new Garmin Firstbeat Sports Physiology Features – What’s Next? [Speculation]
Possible Avenues for new Garmin Firstbeat Physiology Features
Garmin’s acquisition of Firstbeat undoubtedly proves that physiology insights are a great way for Garmin to attract and keep buyers of its sports tech products.
But what next? As you can see from the image above of the Fenix 6 features from 2019, there was an impressive range of features even then, yet they share one trait as they are simple snippets of information about you. No complex insights, no fancy graphs and no complex maths. Simple, single numbers. And not too much has been added since. Perhaps Garmin/Firstbeat is running out of ideas?
Perhaps not. Daily Workout Suggestions were added late in 2020 and that was a significant and useful addition for those of us who want to know what to do ‘today’ based on our current likely training status. Then the Fenix 7 added RealTime Stamina, displayed Race Predictions as a trendline and the Health SnapShot.
I’m sure there were probably a few other featurettes that I can’t remember, though it seems like we can expect one significant new feature about once a year – RealTime Stamina and Daily Workout Suggestion both count as significant additions.
Before I make suggestions about possible future development, let’s look at what you guys and girls like. The results are surprising as this poll with 100s of responses shows. The most popular feature is VO2max. Whilst we might doubt the accuracy of derived measures like VO2max, you all seem to love it!
Regular readers will expect me to suggest Running Power here, but I’ll gloss over that for now and present you with other possibilities that are more based on physiology and hence on Firstbeat’s strengths.
dfa a1
Detrended Fluctuation Analysis (dfa a1) is an HRV-based method that looks at changes in your heartbeat whilst you exercise. It’s complex maths but two simple outputs are physiologically-derived estimates of your anaerobic and aerobicthresholds. From that information, your ‘correct’ training zones could regularly and automatically be derived and updated.
Likelihood?: Garmin is almost certainly looking at this. I wouldn’t expect Garmin/Firstbeat to introduce anything here until 2023 at the earliest.
Breakthrough Performances
Q: What is a breakthrough performance?
A: It’s a personal, performance-based PB/PR over any duration and to be valid it includes the effects of fatigue and the fading, positive impact of past performances.
A simple cycling example would be a new 6-minute power Breakthrough. That breakthrough doesn’t necessarily have to be your ‘best ever’ 6-minute power but rather that it might mark an improvement on what a model predicted you could have done a month ago. Something like that. You could also have a breakthrough based on heart rate or grade-adjusted running pace.
We’ve already seen that Garmin customers like to watch VO2max increase as a mark of progress and that Garmin has changed race time predictions to now show how they trend. Both of these are examples of how we might monitor our sporty selves getting better.
But they’re both shadowy metrics. Do you really understand VO2max and how it impacts your race times? And maybe vVO2max is better but you probably haven’t heard of that one. You’ve seen race predictions before and perhaps come to the conclusion that they are rarely valid for you and the results of your training. Thus, the meaning and relevance of a breakthrough or PR/PB are more immediately obvious.
Plus, VO2max and changes to predicted race times seem to occur on a geological timescale. It seems to be ages before anything happens to either of them and that’s certainly not motivational.
Several existing sports platforms already show breakthrough performances – for those, I use Xert (free trial), Strive.AI (free option) and, to a lesser extent, Golden Cheetah (also free).
It’s more likely we will see Garmin introduce breakthrough performances sooner rather than later. They meet the criteria of being digestible snippets of information that need to be derived in a fairly complex way. These factors conspire to make it difficult for a competitor to copy and improve. Garmin has more resources to do a great job.
Likelihood?: These are likely to be introduced. It will be relatively straightforward for Garmin to rejig the behind-the-scenes maths that it already uses. Breakthroughs nicely complement existing features like Workout Labels. eg a Workout could have a label of [VO2max] and a [pace breakthrough for 5’15”]. Furthermore, a power-duration breakthrough is a more concrete stat that is mostly factual and which clearly marks progress.
