We seem to be in the “I can do that too! Let’s make a company/service/app!” phase of this so more and more keep springing up. Who knows how many will last ultimately.
Can’t decide if this is a super niche target market “I want all the science”, or actually just a big research project. Collect lots of HRV data from current-level consumer devices and see if ML can figure out any useful patterns?
1 Like
AI Endurance has been doing HRV stuff for a bit already, though I know of noone using it in general.
^^THIS^^
There have been many companies that have come and gone with adaptive training / AI training offerings. Xert’s been around for 7 years and isn’t going away which says a lot. It does far more than just prescribe training and does it in completely novel ways.
Indeed, we have some new things coming, some of which are related to training and others that are related to data analysis, all very innovative and disruptive to the entire space. 
Stay tuned.
5 Likes
How many year round subscriptions do you have?
We don’t divulge our subscription numbers but we successfully bootstrapped the business within 2 years and the majority of our users opt for a yearly subscription over the monthly. Our growth is almost entirely word of mouth.
2 Likes
@oldandfast I’m very interested in your experience with this app once you’ve had a chance to try it out for the two weeks. I read a little bit of the blog and I think it’s pretty interesting to see what their training plans are like compared to TR.
1 Like
I’m glad entrepreneurs are staying alive, but your lack of investment in a modern interface and the fact that you are now behind after having 3-4 years lead in modeling, does not make me optimistic. Good Luck and hoping the next big iteration delivers.
I am a little confused by your statement or simply ignorant. What do you mean a modern interface?
I have used Xert and found it to be intuitive and offer literally endless possibilities of self-coaching. You just need to know how to use the system itself which I think they are trying to resolve. That does not take away from the UI that it has now which is pretty straightforward.
The program/system is about as informative and predictive as you can get right now and does give you the tools to make a decision (rather its right or wrong).
Are you maybe referring to how EBC has to be used to power any indoor workout?
7 Likes
Dear @oldandfast,
Thanks for the kind comments and review.
WRT DFA alpha 1 HRV, we’ve pressed pause on this integration. We went down that road alongside a number of researchers in the field and learnt some big picture things that told us to stop. Believe me, we want VT1 as much as anyone, but this method doesn’t lead to accurate/reliable results across the board as we need them to. More coming here but this is on the back dev list for the time being.
Many have commented in the thread on the blog posts (thank you!). The two that weren’t mentioned that we feel quite proud of due to their real world utility were our Workout Wizard and our Workout Reserve. Technically these were very challenging to realize but again, extremely useful in your day-to-day as athlete/coach.
Finally, we are running a free course, The Science and Application of Endurance Training using AI Platforms, on https://hiitscience.com/, for anyone that is interested in the underlying function of Athletica. This is predominantly for the coaches out there but hobbyists will also find of interest.
Please reach out if I can answer any questions.
All the best,
Paul
7 Likes
Thanks for dropping in!. A few questions:
-
If no DF1a is used, what are you measuring with the initial ramp test?….given that you also have the FTP test and the rider profile test.
-
I had some glitches where I populated a training plan, added my fatigue numbers, selected to opt out of testing……the training plan still included the testing week.
-
It not clear if any adaptation is happening when I added my information. For instance the system recommended 2h aerobic, I added the data of my usual 6h ride. Aside from saying “above”, I don’t know how this affected the following workouts.
-
Just a general comment about interface. I know a developer(s) is expensive, but it’s a bummer when a good product doesn’t get traction bcs of the clunkiness of the front end……check the good people at Join, they have a clean app, the plans and adaptations are ok (but opaque), but I bet just the “feel” of the app gets subscribers. Otherwise you stay in a negative feedback loop of no growth, no revenue to improve app.
-
Are you able to articulate why the Ai is making the adaptations it does?. Or is a black box at this point?
Thanks for the app and the info you put out there.
Old
1 Like
Hi @oldandfast ,
- You found a glitch (thank you!). As mentioned, we were doing DF1a with the data but no dice. I will remove from template now. I missed that one.
- Our default for entry into the plan is to request that you do the test week so we can be confident around your zones and our prescriptions. Thereafter, we allow the opt out to function.
- The above data would down-regulate subsequently planned training in line with what you’d expect. We’re using ratios of ATL/CTL and a few other common sense calculations. As you might expect, the longer you’re on the plan, the better it knows you.
- Yep - we know. I’ll take full responsibility. I’m a scientist and I was never able to ‘sell’ my vision to the right investor to allow us to bring in the front-end person we needed. Now that we have a new CEO with this skillset, we’re in a financial environment that’s very challenging. Stuck between a rock and hard place. Please let me know if you have a solution you could suggest for me.
- We disclose just about everything in my new course and ebook on the topic. Its 100% free so feel free to go through. Endurance Secret
All the best,
Paul
3 Likes
That sounds overly and unnecessarily complicated. And why is it so important to pinpoint VT1 with such precision? On easy endurance rides, it is super easy to stay below VT1.
@plaursen thanks for the Athletica blog. Are you able to provide more details on the OmPD model, and how it fits the curve to predict the boundary between moderate and heavy domains? I’m the guy that posted a hacked curve above in the thread: Another AI training app: Athletica.ai - #17 by WindWarrior
By the way, decided to buy the textbook after seeing these HR zones
match my perception/data tuned HR zones. Hey I know its a bit of luck rather than some cosmic message, but go figure, and now the book is off my Amazon wish list.
Thanks for all you do.
Cheers!
Hi @WindWarrior,
Pleasure and thanks for the kind comments also. I hope you enjoy the book 
To your question, we follow Puchowicz et al. (2020) pretty closely in terms of our methodology. This probably isn’t the right place to cut and paste the methods section. Do you have access to the paper?
Puchowicz, M.J., Baker, J. and Clarke, D.C. (2020) “Development and field validation of an omni-domain power-duration model,” Journal of Sports Sciences, 38(7), pp. 801–813. Available at: https://doi.org/10.1080/02640414.2020.1735609.
2 Likes
thanks, if possible I’d appreciate a copy of the paper. If you are ok with that, I’m sending my email via private message on the forum.
1 Like
Honestly, given the current environment and your competition, if you can afford it, go the intervals.icu route. Make your app free, create a big community in a forum and they will give you their data, find the bugs, tell you the features they want and spread the message….some of them will donate regularly once the app is fully functional. Then you can start changing the pricing structure slowly, partial features…etc.
All this needs a very good programmer that is also a partner and it’s a cyclist/endurance aficionado, and this person needs to be convinced that even if there’s no pot of gold at the end of the rainbow, the effort is worthwhile.
1 Like
Hi,
is there a way to have my rowing workouts counted into your workouts and training plans, or is it just Cyc/Run/Swim?
Hi @rentagreement. All workouts you do, including rowing, are counted in some way towards your overall training load and subsequent prescriptions.
1 Like