Watchletic, DFA a1 for the Apple Watch and Wear OS

Updated 9/11/26 - eval of version 3.2 beta

Every so often a new sports monitoring app catches me by surprise. As a Pixel Watch user, I have a weekly ChatGPT search scheduled for any developments in HRV or Bluetooth HRM for Wear OS. A few weeks ago I was pleasantly surprised by a chat notice that a "new-ish" app now supported Bluetooth HRV and DFA a1 for Wear OS (and the Apple watch). 

 

 

 

The app is called Watchletic, and in this post we will take a quick peek at its ability to track DFA a1 during a sample indoor trainer session. Since the app is still in development and the DFA a1 implementation is still being fine tuned, no formal stats will be presented. However, in a future post, a more formal evaluation will be done. As usual, this is N=1 so YMMV.

The app was installed on my PW5 watch with no issues and pairing with a Polar H10 was quick. The screen interface is also straightforward to use to create custom fields on your watch:

 

Here is a screenshot of the entire session from my Fold 8:

Here is a different session showing just power and DFA a1:

 

The lines are color coded with stats displayed where you put the cursor (see the arrow above) The recorded file is downloadable as a .fit file. That .fit file can be uploaded or analyzed as well, here is the file data as interpreted by Runalyze:

 

• It should be noted that the Runalyze file DFA a1 values/plot is derived from the RR output of the H10 (from the .fit file). I used that same .fit file for Kubios vs Watchletic a1 for the comparison.

The following 2 hour session was comprised of a warm-up, a submax ramp (up to my first threshold), and 5 sets of blood flow restriction (2 min on - 2 min off with compression at 75% AOP).

Since we have Runalyze on board, looking at the O2 sat drops through the BFR is always fun:

 

(Yeah, the first set had minimal O₂ desaturation.) 

How did DFA a1 fare between the RRs analyzed with Kubios vs Watchletic?

Here is the entire 2 hours. The max artifact was 1% or less, so we don't have to be too concerned about the "better" Kubios automatic correction method except in one spot, as we will discuss below.

 

A couple of comments before we zoom in to the ramp and BFR intervals:

• The output data is now in a CSV, eliminating previously present timing issues. Many thanks to the developer for doing this.
• The app does use Kubios detrending methodology and threshold artifact correction. My default for Kubios is "automatic" correction, which is more advanced. This is important as the alignment between Kubios and Watchletic deviated in the ramp.

Zooming in to the ramp/intervals:

 

Artifact correction threshold enabled in Kubios

 

• The green circle shows the deviation. Why does this happen? It turns out to be purely from different artifact correction methods. If I switch Kubios over to "threshold" methodology, that transient a1 rise shows up, just as in Watchletic.This is important to remember - if you have many artifacts, most software will have a DFA a1 bias upward.

Blood flow restriction sets:

 

• This is the most encouraging image of the bunch. Why? It shows virtually identical dynamic range of a1 from high to low on the BFR intervals. It's been my experience that apps that don't do well vs Kubios have little issue in the 0.8 to 1.5 range but can't reach the low nadirs of the anticorrelated ranges below 0.5.

What else can the app do, HRV wise?

It also has the option of reporting RMSSD for those interested in resting HRV. It's still using 2 minute windows but can display the running average:

 

• The RMSSD (and artifact %) as a data field will show on the watch display, making it useful for resting and recovery usage:

 

• The "Next" circle is a button press for easy interval demarcation and interval timer reset

How are the comparison stats?

The following stats are from 2 sessions: the Ramp/BFR/5 minutes in high zone 2 and another session in zone 1, with a 5 minute mid-zone 2 interval.

Ramp/BFR (2 hrs):

Regression:

 

Bland Altman with minimal proportional bias (at high a1):

 

Zone 1 session with a 5 min zone 2 (60 min):

Regression:

 

Bland Altman:

 

• Both sets of raw data are not normally distributed, so the regression is "unofficial" (for you "real" science people out there), but I think you can see that the correlation is very good to excellent.
• The Bland Altman plots also show good agreement with a tiny bit of bias, possibly due to the artifact correction issues mentioned above. Both sessions were similar in this regard. 

Summary:

• Apple Watch and Wear OS users now have an option for DFA a1 display and RR recording  while exercising. This is a very reasonable option to both display DFA a1 in real time, record RRs for later upload and show the session graphics on your phone afterward.
• For those interested in RMSSD, artifact and cycling power that is also available in the smartphone and watch display. 
• So far the "agreement" with Kubios HRV is excellent and at least in my case, similar to Fatmaxxer.
• Bottom line - this is an exciting development in mobile HRV and DFA a1. Use cases include times when you can't take your phone but have a Wear OS or Apple Watch (trail run, distance run). Additionally, if one wanted to monitor a1 and RMSSD (and artifacts) inexpensively, get this app and a used Pixel watch 3-4 or used Apple Watch, and you are good to go!

All tested DFA a1 apps

 ....Blog Index....