Micro Adapted Workouts
Garmin already has a great infrastructure for digital training plans. For example, you can create your own plan or you can buy a plan on a third party platform and it can almost invisibly sync to your Edge or Forerunner, where you can see the plan in your device’s calendar and then execute today’s workout and have it guide you through the rigours of whatever is demanded today. Your device might even be able to control resistance levels on a bike trainer that are pre-programmed into a digital workout. Clever stuff.
Garmin does this better than anyone else right now.
However, if you’ve ever compared today’s planned workout with what Garmin produces from its ‘Daily Suggested Workout” algorithm you will often see that the two are wildly different. Which is best/correct? The answer is that we probably don’t easily know. The digital plan models the impact of historical training loads alongside progressively & deliberately increased stress, whereas the Daily Workout Suggestion will be much more short-term focussed, perhaps looking at your HRV levels today and the kind of workouts you’ve recently executed before making a recommendation.
I think Garmin has more work to do there to tie together modelled ability with your actual ability and actual readiness. However, I’m not even talking about that when I say that ‘Micro-Adapted Workouts‘ could be introduced. I would class a micro-adapted workout as a single workout that adapts recovery periods and effort durations dynamically within a workout. Xert already does this (I think correctly) as do Garmin to a lesser extent eg allowing a rest period until HR falls to a recovery level.
For Garmin to produce something similar to Xert would be a quite significant piece of work that might even require fundamental changes to the elements of workouts.
Likelihood?: Low.
Workout Compliance/Workout Scoring
Stages Cycling (Link Platform) and Stryd already have workout compliance in the platforms, as do other platforms.
Compliance is a way of scoring your completed workout against the original target. The target could be distance/duration based or it could be spending a certain amount of time in a certain zone. These are pretty fundamental things that our coaches tell us to do and which coaches will probably keep a track of. However, if you are self-coaching or following a plan, then it can be highly useful to both see how well you did in an individual workout as well as how that contributed to your overall goal. Plus a certain workout score, be it good or bad, might be the prompt that you need to analyse your workout rather than simply get in the shower and continue the day in blissful ignorance.
Likelihood?: Garmin already includes some measures around this like Garmin Coach Confidence Score but I don’t think that rigorous workout scoring and training progress will find their way to your device’s workout summary page any time soon.
11 thoughts on “new Garmin Firstbeat Sports Physiology Features – What’s Next? [Speculation]”
https://scientifictriathlon.com/tts329/ – a recent discussion of DFA1 related protocols and how it’s not currently a useful metric outside of these ramp-up protocols.
Even with Polar H10 I get very poor DFA1 data that makes no sense (which likely means poor HRV data)
Garmin don’t really support HRV based metrics well, and the HRV data / metrics are unreliable. As people complain about the OHR sensor in the watch, I don’t see how they get HRV based stuff to any reasonable confidence.
AIEndurance.com is a great idea and Markus is super impressive – but that system is currently very unreliable. Shows I should be running 2h30min marathon if I train well for 12 weeks (i’m at around 3h), and whatever setting I choose I always get the same generic plan, no matter if I choose Marathon, Half, 10k or 5k! Just some threshold runs (way slower than my threshold) and slow long runs. Apparently that’s my way to 2h30min Marathon.
Here is another CIQ data field, alphaHRV, but they struggle to replicate Kubios numbers with a simplified on-watch algorithm within the compute envelope constraints on the device. Also if you are using native Garmin HRV logging over Bluetooth (for accuracy and low artifacts) you have to use ANT+ for alphaHRV to show the real time graph, which increases artifacts.
I think this is super-interesting stuff. oHR is clearly totally useless for this analysis. It is pushing the limit of the H10 for field analysis. The only other device that is good enough is the Movesense which is meant as a hardware development SDK or for research. It’s early days.
However my advice to everyone is wear an H10, pair with Bluetooth, and enable HRV logging to collect the best data possible today. The software to analyze the corpus of historical data is coming.
yup
Have you come across http://www.trainasone.com? They have 3 very interesting stats. The first two are a Speed and Endurance score, and the third is a Goal score that is related to your personal target / race. All are displayed as ‘today’ along with recent trends. It seems that TrainAsONE’s main emphasis is about what you should do (to achieve your personal goals) and not about what you have done. Their training plans are very reactive, i.e. they adapt and change frequently to how I perform. Often telling me to do less than I would think, and I still do great come race day! Also very configurable to build a plan around personal life.
i’ve come across the name but i think that’s about it.
how you describe it seems a sensible way to go about things.
