I spent ages trying to get useful data form DFA. Multiple straps, apps and input from the authors. It was a mission.
The science just isnāt there right now for someone like DCR to burn so much time on this IMHO.
Usually LT1 is at 1mmol above your resting or lowest lactate which usually comes around 2mmol. LT2 usually comes when your lactates start to accumulate rapidly, usually you can feel LT2 pretty easily as that is when it starts to really hurt! This measurement for most is at 4mmol but I have seen athletes in the lab have 6-8mmol and still be at a steady state output. Training for the uphill athlete has a great protocol and some great advice on how to perform a lactate profile.
For what it is worth I have a M.S. in Exercise Science.
You must be new here then. 
Does the TrainerRoad app actually record RR intervals so you can calculate the DFA-a1 from the .fit file afterwards? Doesnāt seem to be the case, but maybe Iām missing something?
Would be great if this can get added to the recording thoughā¦
The TR app doesnāt, but there are half a dozen ways to get the data and analyze it at this point. Own a recent Garmin? RR intervals in the .fit. Own an old Polar V800? RR intervals in CSV. Use the free app Elite HRV? Guess whatā¦RR intervals in TXT. Want realtime dfa-a1? On Android thereās Fatmaxxer and IOS has its own app.
Want to analyze dfa-a1 for free? Sync with Runalyze. Or use the chart in Golden Cheetah. Or use colab with Mr. Altiniās script.
Want to really analyze dfa-a1? Buy Kubios Premium.
dfa-a1 is cool, but becomes very boring after you use if for a few weeks (assuming you already have a feeling for VT1).
Thanks for the info!
I will work with Garmin & Runalize. However it would be much easier if the RR just comes from the TR file, to avoid having to double record your TR ride - Not sure if this is on the roadmap? Sounds like a very simple implementation behind the curtains to just record this additional data stream in the .fit file. And might be a useful metric to be used in AT in the distant future. Any perspective here @IvyAudrain ?
Will a Garmin 530 capture RR intervals in the fit file if paired to a Polar H10?
It does. However make sure you 1. connect over Bluetooth instead of Ant+ to record RR as Ant+ seems to suffer from much more artifacts which can generate bad DFA-a1 values and 2. turn on RR logging via Menu > Settings > System > Data Recording on your Garmin.
OK, found this post today :-). Have an Garmin Dual HRM and just changed settings in my Edge 530 to record the HRV data.
Maybe later today I do a ride and then analyse it on runalyze. What is the extra benefit for using HRV logger for real-time info? But If you do some rides and analyze afterwards you already now your LT1 and you can focus on that value in your rides? Or am I missing sometime?
Not the expert here, so Iāll share my current thinking on the question.
Reasons to use Live tracking through HRV logger:
- Can be used instead of runalyze or other tools to do the testing, to have an immediate view on how the value evolves in function of power or HR (although you only get a value every 2 minutes, and it seems to sometimes overestimate the DFA-a1 value for some people)
- Once you have determined your LT1 through e.g. runalyze, it might be used to course-correct during a subsequent workout. E.g. it might show you that youāre not 100% after a bad night, or when getting sick or being overtrained. And then give you an indication to take it easy if you reach 0.75 DFA-a1 at lower power levels than expected.
I think itās best to do a structured ride to determine the LT1. E.g. 20 minute warm-up, 20-minute ramp from 55-60% FTP to 80-85% FTP, 20-minute ramp down again to 55-60% FTP. And then put the 40-minute ramps-block of data (so excl. warm-up) into runalyze and see what comes out in terms of LT1 power and HR.
Just a note, dfa-a1 of 0.75 is already above VT1 for the two athletes based on the paper.
I see why 0.75 was selected, but some folks are likely at VT1 by 0.85 or 0.8.
I do a simple 1-hour low intensity ride and then look at dfa-a1 after the fact. No need to have anything in realtime.
Iād recommend keep things simple. A single ramp structure workout with a longer step length (e.g. 60 minutes at 5-10 watt increments every 10 minutes). The trend in Runalyze will be easy to see.
Ok, have my ride, but it was not a structured test.
Do you need to alter some settings in Runalyzer? Or you keep the default values in the AEROBIC THRESHOLD ESTIMATION BASED ON HRV?
It was a Z2 ride with 3 times Z2-Z3 10 minutes. But maybe this is not good enough?
