Making Sense of Wearables with Sian Allen PhD | Koopcast Episode 139

trail and ultra runners what is going on what's happening welcome to another episode of the KoopCast as always i am your humble host coach Jason Koop and this episode of the podcast is going to be all about the wearables the rings the watches the straps and the bands that adorn our bodies and they capture physiological information and try to bring insight into that physiological information in order to inform your training and your racing and those of you that have these types of devices you know don't play the fool you know that sometimes that advice and very few times i would say that advice actually hits the mark but more often than not that advice is either lacking context or completely nonsensical and everybody had everybody out there listening to this podcast knows exactly what i'm talking about and so in order to bring a little bit of method to this madness in the space and i think madness is actually a very good way to present it i brought on the podcast today cyan allen who has over a decade of experience providing sports science support to athletes and coaches in the olympic paralympic and professional sports across high performance environments in both great britain and new zealand she began her work working as a physiologist with british swimming

in the lead up to the 2012 olympics in london before obtaining an applied PhD in statistical model modeling which means she's super smart a sports performance from a ut university while working with swimming new zealand and in addition to that cyan also provides a ton of insightful commentary on twitter about all of her various experiments with all of these types of wearables i'd love this conversation i actually love this space because i do think it is very promising although some of it has gone awry recently and so with that as a backdrop let's get right into it here's my conversation with cyan allen all about the wearables when i was driving when i was driving out to wherever i was going to park to record this podcast um i was thinking about all of these different wearables and i i've i've been a coach for 20 years so i've seen this entire cycle and i can't think this is with all due respect i'm gonna level out some criticism a little bit but this is with all due respect to the people who work behind the scenes because we actually you and you probably don't know this we actually worked with several technology companies to help them design various aspects of their technology whether it's what the consumer sees or sometimes it's the footprint and sometimes it's you know how they want to orchestrate the menus or or whatever so we've been involved me and my coaching staff we've been involved in in that at various

stages for several different companies so i'm you know i'm empathetic to the to the to the fight here but i can't think of another area that has over promised and under delivered in sports science as much as wearables have and this is why this is why i say that it's because the technology can be so powerful but yet because it is so powerful we tend to for whatever reason reason we try to double we the whole ecosystem tries to double down on the extraction of that information and then the application of it to to individuals and it's always been really weird because i've always wanted the hardware to advance like we want risk-based heart rate you know risk-based heart rate measurements just as a very crude example to always become much better much more accurate for people people of you know darker skin tones or if they have tattoos and things like that we've always wanted these like really great like hardware changes and in general we've kind of got them like it wasn't that long ago where i had to have a you know hockey size puck yeah to get gps data that's that was not that long ago and now the chips are just so freaking small but still even at that because the technology has gotten so or the hardware has gotten so good for whatever reason it's kind of flipped the script and we tried to extract too too too much on it so i'm glad i have you on the podcast to maybe unwind that a little bit and maybe

how that actually happens and then we can get back to some sort of sense of normalcy hopefully by the end of this yeah i would rip off that analogy a little bit and say that as a as a sports scientist i can't think of anything that has crossed the border between elite athletes and the general public as strongly as wearable technology has and i think that might be where some of the issues you're alluding to come up whereby we almost have this great power in terms of the hardware but we don't necessarily always have the tools or the abilities or the skills available to interpret everything it's giving us and you're bombarded with all this data and it's like okay where's the signal and all of this noise and the more noise you have the harder it is to extract the signal and you know i don't i don't blame people for not having the necessary background or education to be able to pick up on it and also the companies are trying their best to put things out there but they can be very generic and a bit slower sometimes than the hardware so there is what i would term this kind of huge education gap in this space whereby we have a lot of the tools but we're lacking in how we potentially implement them what we pay attention to and therefore like their utility right now because there is like such power in the tools i think it becomes even more pronounced because you you feel like you should know what to do more with them and they should be more helpful than they are because you have things available but it still doesn't really help if you don't know which thing to look at at which time and how to act on it it just makes it maybe more confusing and more anxiety inducing sometimes well this is going to be so good but i'm going to yeah i'm checking my memory card just to

see if i have enough time left over here okay um i promise i'm not going to take up all your time but we probably could go until tomorrow but before we but before we get into it yeah i have a really niche ultra marathon audience and whenever i bring guests on that are not like elite athletes or very endemic to that audience i always want to take a step back and just kind of explain who you are and kind of where you're coming from so can you briefly describe just your background and your experience this area before we start to dive into anything yeah for sure so i started out a long time ago as an athlete myself in football or soccer i was playing in the english women's premier league many years ago and then i sort of realized it was at a stage where i was never going to be able to go pro or anything so i thought okay i'll go into sports science as a bit of a backup and studied that over in the uk and exercise physiology and then i started working as a physiologist with the british swimming team and that was in the lead up to london 2012 and actually a lot of what i was doing at the time was manually collecting many of these variables from athletes that wearables today give us for free and at the click of a button so i would go to the swimming pool in the morning i would take a blood sample from athletes look at their blood glucose i would measure their heart resting heart rate heart rate variability give them a questionnaire to ask them about sleep and so all of that would take a good couple of hours in the morning with a squad of athletes and then i would spend time pouring through the data trying to make sense of it and help them tweak their training and things like that and after a couple years of doing that i started to realize hey like there's so much here i actually need better

