Monday, May 01, 2017

Whoop Strap 2.0 Review: Novel Approaches to Monitoring Recovery and Optimizing Performance

Article by Derek Li MD with Sam and Jake Winebaum
Whoop Strap 2.0
iOS and Android

Editor's Note:
Derek Li is a 2:42 PR (Boston) marathoner and physician from Singapore where he is known as the Running Doctor and as a shoe and gear expert.  He also writes at his own site Running Commentary.

Jake Winebaum is a tech entrepreneur, accomplished masters cyclist and alpine skier who has tracked his heart rate, power, sleep, and performances for decades.  He has competed in the Leadville 100, Transalp Challenge mountain bike races, Tour Transalp, and Everest Challenge road bike races multiple times.

Sam Winebaum is the editor and founder of RoadTrailRun.

Whoop Description
"The WHOOP system includes a sleek, water proof wrist-worn strap that measures key strain and recovery variables more than 100 times per second, 24 hours a day. WHOOP’s proprietary algorithms then process this data to provide athletes an Intensity score, which informs them about the level of Strain on their body and what it means; a Recovery score, which measures the body’s preparedness for strain or exertion; and a Sleep Performance score, which evaluates the hours of quality sleep an athlete got in relationship to the sleep he or she needed. "

About Whoop
Whoop is a venture backed Boston based company. Starting with a focus on collegiate, professional, and Olympic athletes and teams, the Whoop Strap is now available to the public.  Whoop was recently named the Officially  Licensed Recovery Wearable of the National Football League Players Association, the first wearable to be so named by a professional players' association 

This is by the far the most comfortable skin-sensor based HRM I’ve ever tried. The sensor unit is very form fitting, with a gradual curve to suit the shape of the average-sized forearm, and the elastic strap can be easily adjusted to give a snug but not overly constricting feel, and I found that one does not need to make it extremely tight to give very consistent HR readings, without any aberrant spikes, even with vigorous running. I wear it on my right (non-dominant) wrist, and the numbers track very consistently with my left Garmin Forerunner 35 HRM during my runs. The only difference is I have to make sure my Garmin watch is sufficiently tight during runs to prevent aberrant readings from sweat and slippage while I have no such issues with the Whoop.

Charging and Maintenance
I have gotten used to wearing the Whoop 24/7 and even though the battery life is stated to last more than 24 hours, I have gotten into the habit of charging it for 30-45min blocks once a day while I am at work so that battery capacity is steadily in the 40-90% range all the time.
Charging is achieved by means of a black battery pack that snaps onto the primary HR sensor module, so charging can be done “on the go” as it were. (I read that this is the best way to extend the lithium battery life of iPhones and see no harm in applying it to other gadgets). The main concern is making sure the contact points for the battery pack are dry before charging the Whoop, and I am happy to report that even with running twice a day and showering twice a day, I can easily get the strap and contact area to dry within 1-2 hours of getting soaked even without having to deliberately wipe down the device.
As a wearable designed to be worn 24/7, Whoop offers many band color options
How does Whoop Work?
I used to always monitor my resting HR in the mornings, and use that as a proxy for how rested I am, and whether I am adequately hydrated for the day’s training, but the Whoop has added a new layer of insight in the form of a Recovery index. This is calculated using a combination of sleep quality data, resting HR and HR variability during the last 5 minutes of Slow Wave Sleep (SWS) each night. If there are multiple segments of SWS, then the segment with the best quality is used for analysis. The Whoop algorithms identify specific characteristics in the R-R intervals to differentiate the different forms of sleep, and I noticed that my SWS segments often occur very early in the sleep cycle, and apparently this is fairly common among Whoop users.
I suspect it is an evolutionary consequence of the body prioritizing repair and recovery as early as possible during sleep. They are also able to accurately calculate the difference between your Time in Bed versus Sleep Duration. I actually tested it out my lying pretty still for 30 mins after I woke up one morning, but was not able to fool the Whoop into thinking I had slept more. Part of the answer to how they discern this may lie in the in-built 3-axis accelerometer. Medically speaking, we all experience a cortisol surge when we wake up, and this likely causes some sort of spike or signature change in resting HR or HRV, which the Whoop recognises as us waking up.

