still a different paper where, and this is not looking at uh sleep specifically, but they're looking at step count, which they significantly overestimated with the aura rings. And then they also underestimated energy expenditure, and as the exercise intensity increased, so did the discrepancy, so did the margin of error.

SPEAKER_02 36:38

Which was actually like the opposite of what we talked about earlier. So the the faster you go, the less accurate your aura ring is.

SPEAKER_00 36:44

So basically, what I'm picking up from this is what we can say from aura ring is fantastic at tracking ovulation. Everything else is like, eh.

SPEAKER_02 36:53

Yeah.

SPEAKER_00 36:54

That was my impression of the research.

SPEAKER_02 36:56

Yeah. I don't want to make any brands mad because I know that brands sometimes get a little bit trigger-happy with uh We're only reporting what the data says.

SPEAKER_00 37:04

That's it.

SPEAKER_02 37:04

Okay. Okay. We're gonna we're we're gonna be unbiased on this. We're not gonna give our our our own opinions. Okay.

SPEAKER_00 37:11

Next brand is Fitbit.

SPEAKER_02 37:14

Okay.

SPEAKER_00 37:15

Go ahead.

SPEAKER_02 37:15

So Fitbit uh tends to underestimate heart rate, energy expenditure, and step count.

SPEAKER_00 37:20

Heart rate by around three beats per minute.

SPEAKER_02 37:23

This was a very even like three, three, three study. I actually had to read it several times just to make sure this was correct because it seemed odd. So yeah, heart rate by around three beats per minute. Uh the step count was uh three steps per minute. Yeah. Which doesn't sound like a lot.

SPEAKER_00 37:39

But it adds up.

SPEAKER_02 37:41

So over the course of an hour, that might be closer to a couple of hundred, which like that's that's that's bigger than three steps per minute sounds.

SPEAKER_00 37:51

And then over the course of a day, a week, a month, obviously the difference compounds. Uh as far as energy expenditure, we again have a number three, uh underestimated by around three calories per minute. So you're like, oh, that's not very much. However, uh, the example that the authors use was depending on your obviously your body size, your walking speed, and so on and so forth, we're for one hour, that could be the difference between 280 calories burned versus 420 calories burned.

SPEAKER_02 38:18

Which is much bigger.

SPEAKER_00 38:19

Yeah.

SPEAKER_02 38:20

So even if these research papers say that on average they tended to be good, if you actually look at everyone's individual data, there'll be people right at the higher end and then people right at the lower end. And it's possible that you'll be one of those. So your friend, it might tell them very accurately, but it doesn't necessarily mean that it tells you accurately, which is why when people say this device is accurate, a study might say the device is accurate, but it doesn't always mean that it's accurate for all people. It tends to mean that when they look at the average and they pull it together, it was like accurate as an average.

SPEAKER_00 38:54

Okay. Now we're moving on to Garmin, which again is what I'm wearing. So I'm interested in this data.

SPEAKER_02 38:59

Yeah, so this was a like a Garmin specific review paper that depending on which website you looked at was published in 2020 or 2021. Yeah. Which means that by now it's going to be outdated, but this is what you're saying. So heart rate exceeded 5% error rate uh at rest, but got worse with higher intensities. Uh in 95% of cases, the values fell between minus 27 and 33 beats per minute of the ECG, which That's a lot.

SPEAKER_00 39:32

There's a lot huge variability.

SPEAKER_02 39:34

If someone was doing if someone's heart rate was 160 beats per minute on an ECG and the Garmin was saying 27 under that, like that's big.

SPEAKER_00 39:44

Actually, that makes sense because I was telling you yesterday when I was doing a brisk incline walk on the treadmill, uh, which is not, you're not I'm I'm not doing a speed run, I'm not pushing super hard. My Garmin said my heart rate was at plu over 180 beats per minute, but it felt more like 130, 140 at the time, based on my effort. So I was like, I know that's off by a lot.