data analysis skills to understand these data to actually help these athletes make sense of it and so that prompted me to move to new zealand and i did my PhD there with the new zealand swimming team looking at data analytics on sports performance data and athlete monitoring data and those kinds of things and then ended up staying there for for a few more years working through to the rio games and then after that i started to think about there was so much that was applicable from elite athlete data and how they trained that could be really useful for the general public and that prompted me to move to canada and start working with lululemon who are a sports apparel brand over here and they have kind of much more of a broad population remit thinking about how we can use technology and data to help people access kind of healthy behaviors yoga meditation running cycling fitness those kinds of things so i've been there for the last four and a half years doing that that was a whirlwind of background right there and the one the one thing i want to touch on that we both have in common is is we've both been in this position where we have wanted to collect data on athletes and at one point it was really arduous like it was just i just mentioned the you know just gps technology at one point you know those things and that was this was not that long ago this is within my coaching career and i'm not that old i'm only 43 where we had to use this device that was the site and this is what some of the first stuff that came out that was literally the size of a hockey puck which both people can you know identify

with we'd have to put it on the athlete and it never would sink and you know it would pick up very very spotty whenever you ran into any sort of uh sort of sort of tree cover cycling power meters were the same way they're originally wired when they came out you had to have the sensor very very close to the hub of the of the rear wheel if that's how you're choosing to measure it they constantly you know went in and out and it and my point with all of that which is something that you have uh which is something that you have described very well i think in your social media life is that there was a lot of friction in terms of capturing the data and then also trend like trend even just transmitting that data between the athlete and the coach and then kind of making sense out of it that whole process it almost made the technology i'm not gonna say unusable but it certainly made it less useful because the compliance was so hard we'd get 50 of the files or 80 of the files and there are these big gaps within the files and stuff like that and now it's kind of swung to the opposite where where the device manufacturers are are intentionally designing uh many of their products and many of the features of their products to be as frictionless as possible so before we get into that because i think that's a really that's a really poignant area for the athletes and something that you've described very well can can you paint the picture of what the fundamental advantage is of using any of

these wearables in an athletic contest i know that's a big broad it's it's a it's kind of a big broad umbrella but i think you know the space well enough and you've worked with athletes enough to say okay here are some things that we can do with these that we couldn't do beforehand and here's how it actually helps out and we'll use that as a little bit of a context for how to practically implement these things going forward yeah so i i guess i'll make the assumption that as athletes we're interested in performance optimization be it through yeah running faster or getting injured less those kinds of things and so when i think about anything from that lens i always try and think about okay what is it that the best athletes are doing and to try and benchmark that and then think about how can technology or data help us or help anyone potentially get closer to them and one of the ways in which i think wearables can help us do that is pretty much independent of whichever wearable you're wearing or whichever kinds of metrics you're looking at it's more the process of checking in with yourself on a daily basis and what i see the best athletes doing is they have this intuitive sense of how far they can push themselves when to back off because they know they might be getting injured or when they can go a little bit harder and like i've monitored some of these athletes and seen decisions that they've made validated by the data afterwards they've made a decision before i've collected some of the data and

they just have this like fine-tuned sense of intuition about when to push themselves and when not to and i do think that is something that you can grow as an athlete and i think wearables may be one of the best ways that people can actually do that and the nice thing about them is they do encourage you to open an app check your data every day i think one of the things that they potentially don't do right now that i would encourage athletes to do is to have a bit of a check-in with yourself when you open the app and this is something that i've done myself so when i was wearing whoop for example before i open the app i sat down for a few minutes had a bit of a body scan reflected on what i was doing yesterday and thought if i had to give myself a score from zero to a hundred in terms of what my recovery is what would i say it was and then i would open the app look at the score they gave me and compare the difference and start to think about hey why why might it be different and sometimes it's due to like inaccuracies or things that they're not measuring quite right and sometimes it's due to you know other factors that you have contacts context about your day that the app or the wearable doesn't have and what that process does is it helps you kind of pay attention to cues in your body which are some of the signals that sometimes wearables picking up on but sometimes say if you know what happens if your watch runs out of battery during a multi-day race or something or you lose your ring five days before an event and then you're on your own there's there's often times when you really do need to be self-reliant in terms of these data and so it's like how can you use these devices to help you

hone that sense of knowing yourself and your intuition i think that daily check-in is a nice way to do it if you you're able to kind of like add in a bit of a reflective process yourself and become engaged in it in a way so the the way that i would kind of like quickly encapsulate that is that it it it adds value to or it can add value to the to the object sorry to the subjective measurements that you already know about yourself meaning you have all that you mentioned kind of right out of the gate that athletes are pretty good at knowing when they're good and when they're bad most athletes are not everybody but most athletes with a reasonable uh with a reasonable duration of athletic contacts they've kind of like figured that out because they've been an athlete and they've experienced this is adding some objectivity to the to the entire equation you've been very deliberate about that process i mean so much so that you've hooked yourself up to all of these different devices at once and even compared them amongst each other i go back to you know the power meter analogy when all of them started first proliferating and that's what cyclists would do they'd put the polar on they put the power tap on they put the srm on they put you know five or six on they just figure out which ones were accurate and which ones weren't and and it's a very crude way of determining the differences between those you've kind of done that as well can you describe a little bit of that you know trial and error process just with you just to like once again paint this picture a little bit further yeah absolutely and so one of the things i was seeing was a lot of people uh athletes and