Garmin is heavily based in HR training, even though they don’t really surface to the user most of the underpinnings of the system. But then they give no guidance to novice/intermediate users how to get the most out of the system.
No help in how to set correct HR zones and how to keep them accurate as time goes by. Maybe the recommend workouts try to schedule threshold runs to trigger an auto score, but the rest is just left has it is. Does it update top of zone 3? No. What about zone 1-2? Nothing gets recalculated as you gain more and more fitness. Just a generic % max hr age that might not even be correct when you begin running, much less after 5 years.
Most friends that ask me for help have help zones set incorrectly for their fitness levels. Then of course the training effect gets skewed, to the point sometimes it’s impossible to be in the “low aerobic” training zone.
Garmin should get systems in place to estimate the major HR zones, otherwise is doing a disservice to their own training recommendations.
Rui, Garmin estimates LTHR and sets the zones according to it. It is also able to detect maxHR and change that value automatically.
They do this since years, maybew you and your friends should read the manual of the watches you use. All the thinks you ask for are available.
Stream of consciousness writing, bit of rambling, but no, Garmin doesn’t set correctly the HR zones, especially top of zone 1 and top of zone 2, they just calculate a fixed % based on max HR (which they detect, didn’t say otherwise). I can run while breathing only through my nose to a much higher HR then a default zone 2 would set by default (strange way of estimating zone 2 I know but it correlates to DFA1 go figure). Garmin doesn’t try to estimate zone 2, hence the frustrated people trying to run with very low HR to get a low aerobic effect as recommended. Sure, lots of people could benefit from running slower, but the way they have their zones fixed is doing a disservice to the users. You can’t just get a max HR, apply some % and hope to get meaningful dinamic zones as years go by. Ymmv.
IMO, Garmin automatic LTHR estimation is not nearly as good as TrainingPeaks. I think TP is the practical gold standard. Even in the free version, it will detect LTHR and notify you. Then use the Joe Friel zones tool to generate your zones and copy them onto Garmin by hand.
The Training Peaks LTHR and intensity zones correlate well with what Stryd estimates for critical power and zones and poorly with what Garmin generates — which is independent corroboration.
Turn off the max HR and LTHR detection in Garmin if you are using Training Peaks.
a lot of those Firstbeat things didn’t work for me, especially when doing running workouts with the Forerunner 945 and cycling workouts (duathlon like) with the Edge 830.
There were always problems with Recovery Time, Training Status, Daily Workout Suggestions and so on, now i sold my FR945 and never will buy a Garmin Device
https://scientifictriathlon.com/tts329/ – a recent discussion of DFA1 related protocols and how it’s not currently a useful metric outside of these ramp-up protocols.
Even with Polar H10 I get very poor DFA1 data that makes no sense (which likely means poor HRV data)
Garmin don’t really support HRV based metrics well, and the HRV data / metrics are unreliable. As people complain about the OHR sensor in the watch, I don’t see how they get HRV based stuff to any reasonable confidence.
AIEndurance.com is a great idea and Markus is super impressive – but that system is currently very unreliable. Shows I should be running 2h30min marathon if I train well for 12 weeks (i’m at around 3h), and whatever setting I choose I always get the same generic plan, no matter if I choose Marathon, Half, 10k or 5k! Just some threshold runs (way slower than my threshold) and slow long runs. Apparently that’s my way to 2h30min Marathon.
But what if it is?