SPEAKER_02 40:08

Okay. Uh and then the interesting thing here, now keep in mind that not all exercise is the same and not all walking is the same. Yeah. So just because something is accurate when it comes to cycling doesn't mean that it's accurate when it comes to walking. Right. And if it comes to walking, doesn't mean it's accurate for all walking. So the let's say zero to five step difference over five minutes of walking. We're doing sorry, doing steps now. I'm fast. Yeah, it's okay. Covering steps, yeah. Steps, uh, zero to five steps difference over the course of five minutes, which is small. It sounds very, very small. Yeah. Um however, that changed to 16 steps when people were carrying bags and 37 steps when pushing a stroller.

SPEAKER_00 40:49

And I wonder if that has to do with the difference in what your arm is doing.

SPEAKER_02 40:54

Yes. Well, I I say yes very confidently. Like in my head, yes.

SPEAKER_00 40:58

Your hand is on the stroller, it's not swinging naturally like you would when you're walking, therefore, it's hard to maybe pick up when you do take a step.

SPEAKER_02 41:07

And I think that's actually a great thing for people to remember because if it is measuring your step count based on your arm activity, in theory, what that means is running technique would make it make a difference. Yeah. Uh if the amount your arm swings is registering as steps, you could change the step count just by how vigorously you're moving your arm. And I actually noticed this if I run on a treadmill, uh, my watch will often, even though they're supposed to sync, yeah. Like it sinks to my running app, this can sync to my treadmill, blah blah blah blah. Uh, my running watch will often underassume or under-measure how far I have run.

SPEAKER_00 41:47

Interesting. Yeah. Also, very niche example. Imagine if you have like a broken arm.

SPEAKER_02 41:53

I've I actually thought that. I didn't think.

SPEAKER_00 41:55

And then so obviously your arm is stationary like this while you're walking.

SPEAKER_02 41:58

I need someone who's broken their arm but won an activity tracker to chime in and on that arm.

SPEAKER_00 42:04

Yeah on that same side arm.

SPEAKER_02 42:05

Yeah.

SPEAKER_00 42:06

Who knows?

SPEAKER_02 42:06

Speed if you're someone that's running, there they uh measured this out as if you're running on a track and it had again like decent agreement from 0.84 to 0.99 while running at different speeds on the track. People know this when they race because you often have like you'll have like chip time and then you'll have your watch time. So how many people finish a race? They look down and they're like, oh, it's not actually quite right. Uh and people will often say, well, it's when you're racing, like depends on the line that you take in the race, but it also depends on how accurate the device is.

SPEAKER_00 42:37

Yeah, right. Uh and then last one for Garmin was energy expenditure, which oh man, okay. Errors range from low to substantial across multiple activities. In most cases, the mean average percentage error was deemed unacceptable.

SPEAKER_02 42:54

Yeah, that so this is a running theme. Like generally speaking, when I have talked about uh watches not being reliable, uh calorie burn is is the thing that I talk about. But the reason I talk about that is because calorie burn is often the thing that people rely on. They're like, oh, I've burned 3,000 calories today. Or actually, more likely is I went to the gym and I burned 400 calories today. Yeah. I don't trust calorie expenditures from watches, regardless of device.

SPEAKER_00 43:20

And also I will say, as a coach who uh has in the past, especially quite heavily, worked with clients like with macro tracking and things like that, um, I've actually never once asked them what is the calorie burn according to your activity tracker ever. I do not rely on that uh because of reasons like this. And so I yeah, so I I find that a lot of times the activity trackers grossly overestimate how many calories you burn. For example, um, not that this is not included in here. Um I once was on a treadmill for an hour and it told me I burned 2,000 calories. That would be one hell of a pace. Oh, yeah, and I was you know going at like a decent jog, but it was not a and it was probably close to especially given my small stature, probably close to like 400 calories for the hour. Yeah. If that.

SPEAKER_02 44:10

Yeah. I so that's a that's a very common one because a lot of people will say, like, especially on a treadmill, which I know is a little bit different, but it's very common for people to say, I went to the gym and I burn this many calories. I have seen people who have said I burned a thousand calories earlier in the gym. And we're like you honestly very, very, very, very, very probably haven't yeah. Even close to that. Correct.