friends i was training with getting some of these scores come up on their devices and sort of making crazy decisions like i need to get the highest score that i can in terms of my strain or my recovery and i was thinking well hang on a minute like what's going on under the hood here what's underpinning these scores why does that not make sense to people like let's be a bit curious and try some of these things out myself so what i ended up doing was i was wearing the apple watch the whoop band and the aura ring at the same time and i had them on for over a year all together and one of the things i was doing was downloading the data from each of them and comparing the values each day to see how different or how similar they were and i started to plot some of the values do some kind of basic data analysis looking at how well they were related to one another and then thinking about when i saw things that didn't make sense like hrv from two different devices wasn't lined up for example going a bit deeper into some of the methodology to say or why is it that these values might be slightly higher from one device than another and also looking at the data for some of these things that didn't make sense related to how i felt or how i performed on certain days and using that again to go deeper and say well why is it is there something you know the device is doing or something that the algorithm is doing but it's potentially interesting for people to know about it in terms of some of the nuances for how they interpret their data as well so we'll start with we'll we'll continue with that where does it go awry

both in the actual raw collection process but then also how the athletes start to interpret it to your earlier story where they're trying to get the highest score their highest readiness score their highest fatigue score whatever they think that it correlates to performance or so or something like that let's kind of take that in two two pieces to start to discuss like the pitfalls of using all all of all of this data both from a method methodological standpoint and then also from an implementation standpoint yeah so i guess the first point at which it could go awry is at the source of data collection so to give an example if you're wearing a wristwatch that's sampling uh using ppg there so optical light sensing if you have it positioned over a particularly bony part of your wrist let's say it doesn't have good access to the blood vessels from which it's trying to read the data so you may get some inaccuracies there and different devices have different sampling rates so how often they take a particular measurement and that can affect the inaccuracy just of the very raw data and then probably the next layer is from that signal that a device is picking up the company is layering algorithms onto it to compute certain values so one of the the biggest myths probably is that every value that you see in each device is created equal so hrv from garmin and whoop is the same and that's not true because each our each company applies an algorithm first to screen for any errors

or artifacts in the data so missed heartbeats for example and then they'll apply algorithms to calculate values over certain time periods so some will use the whole night of data some will use certain other periods of data and so you're seeing different values which potentially can mean different things for different people so that that's another source there and then sometimes you're um combining these variables to give an overall metric or an overall score like a recovery score or a stress score or a readiness score and so that's actually interesting because when you if you have any kind of error earlier on in the process when you combine different metrics and different scores something happens called compounding which is basically just adding sources of error together so you get more and more error the more you kind of transform the data and by the time you get to some of those scores it can potentially be confusing because you don't necessarily know where they're coming from you don't know everything that's contributing to them how much error is involved in them you don't know if if my score goes up by two points is that something that i should pay attention to do i need to wait for it to go up by 10 points and you don't really know what's under the hood so it's really hard to interpret and to act on i'm always i'm always reminded of one of the original attempts at this or at least one of the original modern day attempts at this which was a system called omega wave that started to gain a lot of traction kind of in the late 90s in the in the early 2000s that that it was one of the at least in my

coaching career it was one of the first companies that tried to take this multifactorial approach of we're going to look at these different areas of your biology or physiology with this type of snapshot and they're using heart rate very they were using heart rate variability and ekg measurements and things like that we're going to take all these and combine them into whatever with whatever alchemy that they decided to combine it with and i'm using that word very deliberately and we're going to give the coach and the and or the athlete this stoplight system right red yellow green this is how you're you should work out hard this day you should take it easier on this day and you should take this this uh this day off and i think the the evolution of that over the course of years is a really good reflection of the industry because they took that initial red yellow green stop uh stoplight system across the across the entire athlete for everybody every type of athlete wrestlers combat sport athletes endurance athletes you kind of name it all got the same stoplight system now they've divided that stoplight system up into kind of like four quadrants right a strength quadrant a power quadrant an endurance quadrant and i can't remember the other one off of the off of the top of my head but my point with that is is they're trying to say that we can't put this into one singular stoplight system let's try to create some more detail and create a stoplight system across these different areas of physiology to which i still go back you still have the same fundamental problem you're still combining

things that shouldn't be combined into a quote-unquote score that has very little basis to it yeah and i would say probably my biggest fundamental issue with these kind of scores is the question of validity right and what i mean by that is as far as i know there hasn't been any research studies or published evidence that say if i change my training based on this score i will perform better than i would have otherwise or perform better than i would have if i change my training based on the raw data so like an hrv metric on its own or rpe for example and there are some studies that have looked at doing that with hrv successfully and so until those papers or evidences is shared i still have this big question mark in my mind of yeah like yes it would be so much easier for my life to use the score and change how i do my training but is it any better and that often comes you know months weeks down the road if you're looking at doing a big race or a big performance at the end of the year you know you don't get that validation straight away even if you feel like the session went really well so another um big issue for me with these scores is is it are you recovered what are you recovered to do or what are you ready to do and being ready to run a race or being ready to adapt to a training session or being ready to do a session without getting injured are very different things and so even if

