Heart Rate Extremely Inaccurate, Microsoft Says Band is Entry-Level Device

[Stodds1967]

I've been very interested in the accuracy of the heart rate monitor as well and this morning was my first test. For reference, I use the Mio Alpha optical heart rate monitor and I've been using it for months. I've found the Mio Alpha to be more reliable and consistent than any heart rate monitor I've ever worn and have been incredibly satisfied with the results...stated differently, I trust the Mio Alpha.

Today I wore the Band on my left wrist (I'm right handed) with the display on the inside of my wrist. I placed the Mio Alpha above the Band and did a simple run on the treadmill where my average heart rate was 144. I snapped pictures at different times and was pretty amazed at how similar the readings were. I discovered that my heart rate was pretty close until I got above 150. The heart rate on the band was then 5-10 beats higher than the heart rate on the Mio Alpha. I stopped running to take a picture but discovered that after I stopped running within 5 seconds the 2 heart rates would be nearly identical again. This is my first test, but for all intents and purposes I'd say it's good enough for 95% of the people who exercise and want heart rate, and it is WAY more convenient than a heart rate strap.

I'll let you know as time progresses how the comparisons go.

 

[jwpear]

You do realize that this product has barely been out a week, don't you? They haven't even had time to replenish stock after the first batch sold out.

If there's still radio silence from Microsoft a month from now, then you might have a good argument but expecting feedback this quickly seems unreasonable to me.

Microsoft didn't advertise they were throwing this thing into the market in beta form until after folks had already put down $200 for it. It's my understanding that they did so to get addition data to fine tune the device and software. I'm okay with beta testing it, but I think they owe us frequent communication on what the data is telling them and how they plan to tweak. So no, I don't think a week is too soon.

 

[tbsteph]

When, where did Microsoft ever say the Band was being sold in beta form before or after it was offered for sale? I don't doubt they are monitoring usage to improve the device/software. But, isn't that a common practice for any device/company. I think you owe a us a link.

 

[DroidUser42]

When, where did Microsoft ever say the Band was being sold in beta form before or after it was offered for sale? I don't doubt they are monitoring usage to improve the device/software. But, isn't that a common practice for any device/company. I think you owe a us a link.

I think he's referring to the "Why Microsoft Band is in short supply" story in WindowsCentral.

"Microsoft Band is best thought of as a demonstration device for Microsoft's sensor and software technology."

That's not an isolated sentence. That theme is repeated in the article.

 

[DroidUser42]

I think that this is made worse by the fact that people tend to check their watch right after an intense stretch, e.g. freewheeling down a hill (me), or after completing a set of reps, etc. This makes things seem extra bad because it appears that the band loses lock when the heart rate changes rapidly

That may be true, but for the record, I was checking the band during the run when things were fairly stable.

I did notice that you're wearing the band closer to your hand then I do. I may have to try that position, but it seems like it limits my hand movement and puts the band under considerable stress. I normally wear it on the other side of the bony protrusion on my wrist. However, that's quite a bit more sensitive to the top/bottom pressure The Band puts on the wrist. If I make it too tight, my hand starts feeling numb after a bit.

 

[ZvjerOPC]

Do you guys believe that enabling additional sensors in the future could improve accuracy of the band? I was just wondering so because the new Jawbone UP3 boasts some galvanic skin response sensor for the heart monitoring and that sounds a lot like MS Band skin sensor which appears to be unused at the moment.

I'm nowhere near a sports pro as some people on this thread appear to be, but I would like to pick a good device for my basic fitness needs (cycling, skating, running) and I would really want it to have a HR monitoring because that's all that I miss on my Lumia 930.

Speaking of Lumia, it appears that Jawbone does not really supports Windows Phone while Fitbit has a decent Windows Phone app but lack a lot of features (namely MobileRun) that are only iPhone supported.

So in the end when I am searching for a good HR wristband monitor to go with my Lumia it appears that MS Band might be the ONLY choice, no? P.S. I'm from Croatia so getting one might pose a bit of a challenge but as long as it can do metric measurements it will do fine.

 

[sketchy9]

I discovered that my heart rate was pretty close until I got above 150. The heart rate on the band was then 5-10 beats higher than the heart rate on the Mio Alpha. I stopped running to take a picture but discovered that after I stopped running within 5 seconds the 2 heart rates would be nearly identical again. This is my first test, but for all intents and purposes I'd say it's good enough for 95% of the people who exercise and want heart rate, and it is WAY more convenient than a heart rate strap.

I'll let you know as time progresses how the comparisons go.

My experience on a rowing machine as well. The HR tracks pretty well with a chest strap monitor until I get above HR of 150, then the Band just sort of loses it. It will climb sky-high (HR of 175??), so that makes everything else inaccurate. I'm going to experiment with different positions, tightness, etc.

 

[Richard Servello]

My experience is that the HRM on the band can be way off on the elliptical. The elliptical HR shows 90 while the band is in 125-130 range, while I'm holding the moving handles. When I grab the stationary handles the band readings go down to ~90. And yes the 90 is the correct one as I get a medication which practically keeps my HR low. I'm wondering if the band is using the accelerometer to help the HR "calculation" when the measurement fails... The 125-130 bpm looks completely believable under normal circumstances.