SPEAKER_00 44:33

Okay. Last one, and we left this for last because it's actually the number one most popular brand, which is I think it is.

SPEAKER_02 44:40

Uh according to a paper, yeah. Apple devices. And the other reason I left this to last is because this is actually the newest research paper. So this particular paper is from earlier this year, literally a couple of months old and that means that it's newer than the umbrella review that has been updated elsewhere. Yeah. So I think it makes sense to do the newest newest one last.

SPEAKER_00 45:00

Okay. So this paper looked at 14 different health metrics from all the Apple Watch models from series nine to ultra two.

SPEAKER_01 45:10

Yes.

SPEAKER_00 45:10

Okay. Heart rate was the most frequently validated metric.

SPEAKER_02 45:15

Yeah. And that's important to know because if a device has I don't even know how many functions that this has, but if it has 10 or if it has 12, they might validate one or two or three most commonly and then some others tend to be neglected.

SPEAKER_00 45:28

Okay. Let's dive in. Heart rate.

SPEAKER_02 45:30

Heart rate so mean difference in resting heart rate was again very similar to what we had earlier minus 2.5 ish up to 3.6 ish. So generally speaking like three beats per minute in either direction pretty good. Yeah. Very good. Well I say pretty good. It's very good. Very good. Yeah um okay next one was step count overall was moderately accurate they found no uh differences in accuracy necessarily based on walking or running speed uh in the in the heart rate before we move on to step count uh one of the things that I thought was interesting is during exercise the percentage error was lower than 10% uh but it tended to rise as exercise exercise intensity increased which is kind of what you find with your work which is what I said earlier but it was the opposite of a different one so even if it's accurate at lower intensities it doesn't necessarily mean it's accurate as higher intensities. Yeah. And we will have put some like images on the screen over the course of this video.

SPEAKER_00 46:31

Next up is sleep.

SPEAKER_02 46:32

So just like we have seen before there tended to be good differentiation between being asleep and being awake however poor differentiation between physiologically similar sleep stages. So my watch might be pretty good at telling me when I fell asleep and when I woke up but much less reliable at differentiating when I was going through different stages of the sleep cycle.

SPEAKER_00 46:54

Yep. And then the last metric we'll cover for Apple Watch is energy expenditure which they found margins of error were quite large both during exercise and at rest. So similar to the Garmin. And then the mean absolute percentage error reported values over 20% in at least one exercise condition.

SPEAKER_02 47:15

Yeah so it might be more accurate if you are cycling for example I'm making that up I I'm not reading it from the paper.

SPEAKER_01 47:21

Yeah.

SPEAKER_02 47:22

But then walking or running or lifting weights or whatever there can be different margins of error but they were over 20% at least one exercise condition. Yeah.

SPEAKER_00 47:32

So once again energy expenditure very very very hard to trust your watch yeah last thing we'll say about energy expenditure for Apple Watch was that the mean absolute percentage error ranged from 9.71% during running which is around 10% to this is painful walking was 152%.

SPEAKER_02 47:55

Yes. I actually had to check that twice after I made notes I had to go back because I thought I had misread it.

SPEAKER_00 48:01

Yeah.

SPEAKER_02 48:02

That is quite a big discrepancy.

SPEAKER_00 48:05

That is quite inaccurate by a lot.

SPEAKER_02 48:07

Yeah.

SPEAKER_00 48:08

So basically we have a wide range of error rates depending on the brand the device and the metric that you're looking at that's kind of the TLDR.

SPEAKER_02 48:19

Yeah out of interest how many steps has your watch does your watch say okay so I've been intermittently I don't know this is not considered an arm swing but I've been intermittently shuffling so at the beginning of the episode 50 minutes ago whatever you had apparently done 255 steps.

SPEAKER_00 48:37

It is now saying 3171. Oh Jesus obviously I've not taken a single step so what is that like you're uh 600? You're looking yeah I've taken apparently according to my garment I've taken 600 steps while we've been recording this episode.