Data Analysis
Over the past few weeks, I have found the Recovery index to have impeccable positive predictive value in terms of performance. I have never had a day with good recovery end up going badly no matter how hard the workout. In terms of future negative predictive value, however, it has been a bit of a mixed bag. I had one particular day of low Recovery of 27% on 27th March, on a day which called for a moderately hard track workout.
Even though I did not know about the low recovery until after the workout as I did not sync the previous night’s sleep data before I did the workout, I could not complete the workout. There was nothing in the previous two days that really suggested I would struggle with a workout on this day (except perhaps the long run 2 days before), but the Whoop predicted it.

 On the other hand, I had a string of low recovery days on 7th April (23%)
 although I felt strong during an easy run that day, and 8th April (11%)
but turned in a decent performance for a relatively high intensity long run on the 8th.
Strangely, my recovery improved on 9th April despite the hard effort the day before.
It continued to improve on the 10th.
I think the Recovery Index does an amazing job despite not having to take into account the amount of cumulative strain from the previous days. I have since learnt that any naps that are logged during the day will be taken into consideration when the system determines how much sleep you need for that night. There are numerous training models that rely in some way on the accumulation of chronic training stress, e.g. Training Peaks and their ATL (Acute Training Load) and CTL (Chronic Training Load) to predict fitness trends. However, the Whoop is able to bypass all this and come up with reliable data by simply analysing trends in sleep-based HR and HRV, which is most remarkable.

Subjective Input and the Role of Muscle Fatigue
Another interesting point of note is on the user input regarding subjective feeling of how rested one is the next morning,

and the activities one did before going to sleep.

It seems that the input data is not taken into account at all when calculating Recovery, but is used more for one’s own reflection on how subjective feelings of recovery relate to the calculated score. Sometimes, I get a noticeable discrepancy, e.g. I may feel sore the day after the long run of the week, but my recovery index is pretty high. It goes to show perhaps some limitations of using only HR-based data, as it only gives the cardiovascular status without considering the musculoskeletal health of the individual. I personally feel that the latter, while perhaps less important for an endurance athlete, cannot be ignored from a performance or injury-prevention perspective. Feeling sore can impact on running gait and economy, and affect injury risk even if the heart is ready to perform.

Knowing what I know now about the powerful positive predictive value of the Recovery score, I am more cognizant of syncing my data first thing in the morning to get the score before I do my runs, but I also give due credit to my subjective assessment of how I feel. I am wary of falling prey to the self-fulfilling prophecy of what the Recovery index might say about my state of fitness without giving due credit to what my body feels I can achieve. As an example, on 30 March, I had an important time trial planned on the track, and I deliberately decided to do the workout before calculating my Recovery score, and managed to produce a fairly strong performance (in fact it was 5000m PB on the track), off what was apparently 48% (moderate-range) recovery.

Moving forward, I probably need more long term data to see if there are any specific indicators for poor performance that the Whoop may have picked up (e.g. high resting HR specifically or a specific threshold of low HRV regardless of resting HR), apart from the Recovery Index score. This is, I suspect, highly individual, since I sometimes produce good performances off very low Recovery indices. The scientists at Whoop are discouraging frequent hard training on low recovery, because they feel that it is a recipe for overtraining, and they may in fact be right. I have had 3 red days (i.e. days of low recovery) in the last 30, so I’m not really in the best position to test that out, and I’m not sure I would want to either!

I do not see any discernible trends that can correlate the amount of daily strain to the next day’s Recovery Index, and it is likely that hydration, diet, and sleep quality are more important than the amount of strain in deciding one’s recovery.
The black arrow shows how the correlation occurs; you start from recovery, then impose that day’s strain, and then move to the next day’s recovery and so on. As you can see, using two different starting points over the course of the last 30 days, there wasn’t any clearly defined correlation between day’s strain and the next day’s recovery