It uses a number of sensors to determine activity level...which actually makes it more accurate than competitors.

 

[Richard Servello]

My experience is that the HRM on the band can be way off on the elliptical. The elliptical HR shows 90 while the band is in 125-130 range, while I'm holding the moving handles. When I grab the stationary handles the band readings go down to ~90. And yes the 90 is the correct one as I get a medication which practically keeps my HR low. I'm wondering if the band is using the accelerometer to help the HR "calculation" when the measurement fails... The 125-130 bpm looks completely believable under normal circumstances.

I've been very interested in the accuracy of the heart rate monitor as well and this morning was my first test. For reference, I use the Mio Alpha optical heart rate monitor and I've been using it for months. I've found the Mio Alpha to be more reliable and consistent than any heart rate monitor I've ever worn and have been incredibly satisfied with the results...stated differently, I trust the Mio Alpha.

Today I wore the Band on my left wrist (I'm right handed) with the display on the inside of my wrist. I placed the Mio Alpha above the Band and did a simple run on the treadmill where my average heart rate was 144. I snapped pictures at different times and was pretty amazed at how similar the readings were. I discovered that my heart rate was pretty close until I got above 150. The heart rate on the band was then 5-10 beats higher than the heart rate on the Mio Alpha. I stopped running to take a picture but discovered that after I stopped running within 5 seconds the 2 heart rates would be nearly identical again. This is my first test, but for all intents and purposes I'd say it's good enough for 95% of the people who exercise and want heart rate, and it is WAY more convenient than a heart rate strap.

I'll let you know as time progresses how the comparisons go.

View attachment 86863View attachment 86864View attachment 86865

The band has more sensors than just heart rate...i trust it over the Mio.

 

[greyskytheory]

That may be true, but for the record, I was checking the band during the run when things were fairly stable.



I did notice that you're wearing the band closer to your hand then I do. I may have to try that position, but it seems like it limits my hand movement and puts the band under considerable stress. I normally wear it on the other side of the bony protrusion on my wrist. However, that's quite a bit more sensitive to the top/bottom pressure The Band puts on the wrist. If I make it too tight, my hand starts feeling numb after a bit.

For me I like to see my hr after I do a set of something quite taxing. For example, today I was doing my bear complex routine and using my Suunto ambit 1 with chest strap I was able to see I got my hr up to 196 which makes me happy. I always try to get my hr above 191 which is my max hr. With an accurate hr I can see real time which exercises, movement etc are getting my hr up and I can then adjust my routine. Again, in my testing with my fitness routine the band does not accurately track rapid rises and drops in hr when doing interval training. For steady state training its fine but it is not reliable for me and the dynamic routines I do. After speaking with customer service they are worried then band my be too loose since there were times it stopped tracking my hr all together. Hopefully a smaller size will fix the issues and hopefully its not so tight it becomes uncomfortable or the band snaps. Overall I love it and with a few updates to address other user issues/requests it will be a top of the line device. Hopefully Microsoft doesn't "Zune" the device because the Zune HD was great then they buried it.

 

[Upstate Dunadan]

Today was my day off from working out at the gym, so I decided to jump on the elliptical (SmoothFitness) we have at home to see how the Band compared to the elliptical's built in HR and Calorie analysis.

I was not expecting them to be in sync, but I was surprised at how close they were over the course of a 45 minute (steady state) routine. Here are the results.

Band High HR: 137
Elliptical High HR (observation/estimate): 142

Band Avg HR: 128
Elliptical Agv HR (observation/estimate): 120-135

Band Calories Burned:559
Elliptical Calories Burned: 514.7

A few comments:
1. My elliptical observation/estimates were written down before I pulled up the Health App to review my Band results.
2. The elliptical HR seemed to take longer to stabilize and lock in than the Band did.
3. Of the dozen or more times I glanced at my Band to compare HR to what the elliptical showed, each time the Band was within a few beats.
4. Several times during the workout I lost HR on the elliptical which then took a short time (10-15 seconds) to lock back in. The Band never lost it.

Overall I was surprised and very impressed. Makes me feel much better about the readings I'm seeing at the gym, which is where I am 6 days a week (3 on, 1 off, 3 on).

 

[SteveVII]

In response to trusting the band over the mio due to more sensors:

I have yet to see anything that supports this...have you? From what I have seen, calories burned seems to correlate to heart rate. I took a spin class today and my polar chest strap reported that my average was 11bpms higher. The band also reported about 80 less calories than the chest strap. I'm curious how the band decided that even though I worked my *** off, I somehow only burned the same number of calories as a brisk jog.

It's pretty clear that this heart rate monitor is very temperamental and inaccurate. I've found that it is especially bad with elevation changes (running up or down a hill, being on the elliptical, etc.). I found it odd that it was so off in my spin class where my hands are perfectly still. Sometimes I feel like it doesn't try to lock in the heart rate unless it senses motion in the arms. Then again, the number shown when it is locked is bs as it typically takes 10 seconds to normalize. I don't have any trust in the 24 hour chart it shows in the app.