you're doing strength power endurance those are still very different training goals and so i think there are there are a lot of subtleties and layers to it that are hard for technology to always estimate at a very individual level which is where i go back to this kind of education gap of like oftentimes needing someone like a coach to help you interpret these things to make like decisions that you need to make that are actually best for your training on that day yeah so so when i boil this down to a coaching perspective i really want to hear your your your thoughts on this is you have three basic fundamental ways that you can drive the training process and these are not mutually exclusive as a coach you're trying to you're trying to determine how much of each flavor you're pouring into the cup so to speak and some coaches will choose entirely to do one flavor some athletes will choose entirely to do another flavor and some will kind of you know blend them all together but the first one is is just looking at what the research says about the general architecture that you should deploy with an athlete kind of independent of everything else and we have all of these intervention studies with block training and high volume and low volume and vo2 max work and threshold work and polarized training and things like that to inform that type of strategy of where we're just looking at architecture right and if you look at those studies right they're they don't care about people's heart rate

variability or their body temperature or anything like that they're like this group is polarized and this group is not polarized let's put them through the intervention and let's see what happens that's the architecture group the second one is let's look at what has happened in the past to drive the future and that's the ai movement that's starting to crop up right now so let's look at all the training that an athlete has done previous to this and look at when they were doing really well and really poorly and let's aim things at the things that they were doing really really well in the future right so using this i'm going to loosely use the word you know artificial intelligence to drive the training process but it's looking at pat the past with a particular athlete to drive what they're doing in the future the third one is kind of your wheelhouse right we're looking at the physiology of present or maybe recent right is a better way to put it we're looking at the physiology of recent however you want to do it heart rate variability sleep scores recovery algorithms or what readiness algorithms or whatever they are we're looking at that to drive either in whole or in part what the athlete should do in in the future and in my estimation the failure is overly relying on any single one of those and ignoring any single one of those that you have to start to blend a little bit of all of those to really get the right to really get the right mix of it and i know that's the total like wishy-washy deal because i'm like picking all of the above but but when i see this unfold that's kind of where i've

not bifurcated it but trifurcated it into these different strategies that hit that that is that have emerged in practice over the years yeah well i think it's fair because if you think about humans we're complex organisms you know and human behavior is very complex so to reduce it down to like yeah life would be so easy if it was one metric or one particular way of doing things but if that was the case it wouldn't be any fun and everyone would be the same level wouldn't they so it's that it actually makes makes a ton of sense and i think in some of the things that you touched on there are two different things that are important to to think about one's like external load like what's prescribed or what's on paper and then the other is internal load and for the same for two athletes they could experience those two things very differently and even for the same athlete they could experience those two things very differently on different days and so even for that ai model it's potentially trained on what's been done in the past but if you haven't been injured before in your data you're suddenly in a whole new scenario and you don't have the right kind of training data in the model to inform what you should do next and so one of the other things that that also kind of creeps in is like load isn't just from training it's also from life like if you have live stuff going on you could have way more stress and that is impacting your training so it's not like one of the things that some of these apps often do is sort of assume that loading and recovery is just related to to training and that's fair enough because they can't always measure many of these things that are going on in life but if

you've had a really stressful time at work or with family life that's also baked into how you're responding and how you're kind of adapting or not as the case may be and so you have to think about those things as well which makes it yeah a complex and alive situation where you do almost need to be juggling all of these variables and asking yourself the question like does this make sense and always doing that like i think it is kind of a trial and error an iteration process really like you're sort of um yourself trying these things seeing what works um not being afraid to try something else if it doesn't work and kind of just keep like keeping paying attention to the to those trends and how you are responding and it is kind of that very much iterative learning process i would say yeah the the external loading piece has always been fascinating to me because i can remember the boom that happened when acute training load and chronic training load kind of like took center stage in the endurance uh in in the endurance space and so much so that it kind of like in my in my in my humble opinion it it it it far too greatly dominated people's thought process about how to actually train yes it is important yes it is insightful absolutely but you can't like algorithmically design a loading ramp for an athlete and have it like go true to form because you're only taking into account a single variable and then the endurance

space we get it gets all messy because you're combining load from different areas to mean the same thing endurance and neuromuscular and anaerobic and things like that that's that's kind of another story so let's move it down to practical right we've kind of like painted this picture that it can be very problematic right using all of these different devices to help inform training but i'm of the opinion and i think you are as well and you speak for this that there are a lot of insightful things that you can gather from pieces of technology in the wearable space if you curate that list correctly and so you've done a lot of experimentation on this right i'm a coach let's just pretend that you know we're in the scenario where you're consulting me on what i should do with the group of endurance athletes the group of ultramarathon athletes specifically in terms of if they could just do anything right and they kind of can they could just do anything what would you have them what data would you have them acquire on a daily or frequent basis yeah i think if we kind of assume that the the friction side of things it's easy enough to collect any of these these things or wear any of these devices in terms of um variables i one of the ones that i found most useful for me certainly was tracking body temperature and that is was useful for me from the perspective of being a female athlete looking at menstrual cycle