SPEAKER_02 48:52

And it like obviously that's a slightly dumb test because she's intentionally moving her arm but the whole point is if it's sensing your steps based on your arm movements then it shows how there is room for error based on other factors.

SPEAKER_00 49:05

Well actually someone someone on on social media also messaged me that they were at their park with their child and they were pushing their child on a swing and while they were standing stationary for like 30 minutes or whatever it then said oh you've hit your step cunkle while they've been standing you know let's see mine.

SPEAKER_02 49:21

How many steps have I taken 1628 steps. So even though I haven't been shuffling I have intentionally not been shuffling I've been sitting as stationary as you would expect me to sit for a whole episode I have actually apparently walked uh 223 steps which is really funny because all you've been doing is gesticulating here and there throughout this episode and not even that wildly I would say but it's still tracking. So this is an example how if you do like a walking study where you walk on a treadmill for five minutes and you are measuring it against someone literally counting it step by step any margin for error is when you are doing your designated activity. But when people are wearing a watch all day it isn't just your designated activity. It's like a free living situation. Meaning every time you are shaking your protein shaker, every time you're washing your hair every time you're scratching your back every time you're gesticulating or whatever. Oh you're you're walking yeah the interesting thing like in theory if it is if uh it knows where you are which is what you would expect or some device would know where you are it should know that you are not moving. Right. Or like devices with GPS you hopefully would say oh you actually haven't moved. Right. Maybe it just thinks you're walking on the spot.

SPEAKER_00 50:39

And then going back to the data about like strollers if you're pushing a stroller you can see how not having that arm swing would then make it even more inaccurate because they're like oh I guess you're not I guess you're not moving.

SPEAKER_02 50:52

Yeah. So I think now we should round it up. So there were obviously a lot of like statistics and percentages a lot of numbers there will be bookmarks in the or chapters in the in the video version to make it easier for people to want to find your device. So let's do some like cliff note summaries. Okay. I want to say first and foremost please please please keep in mind especially if you're a fan of the brand or you work for a brand and you're angry at us understand that when we have talked about research papers they are probably outdated meaning that for example like if we're talking about Garmin's and Fitbits etc these are research papers from two, three years ago meaning we are talking about devices that aren't even being sold now. But that is one of the issues with activity tracker recommendations.

SPEAKER_00 51:37

And we hope and trust that your accuracy has improved over time. Of course yeah so don't don't get mad at us and again we're still fans we wear our devices every day still so okay practical takeaways here's what I'll say activity trackers can still absolutely be beneficial even though we both know there's a margin of error for basically every single metric we find it still helpful to be like oh according to my garment I've only taken roughly 2,000 steps and it's 5 pm and then it it might then encourage me to go out and you know go for a 30 minute walk when I would maybe otherwise not even have that on my radar.

SPEAKER_02 52:15

Yeah so I think in my opinion even though we have painted sometimes occasionally like quite a bleak view of of the accuracy and uh like accuracy but also just something can be accurate once it doesn't mean it is always as accurate. So people make a mistake of thinking oh if it underestimate estimates my energy expenditure or overestimates it I'll just add a couple of hundred on but that implies that it'll be the same every single day. Right. And that is not the same reliability is different. So something can be inaccurate but then inaccurate in a different direction the next the next time however things do not have to be precise to still be useful. Yes. So if something tells you that you didn't sleep much last night it doesn't necessarily matter if it's calculated the minutes perfectly if it still means that you take note of your sleep time and then you make an effort to improve it. Right. So even last week it told me that I hadn't slept as much as I normally do. And realistically like I probably knew it but my watch telling me that is like you should try and go to bed earlier. And I went started trying to go to bed earlier. Yeah. Meaning that your device is giving you feedback which can then increase your adoption of health promoting behaviors which is ultimately kind of the whole point of activity trackers anyway. Yeah.

