Adjusting training plan based on recovery metrics? HRV, Training Readiness, Recovery Time etc

Triathlon Forum
1

Looking for thoughts from athletes and coaches experienced in this area, do you regularly modify your training plan based on morning recovery metrics (such as Garmin Training Readiness, HRV, Body Battery High, and Recovery Time, or the equivalent from Oura and Whoop)? As an example, let’s say you have a threshold run on the calendar for today as part of this week’s training schedule. You wake up and your Garmin Morning Report tells you have a Training Readiness in the low range (1-25) and/or your overnight HRV has dropped below your 7 day average. Do you change your plan for today’s training? Or do you power through like you used to? I am wondering if there are improvements to be gained by making modifications based on these metrics or if they should only be considered as a validation for when I feel really lousy, and otherwise just ignore the variation in recovery metrics and stick to the plan. Have you had good experiences where the recovery metrics drove training decisions and this seemed to be an improvement (greater training adaptation, less illness, more and faster race finishes) over the old fashion way of just listening to your body, and bringing a strong mental game to just show up and do the work.

As context, I have been using HRV and Training Readiness for about 18 months, all of which I was not training for a race. I’m trying to figure out for my races in 2024 if Training Readiness and HRV are useful inputs for training decisions or more of a distraction that could lead to not train as hard as I have in the past (When Garmin tells me I have “3 days” of recovery time after a couple of hard workouts, it tends to make me take the foot off the gas for a little while.)

2

I don’t. I’ve found them to be, in general, accurate but not close enough for me to trust 100%. By them I mean HRV and resting heart rate. For me, resting heart rate seems to be the most accurate. Tired = high. Rested = low. But it can also be influenced by travel, heat, food, hydration, etc so it’s not quite reliable enough for me to say okay 5 days of high HR, time for rest. Though I’ll use it as a factor to consider when also looking at the training.

3

I had a period of 6 weeks or so where I was feeling exhausted and struggling with my training. Making no progress, feeling drained. Then I started feeling a bit better and things went better. Looking back, I can see this correlated with a massive dip in my HRV from averaging mid 80s, to low 60s with some 50s in there. I was in a period of particular stress at work, and I am sure retrospectively that I also had some illness that took a long time to shake.

So from this then I don’t especially use the training readiness, but it does seem to be pretty close as an indicator to how I feel when I get up, but I do look at the last week or so of HRV and use that as an input to lifestyle choices (make an effort to get early night, get some down time) and feedback to my coach in terms of if we can push on, or might want to take a week or two in ‘steady recovery’ and focus on quality over quantity.

As a side to this then can I expand the question to the Stamina during an activity? After a lot of sessions then I’m finding it not too bad. I think it tends to over estimate stamina at high intensity, but is pretty accurate for endurance (eg +2hours run +4 hours bike). After a really long run or ride then if I retrospectively look at the data if I was at the ‘I’m done, just need to drag my arse the last 5km to get home’ sort of activity then either the stamina ran out at 3km to go or was on 1 or 2%.

4

In theory, trainings plans are designed with a prediction of how much load/stress a body can handle (in the form of weekly training duration and intensity distribution.) Of course that varies by age, health etc. If a plan pushes too hard for an individual, then they are overtraining which is detrimental (leading to a slowdown, reduction in energy, higher threats of injury or illness.) But a training plan is just a static prescription of training load and prediction of the individual’s (body and mind) to absorb and grow from the training. What the training plan and even TrainingPeaks TSS/ATL/CTL don’t know is how your body is handling the training load. Is the training putting your body in unproductive overtraining state? Or perhaps the training alone isn’t of a level that would put you in an overtrained state, but other parts of your life (such as lots of travel, jet lag, poor sleep, personal or work life stress, a chronic or acute injury, booze, etc) accompanied by the training are putting you there. It seems to me that an individual with a training plan (and a strong sense of commitment and no access to HRV (and the other recovery metrics) or a coach to notice declining performance) might not be able to acknowledge that the sum of all of this (training + life stress) might be putting them in an overtrained state, where they are no longer adapting with increased speed and strength outcomes. And HRV (and the other recovery metrics) might be a big improvement in this area where now it’s not subjective of “how do I feel today”, but more objective paying attention to the recovery metrics which will help 1) to improve one’s intention to optimize recovery with good sleep, lower alcohol consumption, relaxation, naps, etc and 2) to do align the hard workouts for the days you are actually properly recovered from the previous training (including the ways life and work stress can delay recovery.)