tracking that was really interesting but then also as a leading indicator or a kind of confirmation of any kind of illness events and for me that body temperature really was this kind of red flag variable and i know sometimes athletes can be in the space of like i'm a little bit sick should i train shouldn't i and some some people depending on their personalities will want to train no matter what and you kind of have this objective measure where you can say no you really shouldn't today because there's something going on and it's it's so obvious when there is something going on in that variable that it makes it really really useful for me um and then looking at other variables i think acutely hrv is the one that is most tends to be most sensitive to all types of stressor going on in your life and i think again you have to be able to distinguish training load from other stresses that you have uh and when you're interpreting it and look at look at some of the nuances to it but that's one that does seem to have some robust evidence in some in terms of being able to tweak training to sort of optimize adaptation so that's like day to day i think that one's really valuable and then resting heart rate is something that probably i would look at from a longer term trend perspective in terms of like looking for some of those improvements in aerobic fitness potentially and then it may show some of those big stresses in terms of like illness and and those kinds of things as well and then sleep is an interesting one because i think that's very much a behavioral one and if we talk about like controlling the controllables like you can't necessarily control how well you sleep but you can control oftentimes the time you go to bed

which is which is such a key variable for setting up how how well you can sleep so i think that's interesting because that one's quite actionable as well so those those are some of the most common metrics i think the that people are tracking some of the the foundational ones and if you're just measuring those and looking at the raw data in those that can probably be enough to get a decent perspective on kind of how things are going i want to bring out the caveat of the raw data right because ultimately the combined scores that we're mentioning they're they're taking you they're usually taking all of these and in many cases some of them and combining them in some form that's proprietary so like you said we don't really kind of know what's under the hood what gets confusing to athletes is is now you have these additional four data points temperature heart rate variability resting heart rate and sleep and it's like what do you do with them yeah exactly because you have the what's the what's the action which is i mentioned sleep is as being something is actionable like resting heart rate looking when you're looking at like longer term trends like that's one where i would just like set it and forget it put it in the background a little bit you know look at don't you don't necessarily need to look at it every day if and the the body temperature is like that red flag metric where you know probably 90 of the time you can ignore it but when something's up with it you know that's when you pay attention to it so but then you've got something like hrv where if you are looking to make an action in terms of like tweaking training and you're looking at what what state is your body and in terms of that physiological

status like if you were just paying attention to to one of those potentially you could pay attention to that one and use that um as much as uh uh it's it's a very nuanced metric in that higher isn't always better but there are some very good apps like hrv for training and those kinds of things that can really help you interpret these apps and understand for you what's normal um what's a change that's meaningful for you and when should you act on it rather than another thing that many of these other apps do is they're very generic they're not necessarily looking at how much you've changed compared to where you were last week last month whatever and so if my hrv goes up by three points is that a meaningful change for me or is that just error of measurement in the device and so one of the nice things that you do see in these apps like hrv for training is what your normal range is and so you know when you're outside that above or below it then there's something up and then you should think about okay maybe i go a bit harder or i take it up take it a bit more easy yeah from a from a practitioner standpoint i i i treat things very the same very similar where sleep is the behavioral cue and more often than not you're just trying to educate the athlete on make sure you're going to sleep at the right time and you know you're waking up at the right time and things like that and then if we have any sort of uh like sleep banking i'm going to use that word and somebody's going to yell at me for it but i'll still use it if we're using these sort of sleep banking types of interventions

we might use that as a as an anchor point for that but it's only going to it's only going to impact training if we see something super dramatic in it outside of that temperature it's almost like a it's almost kind of like a binary thing right if you like you mentioned if you see it it's a leading indicator of being sick if i see that i'm just going to back off immediately because even if it's wrong i'm going to make that training up three four weeks down the line anyway and i'd rather not just roll the dice you know that's yeah there's a risk involved yeah exactly i just look at it okay fine we'll just do it do it at some other point but then any of the other things that are getting that are being used to drive the training process i've almost just treated as like a terrain trap avoidance where if i have a really big training block or really big training load planned for an athlete and all the variables are either not normal or trending negatively heart rate variability kind of being the the king or the queen of those variables then i'll just shift things around until that is the case other than that if it's not giving me a clear like negative indicator i just kind of let the thing play out because it's just you can always chase you can always chase those variables around and try to optimize it and it goes back to my earlier statement of do you let architecture drive things or do you let physiology drive things or you just let ai drive things i don't think you can let any of those dominate at any one point in time unless they're really strong indicator