Listen, this is a well rounded device for someone who has no clue about their body. I would get one for my mom if it wasn't so clunky. But for me, it definitely doesn't live up to my expectations. I'm going to keep it...I paid for my band selling my extra one on eBay so it made it worthwhile. I definitely enjoy the notifications as well. That being said, I'm really hoping Microsoft can figure out a software miracle and proper documentation or this thing just isn't going to do well.

 

[jwpear]

When, where did Microsoft ever say the Band was being sold in beta form before or after it was offered for sale? I don't doubt they are monitoring usage to improve the device/software. But, isn't that a common practice for any device/company. I think you owe a us a link.

See here: With Microsoft Band, a sense of company's health | Brier Dudley's blog | Seattle Times

 

[jwpear]

Ran a 5k on Saturday with my 8 year old. I forgot to take my strap to compare. The Band didn't seem to match my heart rate while running, but looking at the splits, seems like it might have been okay. But this was a more constant activity. Frustratingly, it was never able to lock in the GPS at the race location.

I worked out in the gym this morning. I wore the Band and my Cardiosport watch on the same arm so I could compare and manually check my pulse while watching how both responded to what I was doing. The strap consistently responded more quickly to changes--it was noting changes in real time. The Band was consistently slow to catch up. It almost always took it 10 to 15 seconds to catch up with my strap and the pulse I was manually checking. I know this because I was watching both as I checked my pulse for 15 seconds.

I noticed the Band also often indicated a hollow heart, suggesting it didn't have a good lock on my heart rate. Would love to know how to get it to stay locked in.

Daniel's article on how often the Band measures your heart rate (every second during a run or workout) suggests that it should track more closely. I don't know if I have a bad sensor, improper fit, or the Band just isn't measuring as fast as Microsoft indicated to Daniel. See here: This is how often the Microsoft Band checks your heart rate | Windows Central

 

[DroidUser42]

Frustratingly, it was never able to lock in the GPS at the race location.

My experience with GPS units in general is that it doesn't get a lock if you're on the street in a high-rise district. It needs to see quite a bit of sky.

One thing I wonder about is if the GPS is just GPS or AGPS. The "Assisted" mode found in some smart phones uses some downloaded information and cell tower data to jump-start the lock-in process. Since the band doesn't have direct connection, it would require the phone App to feed it that information.

 

[dbregman]

I worked out in the gym this morning. I wore the Band and my Cardiosport watch on the same arm so I could compare and manually check my pulse while watching how both responded to what I was doing. The strap consistently responded more quickly to changes--it was noting changes in real time. The Band was consistently slow to catch up. It almost always took it 10 to 15 seconds to catch up with my strap and the pulse I was manually checking. I know this because I was watching both as I checked my pulse for 15 seconds.

I noticed the Band also often indicated a hollow heart, suggesting it didn't have a good lock on my heart rate. Would love to know how to get it to stay locked...

I wonder if the HR display on the band lags a bit from the actual. This might explain the discrepancies. When I look at the history it seems more accurate then the band spot check.

 

[swndg86]

I wore my Polar FT60 along with my Band last night at ice hockey to compare the results and they were very close to each other - avg. for Polar 153 and avg. for the Band 147. The only time I saw a discrepancy was when I went really hard and got my heart rate above 180 for a few seconds during some of my shifts. Once I got back to the bench and my heart rate started to go down, they both showed the same rate. Technically, the Polar HRMs aren't that accurate for anaerobic activity so I would actually be curious which one was "more" accurate. I will test it out with a more aerobic activity later but it seems the Band was accurate relative to my Polar.

 

[mhc48]

@Stodds1967, what devices did you test the Mio against in the beginning to come up with the conclusion that it was reliable and consistent?

 

[greyskytheory]

I wore my Polar FT60 along with my Band last night at ice hockey to compare the results and they were very close to each other - avg. for Polar 153 and avg. for the Band 147. The only time I saw a discrepancy was when I went really hard and got my heart rate above 180 for a few seconds during some of my shifts. Once I got back to the bench and my heart rate started to go down, they both showed the same rate. Technically, the Polar HRMs aren't that accurate for anaerobic activity so I would actually be curious which one was "more" accurate. I will test it out with a more aerobic activity later but it seems the Band was accurate relative to my Polar.

View attachment 87099

I had similar results with mine. Steady state exercise seems to cause no problems for the band but interval training or sharp spikes in hr seem to give it problems. I noticed the Band has a discrepancy in you max hr according to the polar. Considering you were playing hockey I'm sure you got it somewhere around 195.

 

[tbsteph]

All three of the links are basically the same article. None of them provide a quote from Microsoft stating the Band was being sold as a beta version. The closet to such a statement is where Microsoft stated they were using user data to improve the device.

Look I'm not trying to split hairs. The OP said "Microsoft didn't advertise they were throwing this thing into the market in beta form until after folks had already put down $200 for it." I find that statement to be patently untrue.