It seems to me that in the future a platform like TrainingPeaks would be able to take the HRV/RHR recovery data and recommend or make adjustments to your training plan according to the recovery data, and that these small adjustments every week could in theory give someone a greater training outcome than the Training Peaks training plan experience today.

Anyone else have any experience with this? Is the technology still too early? Anything to watch out for?

5

After spending darn near a decade trying and failing to find HRV to be meaningful, I feel that my Garmin 965’s HRV is actually meaningful. I avoid maximal efforts of any duration when it’s definitively low.

6

Looking for thoughts from athletes and coaches experienced in this area, do you regularly modify your training plan based on morning recovery metrics (such as Garmin Training Readiness, HRV, Body Battery High, and Recovery Time, or the equivalent from Oura and Whoop)? As an example, let’s say you have a threshold run on the calendar for today as part of this week’s training schedule. You wake up and your Garmin Morning Report tells you have a Training Readiness in the low range (1-25) and/or your overnight HRV has dropped below your 7 day average. Do you change your plan for today’s training? Or do you power through like you used to?

I don’t think HRV is “there” yet to be a metric that drives your training. The metrics are specific to no one, general to all.

The problem, imo, ymmv, with the above thought process is the tail is wagging the dog.

Maybe this is me, ok well I know it’s not bc I talk with enough coaches.

There are 2 segments/types of athletes. Those that can feel what is going on. You can tell them go run 3x800m at 10k effort or half marathon effort or threshold and they can go do that. Or you can tell them 5x15 min threshold on the bike and they can do that. they can properly pace 90-95% of the time.

Then there is the electronic generation who has been told by their watch, hrm, other devices what they are doing, how fast they are going, what zone they are in. They know they ran their last flat 10k at 6:31/mile. Then they get to a hilly 10k and try to run 6:31 mile and blow up 2 miles in bc the first 1.5 mile were hilly AF. They wonder why the first 1.5 miles were so much harder (it’s the uphills - u got to be smart on those early in the race btw). They struggle to feel what they are doing, they struggle to tell the difference in effort between a 20 min threshold and tempo. They don’t really grasp that if their threshold swim speed is 1:22 an 800 at that pace is going to be a lot harder than a 200 at that effort/pace.

They struggle to understand that the physiological ask is greater as the duration increases even if the intensity/speed/effort is the same.

Do you let some device modify you workout before you even walk out the door and try it or do you decide your workout(s) and then modify based on how it’s going, how you feel, the results you’re getting?

The tail wags the dog in the first group. The dog wags the tail in the second group.

7

FWIW, my opinion is the implementation of HRV in Garmin devices is very bad. Overnight average HRV, which is what they use, is going to be highly impacted by things like the size of your dinner, and has low correlation with readiness to train the next day. As a very recent example, Friday night I ate a much larger than normal dinner because I had a marathon sim workout the next day. My overnight average HRV was very low because of the dinner, though when I looked at the graph I could see that it started low and finished the night high. I expected to run well, the Garmin told me I was unbalanced. I had an exceptional run.

Whoop used to do something equally stupid, taking your HRV during some stage of a sleep cycle (which of course is not accurately measured to begin with). The best implementation for evaluating training readiness is probably putting a HR strap in the morning and using HRV4Training. In any case, my n=1 is that HRV, whether implemented poorly or not, seems to have low correlation with the quality of my training that day. I’d say at best use it as one input into the equation, incorporating how you feel as the primary metric, recent training load, life stressors, etc.

8

I think there is a discussion to be had about the future of optimal training with AI-based modifications to plans based on actual recovery metrics, but I’m also just interested in guidelines for 2023 about how we can utilize recovery metrics to modify the standard training plans (I use 80/20) week by week for optimal results… Some of those guidelines might be (I am making these up):

  1. “When you see a decline in HRV and Training Readiness, if that aligns with how you feel, then eliminate the intensity in training for the day, keep the duration of the workouts as originally prescribed if you can, and plan to return to increased intensity in the days ahead.”