I struggle with low DFA a1 HRV values measured in real time during running with an H10. I think the problem is stabilizing muscle activation as suggested here http://www.muscleoxygentraining.com/2021/05/low-dfa-a1-while-running-possible-fix.html
Here is another CIQ data field, alphaHRV, but they struggle to replicate Kubios numbers with a simplified on-watch algorithm within the compute envelope constraints on the device. Also if you are using native Garmin HRV logging over Bluetooth (for accuracy and low artifacts) you have to use ANT+ for alphaHRV to show the real time graph, which increases artifacts.
https://apps.garmin.com/en-US/apps/1a69b10a-1d31-4afe-a32f-6a579ae20d9f
I’m also intending to give Runalyze a go when I have some more free time to you around with it. http://www.muscleoxygentraining.com/2021/05/runalyze-vs-kubios-dfa-a1-agreement.html
I think this is super-interesting stuff. oHR is clearly totally useless for this analysis. It is pushing the limit of the H10 for field analysis. The only other device that is good enough is the Movesense which is meant as a hardware development SDK or for research. It’s early days.
However my advice to everyone is wear an H10, pair with Bluetooth, and enable HRV logging to collect the best data possible today. The software to analyze the corpus of historical data is coming.
yup
Have you come across http://www.trainasone.com? They have 3 very interesting stats. The first two are a Speed and Endurance score, and the third is a Goal score that is related to your personal target / race. All are displayed as ‘today’ along with recent trends. It seems that TrainAsONE’s main emphasis is about what you should do (to achieve your personal goals) and not about what you have done. Their training plans are very reactive, i.e. they adapt and change frequently to how I perform. Often telling me to do less than I would think, and I still do great come race day! Also very configurable to build a plan around personal life.
i’ve come across the name but i think that’s about it.
how you describe it seems a sensible way to go about things.
Garmin is heavily based in HR training, even though they don’t really surface to the user most of the underpinnings of the system. But then they give no guidance to novice/intermediate users how to get the most out of the system.
No help in how to set correct HR zones and how to keep them accurate as time goes by. Maybe the recommend workouts try to schedule threshold runs to trigger an auto score, but the rest is just left has it is. Does it update top of zone 3? No. What about zone 1-2? Nothing gets recalculated as you gain more and more fitness. Just a generic % max hr age that might not even be correct when you begin running, much less after 5 years.
Most friends that ask me for help have help zones set incorrectly for their fitness levels. Then of course the training effect gets skewed, to the point sometimes it’s impossible to be in the “low aerobic” training zone.
Garmin should get systems in place to estimate the major HR zones, otherwise is doing a disservice to their own training recommendations.
Rui, Garmin estimates LTHR and sets the zones according to it. It is also able to detect maxHR and change that value automatically.
They do this since years, maybew you and your friends should read the manual of the watches you use. All the thinks you ask for are available.
Stream of consciousness writing, bit of rambling, but no, Garmin doesn’t set correctly the HR zones, especially top of zone 1 and top of zone 2, they just calculate a fixed % based on max HR (which they detect, didn’t say otherwise). I can run while breathing only through my nose to a much higher HR then a default zone 2 would set by default (strange way of estimating zone 2 I know but it correlates to DFA1 go figure). Garmin doesn’t try to estimate zone 2, hence the frustrated people trying to run with very low HR to get a low aerobic effect as recommended. Sure, lots of people could benefit from running slower, but the way they have their zones fixed is doing a disservice to the users. You can’t just get a max HR, apply some % and hope to get meaningful dinamic zones as years go by. Ymmv.
IMO, Garmin automatic LTHR estimation is not nearly as good as TrainingPeaks. I think TP is the practical gold standard. Even in the free version, it will detect LTHR and notify you. Then use the Joe Friel zones tool to generate your zones and copy them onto Garmin by hand.
The Training Peaks LTHR and intensity zones correlate well with what Stryd estimates for critical power and zones and poorly with what Garmin generates — which is independent corroboration.
Turn off the max HR and LTHR detection in Garmin if you are using Training Peaks.
a lot of those Firstbeat things didn’t work for me, especially when doing running workouts with the Forerunner 945 and cycling workouts (duathlon like) with the Edge 830.
There were always problems with Recovery Time, Training Status, Daily Workout Suggestions and so on, now i sold my FR945 and never will buy a Garmin Device