arrows so if i have really strong architecture i'm not going to let the physiology heart rate variability or whatever it is trump that unless it's super obvious like super obvious that they should be backing off and then it's like an okay deal because i know i'm going to make that training up later down the road yeah and i think this is actually a really good point because you can spend a lot of time as a coach or as an athlete overthinking that oh my god and for me like as a sports scientist we don't just think that like performance is the only currency that we're dealing in time is also such a big currency and energy and effort and you can be expending a lot of time and energy and effort on some of those like very small things where you're like hey what are the big rocks let's put them in the jar first and let's just trust ourselves and you know if something goes wrong then we learn and we iterate and we improve our process but you have to think about what you're giving up as well when you're spending so much time potentially agonizing over some of these things so i think that is like a nice approach from from many perspectives yeah i really like your like your four-pronged approach here of temperature heart rate variability resting heart rate and sleep i do think that from a bang for buck standpoint you're 90 or maybe even 99.9 of the way there in terms of how to how to change alter contextualize training with those four with those four single individual variables not alchemizing them into a score everything else i i just i think it it's just such an advanced intervention

that the use case or the use case or the advantage you start to get to less than one to a hundred so let me pick like let me try to describe that for the listeners a little bit and you can ping on this with your experience whenever i'm using those i i really do feel that it's that i'm changing somewhere on the order of about one to two per 50 workouts so one to two things every couple of months right if we want to put it on a time time perspective if i have all that stuff captured extremely accurately and that and the athlete is communicating honestly about how they're feeling whatever architecture i design i'm changing maybe maybe maybe two percent of it and so when you think about that functionally what that results in terms of an improvement it's not a two percent improvement it's like two tenths of a percent improvement right because it's not a linear you know it's not a linear gig at that point so the the cost benefit of that you could view as very high right that's a huge cost you're aggregating all this data it's a lot of my time and it's a lot of time on the athletes for you know a couple tenths of a percent gain that's worth it in some situations you know we can talk about the olympics and how you know those those performances are separated by tenths of a percent and things like that in almost all other cases not to like you know you know demean or downplay you

know everyday athletes but if you like gathering those that information if it's you know satiating your curiosity great but if you want to talk about the performance outcome you just have to be honest with how much you're trained you're changing and then how much that change impacts the actual performance with one caveat and this is why i look at this through the lens of train trap avoidance if you change something that is a really big deal you don't overload the athlete too much and they have to take weeks off or something like that missing that one terrain trap sometimes is worth all of the hassle of gathering the data because it's worth more than that two workouts that you actually are changing every couple of months so i wanted to know you're kind of like your your take on that on the you because you've written about this a little bit before on the friction versus the reward side of the equation in terms of what we can actually expect from an improvement from looking at all these things yeah and i think it's important to remember that oftentimes these devices aren't necessarily made for you as a ultra endurance runner or you as an olympic athlete they're almost made for everyone and so there are some people who know very little about the physiology that really do benefit they get like you know maybe 50 improvement or that cost benefit analysis ratio is is a lot higher and so it is it's not like completely uses for everybody but you see you have to look at kind of where you are as an individual and how much you do know about yourself and whether you use it as a

as a learning process or not um but in terms of the the friction and ease side of things i think this while harking back to when i was collecting those variables manually from athletes we were asking them to come to the pool an hour earlier than they were training which they were training at 6 a.m so they were getting up much earlier and you could argue that they would probably have been better spending the time sleeping than collecting these variables and so we didn't actually end up doing it every day we would just do it for some intense training blocks or some specific periods and so again you have to think about some of the trade the trade-offs but now that problem is much less and that's due to lower friction that some of these devices have enabled so you're almost getting like free data let's say you're wearing a watch or a ring overnight you don't have to do anything extra to it it's already collecting your sleep your hiv that kind of thing so that's what i mean by friction in terms of how easy is it to use for one how durable is it like how long will it last you don't want it these devices breaking every three months and you have to replace them how comfortable is it like can you wear it when you're training can you wear it when you're asleep without it without you feeling like oh god you know i can't sleep it's too uncomfortable and and then sort of how like um affordable is it i guess because some devices are out of the price range of of many people and so thinking about those things in terms of like yes these days it is pretty possible to kind of like set it and forget it with many of these things which can be good and bad because you want some level of engagement with it otherwise you're just wearing it and maybe you are forget about it and it's not

actually useful but then on the other side of the spectrum you have potentially people falling into this trap of like becoming obsessed with some of the data whereby you're constantly refreshing the app or you're constantly looking at like oh god what's my sleep score going to be and you get so anxious about it that it actually stops you from having a good night of sleep and so when i think about that i often think about what's the kind of temperament of the athlete that you are or the athlete that you're working with and how are they actually set up to respond to or deal with some of these kinds of things and that can be very different for different types of athletes and another thing that i've also seen or worked with athletes on in this space is using the data they're collecting as like a point of confidence so for example what i mean by that is they might actually have some bad scores some bad sleep scores some bad hrv scores and then they do a training session they actually do quite well in that session and that is possible sometimes because i think about what happens when you wake up on the day of a race and your watch tells you oh no you should take a day off you shouldn't race today and it's like when you're an olympic athlete or a high level athlete you have to do that race on that day and so how can you actually use these data to help athletes believe that you do still have the ability to perform well despite one bad day of scores and yes it's different looking at trends over time but i think there are some some valuable uses of of these data to to those extents as well yeah whenever i get an athlete that starts getting very uh that that start that that starts focusing on