Sleep Quality and the Sleep Coach

Sleep quality is probably the hardest to control for the amateur athlete, as how much sleep we get is often the denominator of the day’s tasks and obligations, rather than a key priority to arrange the day around. As such, I find myself under-utilizing the sleep coach or failing to meet up to the recommendations set forth. For example, as I write this review on a Sunday evening during my dinner break at work, I have an easy 5 miler scheduled for after work, which will likely be run at about 10:30pm. I will not be in bed before midnight when all is said and done. I also have an easy 8 miler scheduled for tomorrow morning. Here’s the catch: Whoop’s Sleep Coach is recommending 8h25min of sleep tonight. If, in the best case scenario, I manage to fall asleep at midnight, I would be waking up at 8:30am, and ending my 8 miler around 10am. That’s 90degF weather in Singapore.  More often than not, sleep is sacrificed to a certain extent to accommodate the day’s training or work schedule, and I am sure I am not alone in juggling this balancing act.
After your night's sleep Whoop will provide both a quantitative and qualitative analysis of your sleep and its stages.
Sam's Data
Whoop has an article describing the Sleep Coach, the meaning of the different stages of sleep it tracks, and tips to improve sleep quality at The Locker, their excellent blog here.

I think this is a very compact and well-designed product that would work well for track and field athletes and serious amateurs. The price-point may well limit the consumer market to the affluent or professional sports community, though with the current trends in GPS watch pricing, $500 may not be that much of an outlier in a year or two.
I have some concerns about the viability of the current sensor placement for team sports that may have some degree of contact, e.g. football, basketball or gym work. As these sensors rely heavily on secure contact with skin to get reliable HR data, any sort of contact risks dislodging the HR sensor and affecting the data. Static resistance exercises also reduce blood flow to the arms and can affect readings, though I have not tested this out specifically with the Whoop. The good news is that Whoop has developed arm-based strap for people who do a lot of exercise involving the forearm and wrist muscles such as cycling and weight lifting. I believe providing more placement options for the HR sensor will only open the door to more athlete groups. I suspect we may well see some options for calf or forearm placement of HRMs in the future. Having used the Sosche Rhythm+ forearm-based HRM in the past, I can attest that it is a fairly comfortable location to place the sensor and is much less likely to get nudged during exercise.
Would I continue to use the Whoop in the long term? I will be using it in the medium term, that much I can say. I am not seeing any good correlation between daily strain and recovery, and maybe I’m not supposed to. Perhaps it really has more to do with diet and hydration and sleep than anything else. I would like to see better negative predictive value from the “red zone” low recovery days, and that is the main reason I will continue to use the Whoop. I need more data, preferably 6 months’ worth or more, to say definitively for myself, “yes this works”, or “no it is not giving me enough information”. Ultimately, information is useful only if it changes that way you do something, and so far I haven’t really changed the way I train. It has shown me that I am generally training within myself and not overstretching, but it also isn't telling me where that tipping point is (yet), and steering me away from it. The jury’s still out for me, but what I have seen so far has given me enough reason to keep using it to see if I get more answers farther down the line.

Jake's Take
I agree with Derek's overall assessment.I love the sleep, 24/7 HR and HRV tracking in a single reliable and comfortable device.  HRV is super useful in measuring and anticipating recovery state.  I have used another application for that called ithelete for years.  I also use another sleep tracking device called bedit which measures resting HR and quality of sleep and other stuff like snoring and respiration rate.  Whoop data is very similar to those measurement tools.  Whoop's recovery measure does not correlate directly to Strava's fatigue measures which uses HR and bike power to determine the impact of the frequency and intensity of workouts on recovery.  Whoop's recovery score goes up much faster because it doesn't factor in the cummulative impact of workouts over a long period of time.  Golden Cheetah and Training Peaks use TSS to measure fatigue.

What's missing from Whoop which I get from Strava and Golden Cheetah is a measure of cumulative fitness which I have found tracks very accurately to other measures of actual performance such as Functional Threshold Power (FTP), most often tracked in cycling and the highest power you can sustain for approximately one hour. For a description of the science behind FTP see this article at Training Peaks. 
I did find that Whoop is consistently way off on cycling workouts.  According to Garmin/Strava with HR strap, today's ride was average 120BPM, 1500 calories and a 120TSS (which is a big workout), Whoop had it at 102BPM average, 680 calories and an 8.9 strain.  Both had the ride time the same at 2:40. I look forward to testing the bicep band as I now realize that clenching of handle bars and bumpy roads can affect blood flow and readings.

Since I've been skiing quite a bit this winter and wearing my Whoop, it is filling in an important hole in my tracking.  I was always completely wiped out on Mondays after a weekend skiing.  Turns out that at altitude, going to Mammoth (8000 ft) from sea level, the first night, my sleeping resting heart rate goes from 46-48 to 65 and my HRV goes from 60-90 down to as low as 25.  Backcountry skiing is also a very aerobic intensive exercise at altitude and downhill is anerobic intervals.  My recovery score went from  80 to 28 over the weekend. As a result, I know take it super easy on Monday and even Tuesdays to let my recovery score come back.

