thrlskr:

I wish I had good news for you regarding the use of the Firstbeat Heart Rate Variability (HRV) based metrics such as EPOC, TE and Recovery Time. While the theory behind using HRV and EPOC for tracking training seems to be solid, my experience trying to use several First Beat HRV and EPOC software programs embedded in many different watches leaves me dubious of their effectiveness. I have no experience with the Fenix 5x so I can’t speak to that watch and any improvements to the FB software it might include. I started using HRV software in about 2007 with many of my athletes. At that time the only watches that incorporated it were the high end Polar ones. They contained a program called OwnOptimizer developed by the grand daddy of HRV, Heikki Rusko. The OO program worked fairly well for predicting readiness to train. But it involved a 10 minute test and was to prone to failing in mid test. So it was a hassle for the athletes. I used it for 3 years on a group of athletes.

Rusko went on to start Firstbeat and the man surely knows his stuff when it comes to HRV and collecting neurological, metabolic and physiological data concerning training stress. Trust me when I say that I wanted to believe that this technology would at last give coaches and athletes some objective data to go along with the subjective feelings we normally have to rely on. Since Polar stopped the OO program I have tried several watches by Polar, Garmin and Suunto. The second 2 both buy their HRV software from Firstbeat. I have also tried various phone apps as well.

My conclusion is that these things work fairly well (70% accuracy) for back casting the data. But only about 40% accuracy for forecasting, which is after all what we want them for: To inform our decision about the training TODAY. So if you look at the Firstbeat recommendations in the past you will see some trends in your perceptions and actual fatigue and workout quality. But the trends are not consistent or repeatable enough to have any predictability when looking into the future.

Simple examples may best illustrate what I am talking about;
1) I have had athletes get a big red flag warning that they need to rest that day and then proceed to have one of their best competition performances or breakthrough workouts. After this happens a few times you loose faith in the products reliability
2) On the opposite end of things, many times I have had the FB software predict that a particular day was going to be a great day for a hard workout and the athlete felt tired and clearly needed a rest.

For long term trends in fitness, the technology may prove effective where the scatter of daily data is smoothed out. But for short term day to day use of effectively predicting preparedness to train I have been very disappointed after 8 years of hoping and trying to make it work.

My personal theory about why HRV does not work so well in the real world is that we are very complex organisms. HRV looks only at the Central Nervous Systems response to stress. While that is clearly important, it is only one of several other systems that clearly have strong effect on on your ability to absorb more stress: The endocrine system perhaps being the other most important.

Our friend and colleague Scott Semple who posts here often has extensive personal experience with trying to use HRV predictively. He collected data and made a long term graph of the accuracy of it. I’ll see if I can’t get him to post something here as well.

Your best bet will be to try your Fenix 5x out and see if the data it gives you jive with your feelings in the short and long term basis. I honestly hope that they do.

Scott

Thanks, Scott.

@thrlskr: As Scott mentioned, I did a 2-month “study” comparing ithlete Pro, Elite HRV and an orthostatic heart rate test similar to what Polar recommends. Alongside the HRV app recommendations, I tracked training loads and how I felt in general.

Yes, with a sample of one, the results are anecdotal, but it was enough to tell me to delete the apps from my phone.

In short, the orthostatic test never failed, but it does take some getting used to and some ongoing interpretation. This is not surprising considering the dynamic system that it’s trying to measure.

The most worrisome thing is that both ithlete Pro and Elite HRV prescribed false positives (telling me to train, or worse, train intensely, when I needed to rest).

Ithlete Pro was the worst; it regularly gave me false positives. It seems that something in the ithlete algorithm sees increased HRV activity as a good thing, always. Gradual increases are good, but sudden sharp increases indicate more of a severe stress response than a strong adaptation.

Now, I only use a daily orthostatic heart rate test and six defensive questions about my recovery (sleep, fatigue, stress, illness, etc.)

@cramblda: I think any AeT or AnT prediction is like throwing darts. It may seem to work for the population at large, but getting an accurate estimate for an individual is pure luck.

It’s well worth the time to get a metabolic test done. Or at the very least, use nasal breathing to define AeT. If you need to know AnT, a solo 30′ time trial is probably pretty close.

Similar to formulaic threshold estimations, the recovery recommendations are generic and unlikely to be truly accurate.

However, they do seem conservative–recommending longer-than-expected rather than shorter recoveries–so that’s a good thing.