some of the recovery scores or the heart rate variability scores and things like that i take the first opportunity i have and there will be one because this happens as you mentioned the first opportunity i have to point out when they shouldn't have performed very well according to whatever metrics they're looking at ready scores or hurry variability or whatever whenever that opposite that opposition exists the metrics say they don't perform very well but yet they do perform very well i i over intentionally overemphasize the value of that workout and how awesome they were on that one particular day because what i don't i realize what i don't want to have happen and i've made this mistake of not doing this before and it's bitten me that when they wake up race day and they look at their readiness score and it's i mean you might as well just not race at that point like it's just like it's so it's so psychs them out and that happens in training a lot and i've just found with athletes that get really involved in the data you have to do as a coach you have to do a very deliberate job of pointing those things out as you mentioned as a point as a source of confidence such that when it does happen on race day it's not freaking them out so much because we you and i know we know these case studies that don't get circulated uh very widely and uh amongst the popular press but they kind of get you know regurgitated internally amongst practitioners and coaches of the hundreds of elite and olympic

athletes that either had their heart rate variability or whatever blinded to them and they won a gold medal they they won the world championships they said a huge pr did some world championships final or whatever like all of those stories are out there and we need to be very cognizant of that oh yeah and i can tell you when you go to an olympic games you know you get put in a tiny room and you're sharing with a roommate no one really sleeps well in an olympic game so like you have a lot of people still performing well and you don't see that behind the scenes and it's interesting actually because i had a practitioner from one of the professional sports reach out to me and he said we we built our own dashboard whereby we just hide the values from our players and we like we just look at them as practitioners and if they need to see them the athletes then we show them you know otherwise we just have so many people second guessing things or you know freaking out about it and so there there are certain other softwares i think do that like third-party softwares with some of these big apps and yeah i think there are some of those interesting things that will people will start to create them as well to deal with some of these problems well i my my hope is is once all the once all the dust is settled the pendulum that we were talking about earlier kind of swings to the other side right the pendulum where we're combining things into nonsensical scores we're collecting things either in way we're either collecting things that are not valuable or even counterproductive or we're collecting them in ways that are counterproductive or not valuable but in all of that is an is in an effort to create value out of the product let's not like mince words here you know you're as you mentioned part of the

friction with some of these products is that they're expensive and when they're expensive one of the things that you just have to do as a device manufacturer and i understand this plight is you have to present your product in as valuable as a context as possible and part of creating that your look right now is hilarious i hope people are watching the youtube version of this part of creating that value is first off doing something novel right so recovery scores are something novel or you create a you know here's your purple score kind of whatever it is something that's unique to the device itself and the second thing is is just the volume of data points the types of information that you're that you're collecting and also the frequency at which you're which you're collecting them i get i get that plight i get that plight with device manufacturers but what i'm hoping is is eventually there's some sort of normalcy to like come come back to the middle and what i want to know from you because you've seen all the form factors you know them very intimately you've seen all the different ways of collecting them how would you do it like if you were to create the pendulum in the middle right that you would use for you know the whole basket of athletes kind of where would you start it's it's it's a good challenge and i think um it probably speaks to customization because to go back to what we were saying before right now these devices are made for everybody looking to diagnose sleep apnea to elite athletes and so in that you have very little personalization and so what i think

we'll start to see and this may be with the same hardware but more on the the software side um helping people and maybe it still collects all these variables from you but on the software side it's saying hey like you need to pay attention to this one variable today because you're out of your normal range and your goal is this you want to be ready for this so this is the recommendation that you have and it's just way more personalized to like every individual there's better um data in the algorithms that um is more personalized to you and so i think there's potentially more of that evolution on the software side than the hardware side going forward um and that's probably it's it's it's nice as well because so many other like third-party developers and companies can come and like i know the ultra running market really well let's say so i'll make the app that works best for my athletes because i know what they want they need to see and how they need to see it i'm with you maybe we can get some funding for that but then we're the same problem as everybody else's we have to create too much value around it i think we can start here's what i've always here's here's the thing that makes me bang my head against the wall i think we always start with just sport group and then you can individualize within each sport group but the fact that you know you have a wearable that is treating like you said somebody that's trying to lose 80 pounds with the with the same kind of information dump as an olympian much less a runner

versus a combat sport athlete then you can divide that kind of anywhere like just just start with the sport groups because that's really not that hard i mean yeah i mean i think it's it's the process that you describe that you're doing as a coach probably many other coaches and scientists are doing it within their teams or groups of athletes it's just sort of making that like productizing it essentially which you know you could say that's your competitive advantage so why would you want to do that so that's probably where it's one of the other reasons it's got stuck as well but i think the knowledge is there it would just be extracting it and turning it into something that makes it accessible to people i totally i totally agree it is there i mean i i do think that every all the all the coaches that i know across all the sports have their own you know dashboard to use that to use that word they have their own kind of like dashboard whether it's like in their head or whether it's like a physical you know dashboard that they've kind of made to harness these variables and make use out of it for their particular athletes i think that's the first step and then the second step is then within whatever your cohort of athlete athletes are individualizing it across those athletes to make sense for them in particular with whatever else that you're using architecture and previous training and all that other stuff that that i kind of mentioned but it always has seemed to me that it's that it's not that big of a leap and also advantageous because you know how people want to get spoken to you know by their tribe members you know they wanted they wanted triathletes want to get spoken to by triathletes runners want to get spoken to by runners you know kind of so on and so forth i also think it makes the