  2. “It’s natural to see a decline in your HRV and Training Readiness during this week of the training plan. Don’t be alarmed. Work hard, you’ll be tired from it, and keep in mind that the first half of next week will be a lighter load allowing your body to recover and your HRV to rebound.”

  3. “If you want to use accurate overnight HRV to make adjustments to your training plan based on your body’s response to the training, it’s critical that you do X,Y, and Z test and make sure the ___ settings in Garmin are set to this, so that your recovery time and Training Readiness are as accurate as possible.”

  4. “We have found that Garmin HRV (up through Forerunner 965) is relatively consistent and reliable, but that the Recovery Time feature assigns too much recovery time for extended Zone 3 workouts such as RT, and CT. Don’t be alarmed when your watch tells you you have 58 hours of recovery time after a CT workout. If you feel ok the next day, train as prescribed.” -OR- “We have found Garmin’s Recovery Time to be relatively accurate as an indicator of intense workouts having an extended impact on the body, requiring an extended period limited to low intensity training.

Thanks! Looking forward to the feedback, guidance, and other’s experiences. I know this is new, but surely folks have begun wondering how recovery metrics and training plans intersect.

9

FWIW, my opinion is the implementation of HRV in Garmin devices is very bad. Overnight average HRV, which is what they use

Maybe at some point they did, but my 965 gives me four metrics - 7-day average, overnight average, instantaneous, and highest 5-minute average.

The nighttime and 7-day tell a reasonable story to me. The 7-day is a lagging indicator, and the nighttime and instantaneous a (noisy) leading indicator.

But the story they tell is starting to make sense to me more than my last 5-6 years of trying 5-6 different methods .

E.g. for the below, when my 7-day average was triangles (bad), that was when I buried myself unnecessarily on a backpacking trip. Didn’t eat enough. The squares were me letting myself recover. The circles were me getting back at it with a heavy training block. The second section of squares is me doing a lighter week last week. I had my first return to a “circle” today, which coincides with the first day of another heavier block.

You’re right, the nightime average (dashed) jumps all over the place, but it correlates well enough as a leading indicator that I can typically get a good idea of where my 7-day average is going to go the next few days.

The best implementation for evaluating training readiness is probably putting a HR strap in the morning and using HRV4Training.

I disagree with this for the same reasons you posted above. You’re taking one spot-check of your HRV, subject to all kinds of factors. And it’s a bit of hassle. (I did it for a about a year.)

myhrv.png

10

The 7 day average, which is what goes into Garmin training readiness metrics, is just the ovenight numbers averaged. Averaging 7 days of something with no value does not help add value. Maybe it’d be useful if they used the 5min highest number instead?

The morning measurement gives you the most apples-to-apples comparisons, devoid of stupid things like having a big dinner, and is closest to future training. I also did it for a year though and stopped because it wasn’t adding any value. FWIW, I’ve heard Olaf Aleksander Bu say they don’t find any value in HRV and don’t use it.

11

I have a FR955 and a coach prescribing workouts a week in advance. I do check those Garmin metrics every morning - HRV, RHR, sleep, training readiness (which is a mash up of those parameters and others). But I never modify the pre-planned workout - up or down - based on my Garmin metrics. So, why do I even look at them? I don’t know…

But the Q, I guess, is - could my training be more effective if I ramped it up on high readiness days and took it easy on low readiness days. I’d be interested in data from that experiment, but it’s hard to imagine how you would go about getting interpretable data. You would need a control (or controls) that would not consider training readiness. Some way to normalize for workout intensity. And what would be the output measure? Seems as if it might work for relatively undertrained folks aiming to improve a specific performance or measurement metric (800 time, FTP, etc.). Even then, probably a multi week (multi month?) experiment with lots of hard-to-control factors. Still…interesting to think about…

12

The 7 day average, which is what goes into Garmin training readiness metrics, is just the ovenight numbers averaged. Averaging 7 days of something with no value does not help add value. Maybe it’d be useful if they used the 5min highest number instead?

You might be right there. Tough to find a detailed explanation on Garmin’s site. I disagree about your assessment of averaging, though. Moving-window averages are a tried and true tool for suppressing noise to find a signal in time series data. I use them all the time in my day job as an electrical engineer (though typically in fancier forms than a pure moving average). A moving average is in the toolbox of low pass filters. And with HRV the theory is there’s some kind of slower-moving signal among all the spikes throughout the day.