DFA-alpha1: avg. 1,12 (min. 0,53, max. 1,67), > 0,75: 98%, Valid: 90% Artifacts: 0,7%
Few values around 0,75 around 131bpm and 200W. Maybe I need a better ride.
And it calculates also the aerobic treshold? 243W at 143bpm but I never reached those values?
For me the difference between 0.75 & 0.8 is 3-5W, so in the end it doesnāt seem to matter that much. Donāt know how precise an actual LT/VT measurement is, but will probably also have a confidence interval of 5-10W at least.
The reason I do a slow ramp up and down is because I assumed there would be some lag on the response. Which proved to be true, so doing both averaged out the result. Also I read here that an actual ramp gives better results given DFA alpha1 has a Gaussian distribution around a mean at a constant power. So a constant ramp smooths this out vs. a ramp with steps.
Anyway, probably Iām already overanalysing this, given accuracy / precision of this method is still questionable 
In the same page I mentioned above, I read to decrease the artifact correction threshold to 0.01 and only select the actual (ramp) test as data points, so excluding the warm-up etc.
I think you donāt have a good dataset, and should do an actual test instead of just recording an endurance ride. You should probably ramp through your VT1 point, so you have values below and above 0.75 and there can be a nice interpolation. Now it seems that the model needs to extrapolate to get a value, which isnāt ideal.
I now tried 2 ramp tests, and will try a step test somewhere in the coming days. Even while I still doubt if thereās any training value in doing this, itās fun to read up on the topic and do some experimentation. 
If you have access to Android devices then just get Fatmaxxer for a couple of $. Pair it with your H9 or H10 and get a real time read out and see a graph of HR, A1 etc and see when it nears and crosses 0.75.
Worth doing the slow ramp described above with 4-6 min intervals to get the initial numbers.
Definitely hoping the development team can have the bandwidth to work on run (and swim) support soon!
Thx for the answer, but not sure if my question was understood. Not a running / swimming question. 
In summary: it would be great if TR also logs the Heart Rate Variability data coming from a HR monitor when doing a workout on the trainer. This data is already sent out by Polar or Garmin HR straps but not captured / logged by TR
OPE apologies, I got my forum threads crossed haha.
To speak to features like this getting on the roadmap, itās not that the team doesnāt these āeasierā features arenāt important or useful, we just have to prioritize updates that are the most effective at helping the greatest number of athletes possible get faster in the immediate future (such as incorporating unstructured rides into Adaptive Training, and refining Adaptive Training as a whole). 
Awakening this thread to throw a question to the group. I drew up a DFA a1 ramp test today to play around with it using HRVLogger. The genesis of my desire to try it is post-COVID training my generally reliable HR trends arenāt aligning with RPE and VT as compared to power. So, I was looking for another way to try and confirm where I was feeling LT1 even though HR at that point is higher than normal right now (and probably temporarily so.)
The test (FWIW) confirmed that LT1 is aligned with VT1 and RPE and HR is out of whack somewhat, but Iām curious how any of you would interpret the data. From the DFA a1 graph below, you can see a step down to less than 0.75, which is thrown out as the first data point of that step. The next two are 0.75 and 0.74, and then it dances around 0.75 for another full step before falling off the cliff. So strictly using this data, I could put LT1 anywhere from 200W to 216W for those stepsā¦
Instinct tells me to use the first step where it dropped from around 1 to 0.75, ride just below 200W and call it a day, but curious what other smart folks think. Thoughts?
What I try and do after doing the ramp test and establishing where I think (of hope) my LT1 is, is to pedal really softly for 5 or so minutes to let the body settle down then do a constant power 30 mins or so about 10 watts below that figure and observe what happens to the alpha1. It looks like you did something similar without the rest period. I think its something to do with once youāve gone over LT1 it takes a bit of time for it to come down properly. Probably why doing surges on LSD rides can blunt the effect of them. Have you tried use Runalyze. They do an excellent post ride analysis of the fit file. Lots of ways though of interpreting the data. RUNALYZE - Data analysis for athletes
This was a 5 min step test with 30 sec rests between each step. My interpretation is that my LT1 is between 180 & 190W. Being conservative I would say 180W. The LT2 point is more difficult 210W seems to be the point but 220W gives a similar result - so split the difference 215W maybe