devices stickier if they somehow can change the the i guess the language is is primarily what i'm focusing on the language around the data interpretation for each of the each of those sport groups and that i don't think that's big that's that big of a leap yeah yeah absolutely and i i think there's um you know there needs to be some incentives for that to happen probably uh in terms of as you say like potentially the money is more in the like medical space and in the the sport space and right now but yeah maybe probably in the future i think there will be more personalization more customization and more uh things that are available to people like specifically related to the goals that they're going after yeah it feels like like they're like we're trying to serve two masters right we're trying to individualize it for everybody and we need to just like solve like we need to like narrow down the problem first like let's just get people let's get that let's get the athlete buckets first and then we can work on the individualization but trying to do both at the same time it's kind of i mean it's kind of it's kind of proven out already that it's just a really it's too complicated of a problem to solve right now because the advice that is being dished out in many cases not all cases but in many cases it's so generic for everyone that it's specific to no one as you mentioned at the onset of this podcast yeah you try and be all things to all people and you end up being nothing to no one and sometimes

you think just give me the raw data you don't need to send me a message that goes with it right now because you might be doing yourself more harm than good in terms of how it's interpreted you know it's easy to even insult people or make them feel and not very good about how the training went and those kinds of things that you know potentially they don't necessarily need to do that right now so yeah like as much as as much as you can i would say spend more time looking at the raw data and um trying to make sense of it with with a coach or someone who actually has more of your context so they can make those decisions or recommendations more so suited to you and the quintessential example of that is what i sent you before we got on the horn here it was a message for it was a message for my garmin the day after i ran the hard rock 100 which is a 33 hour run for me telling me that i had i did not have a very restful night's sleep so i should recover more which the advice is actually correct right i mean that's like sure i i'm absolutely gonna do that i'm gonna recover yeah i can't fault you there but my sleep was shit because i ran all night you should probably know that because you can see that i was doing that anyway that's a brilliant place to leave it i i really appreciate this conversation this is this is really fun i think we go all day maybe we'll have to have a second round once the technology improves by a step then we can kind of come back oh yeah yeah yeah we'll have to re-evaluate and how we use it then and that's that's a great place to end it because i think these things are improving rapidly all the time and so it is

very much a moving space and some of the problems that exist today might not exist tomorrow and i am very optimistic about what that future looks like and a lot of the rationale behind me sharing some of these things is to try and help drive these things forward and make them better for people and for certain groups so yeah just thanks very much for for having the conversation and i would just encourage people to be curious and be open but be a little bit skeptical if something doesn't make sense then there's probably a good reason for that if something's too good to be true there's probably a good reason for that as well so use your use your intuition and um yeah just do what do what works for you that's brilliant you you're a very insightful twitter follow i get a lot out of just the the content that you put up put out on that platform where else can uh the listeners find out more about you and where to follow you i think twitter is the best place i'll just try and share as much as i can about what i'm learning and my kind of observations and experiences on there so it's dr sean allen on twitter and i've got some of the threads that you mentioned in the past comparing different devices you can find them on there and reach out to me as well if anyone has any questions awesome thank you for coming on the podcast i really appreciate it and keep cranking out that content it is gold really good stuff absolutely thanks for having me all right folks there you have it there you go much thanks to sign for coming on the podcast today and trying to untangle this big mess that we've gotten ourselves into in terms of all the data that

we are collecting from our bodies through these various wearables and how to make sense of it all and after i got done recording this podcast this one thing uh very abruptly came to light to me we've been trying to alchemize things in sports science and in endurance sports for forever and we can go back to acute training load and chronic training load and kind of the original genesis of this which is the trimp score which is a heart rate based mechanism of quantifying training load and in those cases you're taking all different types of load whether it's a cardiopulmonary load or a muscular load or a neuromuscular load and you're essentially trying to combine them into one score or make sense out of it via one mechanism and when i first started working as a coach and we started using chronic training load and acute training load which are terms coined in the training peaks vernacular we started realizing very quickly that that sometimes the intensity component and the volume component of that load which were combined into the scores did in fact need to be separated and i'm starting to think that this parallel of trying to separate all of these different types of variables where you are collecting them all and they all mean different things across different time frames i'm starting to think that that parallel is going to hold true in the recovery space and in addition to that the wearable space where we're collecting all these metrics and trying

to combine them when in fact we need to be doing just the opposite separating them out in order to extract their true meaning we'll see if the space goes this way like i said it is very early stage i think there is a ton of potential out there but we need to be careful on how we use all of this information so that's the quick caveat for all the athletes out there i like these devices but i think they need to put be put into the right context i appreciate the heck out of all you listeners out there if you love this podcast please feel free to share it with your friends and your training partners and your loved ones i always appreciate getting this out to a much broader audience and one of the great advantages of not taking on any sponsors is i can bring on people like cyan and we can talk about these wearables and the information that they are trying to extract with a completely unfiltered and honest tone so that you guys ultimately get the best information that's it for today folks and as always we will see you out on the trails you