The morning measurement gives you the most apples-to-apples comparisons,

I didn’t find that to be true. I did it for a year and found it to correlate to absolutely nothing. I could be in a huge hole or doing just awesome and I sure couldn’t find any indication in HRV morning reading. And I tried to do it “right”, being motionless, picking the exact same time every morning. I disagree that the morning is any more free of “lifetime stresses” than any other time of day. At least for me.

I’m somewhat convinced that some kind of continuously sampled reading is going to blow away single spot-checks.

And without having read the research, it seems that nighttime would be a little cleaner data than daytime when you’re working out and have all kinds of other stresses. At least intuitively. I’d have to go read “the literature” to learn more.

This Garmin is the first time in over 5 years I’ve had the data tell a story that made any sense to me at all. Including spending about a year with every generation of Whoop.

13

I enjoy collecting all the Garmin data, but I don’t make any adjustments to my training based on that data. I got off my 5 hour bike Saturday and Garmin wanted me to take 3 days off while I still had a 30min run to do and 2 hours of running on Sunday. On Sunday morning my Training Readiness was in the toilet, and I ignored it and had a great run.

14

no way. just smash yourself every day and that will teach you to toughen up. dont make no excuses. more coffee, less sleep. you only are training enough when your dong stops working. train through niggles. train through colds. go hell for leather. get shit done. :slight_smile:

15

I enjoy collecting all the Garmin data, but I don’t make any adjustments to my training based on that data. I got off my 5 hour bike Saturday and Garmin wanted me to take 3 days off while I still had a 30min run to do and 2 hours of running on Sunday. On Sunday morning my Training Readiness was in the toilet, and I ignored it and had a great run.

That’s reasonable, but pointing out there’s a big difference from the raw metrics like resting HR, chronic training stress, HRV and Garmin’s more fudgey “Training Readiness.”

I tend to use the data to confirm how I feel. If I feel like shit, my resting HRR is high my HRV is low, and my chronic stress load is high…well that combines into a story that makes sense.

16

the Garmin recovery score appears to be mostly based on HR during the workout, so it’s not very useful.
Recently my Garmin had some kind of problem with HR and was showing inaccurate low HRs. After a 25k trail race, 3000ft of climbing, with HRs shown inaccurately as 80-90, it gave a recovery time of 3 hrs… so it seems distance, time, and other measurements through the day of HR etc are not much used in the recovery time calculation.
Normally a 5mi training run in Zone 2-3 with accurate HR 130-140 gives a recovery time of 20-30 hours. That seems on the high side. So I don’t have any faith in that metric.

I find the arguments against averaging HRV quite convincing, so a single standard morning or evening measurement is more likely to be useful.
For me it tracks closely with perceived effort and my ‘gut feel’ for readiness and is a useful warning system. On the days when I both feel bad and the HRV is trending low, it gives me enough reason to back off effort for that day - where normally I’d just try to push through…

Before I got old my own ‘gut feel’ was enough to regulate training. Now I can’t reliably tell the difference between exhausted-because-old and exhausted-from-workout/life stress, pushing through has not worked well.

The HRV4Training app does the morning measurement, and also has a set of questions to answer about training readiness. I’ve found it useful to be prompted every day to think about those questions. As desertdude says, in the long run it’s better to be in tune with your own perceptions and use those, than reliance on external metrics, especially the black box metrics that Garmin etc generate out of whole cloth.

17

t
I find the arguments against averaging HRV quite convincing, so a single standard morning or evening measurement is more likely to be useful.

What are the arguments against averaging HRV?

And how are they ameliorated by taking an instantaneous sample once per day?

Edit: Been trying to do “ma resurch” and so far have found this one study which directly compared nighttime averaging vs. a single orthostatic measurment in the morning. They found no differences, concluding, “The present findings indicate that the accuracy and reliability of weekly average values for HR and RMSSD obtained during sleep employing a BCG-based device is acceptable. Thus, day-to-day monitoring of nocturnal HR and RMSSD appears to be a convenient and valid approach for long-term monitoring of young endurance athletes.”

There were no significant differences (according to them).

Of course one study is just one study.

