There was another user above who said they’ve been testing privately with them since last year sometime, I believe. That said, I think we’ll be waiting a bit as well. Nate mentioned they are still tweaking onboarding in addition to the bugs mentioned.
3 Likes
Somewhat fortuitously, I had a 2:15 hour workout scheduled yesterday, almost exactly the length of the podcast. And yeah, this was a good one.
First of all, @ambermalika congrats! When you announced you were pregnant, I spiked +30 W and gave you an imaginary high-five. My wife and I are trying for our second right now, so I can imagine how joyous of an occasion that is for you. Also good to know that you gave birth to your TR baby before going on parental leave.
The podcast did a great job explaining what is going on, and it seems we got pretty much everything we wanted. Can’t wait to try it out!
9 Likes
Absolutely love that Avasarala drop in. chef’s kiss
3 Likes
I may have missed this somewhere in the 1000 posts above, and the podcast:
Will the AT algorithm provide any recommendations on changes to training intensity distribution (I.e. beyond just recommended workout intensity changes within a given plan)?
E.g. “based on your performance and reaction to past workouts, we recommend you do less threshold workouts and more endurance”. (I’m using this as an example, not intending to restart the POL vs SS debate).
Maybe a roadmap item if not in the current algorithm?
4 Likes
Until unstructured workouts are included, I had a thought for a workaround. For a three hour outdoor ride, say I chose Ochoco. At the end, I will have done something that may or may not look like Ochoco. What would that do?
Also, sorry if I missed it, but how would AT handle power meter variation? Meaning 250w on one power meter vs another.
And mountain biking.
I this will be awesome, and get continually awesomer.
1 Like
Number one, I’d like to buy a bit of your enthusiasm for cycling and training, I’m pretty motivated at the moment but that was great.
Number two, I agree that pulling in data from as many other sources as possible, even a ride done on a competitor platform, would be fantastic for the end user. I don’t know if that is good business sense in terms of new users and retention (I’d say it would be) but the functionality and opportunity for analysis would be awesome.
Hi
How will the maths deal with non planned outdoor rides (Sunday club run) will it just take extra load into account as it’s not measuring against a workout and most of us still do this type of ride once a week ?
Ps. I’m really excited by your development and use this technology in other fields
1 Like
I feel professionally obliged to point out that while this is fairly strict on the governance of how personal data of European citizens is used, it is not prescriptive on information security measures. Happy to discuss further in another thread.
@Shoco I think the AT will be super cool for people like us that just want to do the prescribed sessions that TR sets us because with machine learning it will automatically adjust the planned progressions if it sees us struggling/failing/delaying workouts.
Currently if I fail the V02 workout in week 2 I can expect to be in for a world of pain in week 3 unless I know I got my ride preparation/fueling wrong etc. AT should mean we always get sessions that make us not break us!
In fact the only input we will need in future is to say how we felt at the end of the ride. Indeed, the way Trainerroad is progressing they will keep amending our plan however is best to help achieve our goals, Ultimately I think that they might eventually start sending someone round to your house to give you a hug when you need it 
6 Likes
It won’t, yet. But they are working on it as far as I understand.
1 Like
My understanding would be that as part of your plan if you do have an event recorded that you’re aiming for, that the plan as guided by ML would get you to a point where you’re peaking for that event.
This 100%, for me is just knowing that if I do a cheeky ride outside that is harder than TR had planned for me, that the subsequent training workouts are adapted and adjusted.
3 Likes
That’s not quite right, as they say in the podcast the ML algorithm is trained constantly, based on the dataset that is all of us. But more importantly, the algorithm adjusts to us individually as we progress through the training program.
1 Like
The beauty of this release I would imagine is they don’t need a lot of beta testers. They have the data to train the model and apply that model to and see what comes out the other end. Is it what @chad would expect?
The beta testing is the icing on the cake, hopefully just confirming that athletes get the session they would expect from the ML.
It would be interesting know how often @Nate_Pearson expect the model to be updated. Are they always run two, the live one and the next one for comparison?
1 Like
I’m an idiot! Now I got it… the desktop APP!! I was logging into the website on my desktop. LOL!
1 Like
This is what could be a real game changer for me. I’m willing to try (risk wasting) a cycle on seeing how I respond to AT on a TR plan, but three days of intensity a week, even properly adjusted, concerns me. So I am really looking forward to trying it but am nervous.
However, once custom plans and workouts are taken into account…ooooo baby that’ll be nice
5 Likes
We’ve heard this one before. I love TR and the podcast, but from a triathlete’s perspective the development in the multisport direction has really stalled. Adaptive training is no exception. I imagine it’ll affect bike workouts on a tri plan (as long as it’s a Trainerroad plan), but the engine won’t know what should be changed or why.
We’re just not a priority and that’s fine. You know your market. Even for the glorious future developments, triathletes are just seen as a segway to serving the running community. We don’t have the weight.
I’ll continue to pay the $100/year because it’s still a decent deal for a trainer app, especially one with group workouts. Also, there’s no escaping the fact that people like me get sentimental about TR and its team 
4 Likes
AWESOME! AWESOME! AWESOME! Massively impressed by this upgrade. This is simply fantastic.
Just ideal for me as I am struggling with the odd injury. Love the “unstructured outside ride feature”, a I did 38 TT races and a shed load of training outside for long TT races.
Questions:
- Can I look back at previous position? I was really hot at the end of 2019, but am way off the pace this year. Would like to see where I was overall against the 7 zones & levels.
- What if I self change FTP? (eg after injury) Or do I not bother?
- Does it make allowances for A, B and C races.
As a data geek, this is Awesome! Well done TR team.
1 Like
+1 … I was thinking on the survey I’d mark device problem as a limiter and if if there was free-form explain fans couldn’t cool enough. But could there be an ENVIRONMENT option to cover too hot, cold, humidity, pollen count, or some such?
2 Likes
I’m so excited about this! For the past year, probably even longer, my training honestly has mostly been garbage. No consistency, lack of motivation and discipline, out right dissapointing. Even with FTP dropping, the workouts seemed to get harder and harder, just a drag to get on the bike. Even just getting on the bike which normally would be the biggest hurdle, isnt enough anymore. Once on the bike my thoughts directly go to when can I get off. Resulting in a negative headspace this just isn’t a healthy behavior. And planning around this seemed almost impossible.
To have a big part of these issues now taken care of is giving me hope again. Hope to enjoy training again, hope to get faster instead of slower again, hope to move forward again. No need to worry about failing workouts and therefor creating issues with the rest of the training plan, eliminating another big part of a viscious cycle. I don’t to fight through it, I want to work through it.
Can’t wait to try this out and feel like I ‘deserve’ to be a cyclist again.
3 Likes