Sam's Take
I concur with Derek and Jake. I have not taken my Whoop off since receiving it last fall, literally, as I snap on the charger as needed and have become almost superstitious about having it off my wrist even for a few minutes! What I have found most stunning is the impact on recovery of strenuous non "exercise" days. Travel, long days reporting from trade shows, with their late night screen time, and less than ideal sleep and diet have produced by far the lowest recovery scores for me. I have also definitely correlated the impact of a beer to many, late meals and late computer/screen time to poor recovery scores. 

With regards my running, I have discovered the delayed impact on recovery of hard, long weekend workouts or races. I often see a strong recovery score the night after such efforts, then declining over the next few days calling into question for me the usual Tuesday speed work regimen. I tend to prefer to go hard Wednesday traditionally and Whoop has shown me why... I am also using Whoop to more carefully monitor my race tapers, seeking a very high recovery 2 and 3 nights out by focusing on some extra sleep, the night before a race often a lower recovery score due to nerves and more restless sleep. 

Whoop is simple to analyze and follow, yet is backed up by heavy duty science and very sophisticated algorithms. This combination makes it useful and effective. All in all Whoop is the best single "tracker" I have used to date as it accurately gives me the "big picture" of cardiac effort and night time recovery based on the combination of HRV, RHR, and sleep quality and quantity,  all in one simple device. I do wish the price was more accessible 
We interviewed John Capodilupo, Whoop o-founder and CTO for this article and we were very impressed with his command of the latest exercise physiology research and how it was being considered and applied to the product. I would like to see the data feed into other devices for on the go monitoring or at a minimum optionally exported to other apps and platforms, something we hear Whoop is considering.
Echoing the others. what is missing from Whoop is consideration of musculoskeletal fatigue and strain beyond the subjective questions and the recovery score. For cyclists power is a good metric as Jake says, for runners either power or some combination of power, pace, terrain, and heart rate should be considered.
Please feel free to comment, ask questions in the comments section below!
2 Whoop Straps were provided at no charge, one was purchased. The opinions herein are entirely the authors'. 

Purchases through the links below, and all purchases at Amazon, help support     Road Trail Run. Thanks!

Buy the Whoop Strap 2.0 from Whoop here
Use Coupon Code RoadTrailRun50 to save $50


Telemarker said...

In his commentary, Jake says "other measures of performance like FTP". What is FTP?

This product looks extremely interesting. All high-level athletes, particularly those in endurance sports (i.e. where training is almost all about physiological conditioning, and less about practicing sport-specific tactics and motor skills), are faced with the inevitable daily question of "am I training hard enough?" or "am I training too hard?"

Any technology that advances this from being a voodoo art based on intuition, experience, and gut feel to something more rational and scientific is going to be extraordinarily valuable. However, $500 is a pretty steep price to pay up front. If the tech works well, the price point might be acceptable; otherwise this will be a very bitter pill to swallow.

Do any of the reviewers have experience with previous products that attempted to solve the same problem (of quantifying fatigue and training readiness)? A few years ago, I experimented with the Omegawave chest strap and phone app, and found it to be inconvenient (need to put on chest strap first thing in the morning, and then lie down for a couple of minutes). Doesn't sound horrible, but everyday life tends to get in the way. And at the time, their app was buggy and the results frustratingly vague.

Oliver Maurice said...

Here are also a pretty good app for those who want to monitor their kids

Gary said...

Thank you for the detailed review. Regarding negative predictive value, if the goal here is better performance long term, not performance on an individual day of training, it would seem that for a period of time leading up to an event or competition one would need to strictly follow the recovery score "advice" and adjust training accordingly. This would give a better assessment. We would be limited by historic comparisons. Thoughts? I ask because I am also an endurance athlete that has a day job and kids, thus my training schedule gets forced by those external factors you mention above. This brings into question whether or not modifying my training plan based on the feedback of the device will lead to better performance when I need it. Thanks. G

Ultimate Phone Spy said...

Do you need to retrieve deteted messages and recovery data>> Check this website
recover deleted whatsapp messages