But maybe I was doing my morning orthostatic testing wrong. Because I found it to lack any apparent correlation to…anything. While my nightime averages seem to follow what I’d expect them to given my training patterns.

And with all physiological stuff there seems to be a ton of individual variation. I wouldn’t prescribe anyone else what (I think) works for me. Just describing my experiences.

18

Furthermore, the morning 5 minute measurement is subject to the “measurement phenomenon”…ie, the measurement is perturbed by the act of taking the measurement. Conscious awareness of the test, causes some type of autonomic reaction that messes with the data.

Like you, I’ve spent a ton of time with HRV4Training (about a year), and with the garmin metrics (once they became natively available)…HRV, and many of the other related metrics. I did all the same things you did, same time every day, same routine, etc. Same useless data results.

I’ve never found them to tell me anything useful, or anything that I didn’t already know. HRV is mostly a random number generator to me. Often, but not always it moves in the opposite direction (for me) from what most report following hard training. Eg, often it will be 50% **above **baseline the night after a hard day. But, sometimes it may be 10% low or even “normal”.

At best HRV agreed with how I felt, at worst it didn’t. But, I couldn’t ever rationalize NOT training when I felt good, simply because HRV “said so”. Never once, have I gone out for a workout based on how I felt (thinking it would be good, only to go like crap), and later looked back at the HRV data…and went “oh…that’s why that workout went so shitty!” Honestly, I find better correlation to RHR than HRV…but, the same thing applies there. At best, it tells me what I already know.

If I blindly followed HRV, I’d have done the wrong thing more often than the right one. Similarly for the recovery advisor and training readiness.

I will say though, that the only time HRV could have been useful is when I have come down with the flu. HRV dropped through the floor…by like 50%. I felt fine when I work up, but I was sick as crap by noon…and at the clinic getting tested. Never made it to my workout session for the day. In retrospect, HRV was a leading indicator in that instance.

19

Furthermore, the morning 5 minute measurement is subject to the “measurement phenomenon”…ie, the measurement is perturbed by the act of taking the measurement. Conscious awareness of the test, causes some type of autonomic reaction that messes with the data.

Ha! Yeah, I remember in my year trying the morning orthostatic measurement I was always telling myself, “Don’t think about work stress, don’t think about work str…DAMMIT”

20

What are the arguments against averaging HRV?
And how are they ameliorated by taking an instantaneous sample once per day?

arguments against:

  1. measurement error. The reliable way to get HRV is a chest strap, Garmin and other wrist devices are estimating at best.

Because these distant cousins of the heart rate strap use light to infer heartbeats, they are subject to interference from other sources that might affect the light hitting your wearable receiver. Consequently, many of the algorithms that calculate HRV are filtered more aggressively than those that calculate HRV from a heart rate strap. This combination of added interference coupled with more aggressive filters can lead to differences in the measurement. For this reason, I always recommend sticking with one particular app/wearable combo and strongly recommend using a heart rate strap.

Alan Couzens, at https://simplifaster.com/...s-hrv-training-data/

  1. hrv fluctuates widely through the day and night, averaging over a lot of spikes may not be useful.

Alan again,

An important thing to note about HRV is that it scales negatively and nonlinearly with heart rate. For example, if my resting heart rate is 30 bpm (i.e., an average of two seconds between beats) and yours is 60 bpm (an average of one second between beats), there is much more room for higher time variability between the beats for me than you. This is even more true for periods of increased heart rate due to exercise or stress—when heart rate variability can effectively decrease to zero due to the impact of the sympathetic nervous system.

With all of these influences and how heart rate changes throughout the day, it is challenging to interpret all-day HRV in any meaningful way. It is far better to have a controlled, resting test where your heart rate is relatively similar to assess significant changes in HRV.

He’s been doing this for well over a decade with lots of different athletes. I trust his data and opinion.
Marco Altini has some published research on this too, don’t have the links at the moment, would have to search.

HRV is more about measuring systemic stress, than training stress. For amateurs like us who have to work and deal with family, those stresses are typically overwhelming any training effect.

morning HRV responses (w/effect sizes) to stressors associated with…

training: 4.6% (small, d=.36)
illness: 10% (mod, d=.47)
high alcohol: 12% (mod, d=.55)

9M observations from 28k subjects via @altini_marco https://www.mdpi.com/1424-8220/21/23/7932