SPEAKER_00 53:33

At the same time you don't want to overrely on the activity tracker readings. For example there have been many times for those of you with a garment it'll tell you your like your body um readiness score I forget the right term. Yeah readiness score a few minutes into your work your logging your logged workout especially for a run and it might be like minus four relative to your baseline you're at a minus four your your condition is not good today. But if you're feeling great otherwise don't take that as a sign that you need to necessarily scale back your effort for the day or sometimes and I know I think even the Garmin Instagram page they have running jokes about this where you're you wake up, you feel fantastic, you're like, I'm gonna have a great day I'm gonna crush my workout blah blah blah blah and it says your training net your training readiness score is low.

SPEAKER_02 54:17

Or your workout was like unproductive your workout was unproductive right even though you thought it was stellar. Yeah.

SPEAKER_00 54:23

What we're saying is do not over rely overrely on those readings especially if the way that you are actually feeling is different.

SPEAKER_02 54:30

That actually brings me have you finished your point because there's something I want to say on that.

SPEAKER_00 54:34

My last thing was about was that you know you probably know your body better about how you're feeling subjectively than a device is telling you. So don't be afraid to ignore uh what it's telling you. Like if you say your workout's not productive or you're not ready or this and that your sleep was terrible even though you see you actually think you slept great you probably are going to be the better judge of how you actually feel yeah so you have I remember you actually telling me once it told me that I my training readiness was really low but I still had a great workout.

SPEAKER_02 55:05

And some people will say I was going to work out but my training readiness is low so I'm gonna have a day off yeah which don't get me wrong like listen to your body if you need a day off take a day off absolutely but sometimes a watch will tell you that you should have a day off and you can still have a great workout or you can still do something and the point I want to make on that which is discussed I've mentioned earlier in the episode to a degree is when devices are calculating things like training readiness scores or whatever score that your watch is throwing out at you that's not a very well known thing like heart rate everyone knows what it is. VO2 Max everyone knows what it knows what it is. But if a device is calculating its own training readiness or however they want to phrase it's what they are using is an algorithm or an equation that a lot of the time they are not telling you how it's calculated. So for example we launch a device the Ben and Sohi activity tracker XXL amazing 5000 if we had a how amazing are you feeling today we would calculate it on for example your sleep score maybe like your heart rate how much you've moved etc but we are essentially making up we are saying sleep that can contribute 60% sure your heart rate that can contribute 30. We're using an algorithm to say we think you are very ready to exercise today. And all of those things are calculated by the brands and kind of made up like it's often proprietary technology. So don't fully trust it it will often tell you a number which might not necessarily reflect how you're feeling in practice and if you rely on it too much sometimes you might skip workouts that you don't need to skip. Right. For example right so basically don't ignore your actual body's cues and how you actually feel yeah but generally speaking I really like activity trackers I like that they can increase adoption of health promoting behaviors. I like that even with kids and adolescents they can encourage people to walk more and move more if it tells you that you haven't been more active if it means that you'll be more active I think it's great. I think it can be great with the development of technology which can be used in a medical setting I think that's really really exciting. But don't get too caught up on the numbers and if you are getting too caught up on the numbers maybe take take a step back. We have a mutual friend I don't even need to say their name that's not important who I remember making posts saying I stopped using activity trackers and I felt like it was better for my brain analyzing the numbers too much getting too obsessive about step count and minutes spent doing whatever.

SPEAKER_00 57:49

Right right so I feel like it's really helpful to give you ballpark data and also helps identify maybe trends in your activity level and the different metrics that it measures.

SPEAKER_02 58:04

I think it's the same with a lot of things that that measure you like if it if you get something that measures your body fat percentage it's not likely to be fully accurate. It doesn't necessarily mean that it's useful and that's okay. As long as you know that it might not be a hundred percent accurate and you don't get too caught up in it, great. Just make sure that you are using them for positive reasons. Yeah. And if you have any other devices or questions or something feel free to ask us because we cannot answer if you say well I use this I use this is there any research on this or whatever we can answer. We just tried to pick some common brands to give you an example of some of the research out there.

SPEAKER_00 58:41

I guess good so we are still fans but there are limitations.

SPEAKER_02 58:47

Done. That's a good conclusion to be okay bye see you next time high five