🎉 🎉 🎉 Introducing Adaptive Training! 🎉 🎉 🎉

Can’t say I was surprised or disappointed :rofl:

Seriously though, this context is quite refreshing and eye opening. It gives me some optimism that I may be able to simply use PB to set a plan and trust AT to set me straight without trying to break me in half.

Honestly, I still have some reservations with re-adopting the standard work / recovery schedule vs my altered versions that have worked really well for me over the last 1.5 years or so. I know those more frequent recovery weeks are what allowed me to suffer through and, more often than not, get relative successes on those SUPER HARD workouts. But I also know those were almost always a stretch for me, even in my younger years.

All that to say that this is I like what I see right off the bat. Hopefully this will be positive in the eyes of other masters users and anyone who struggled with the “default” TR plan layouts and workout prescriptions.

Just to add to this, the specific workouts will change, but the rhythm of your plan won’t. A recovery day won’t be replaced with a hard workout, a day off won’t suddenly become a day on, and workout lengths will only occasionally ever change, and only very slightly when they do. So it’s really only the specifics of the work itself that changes as your plan adapts.

I would be interested in this as well. Perhaps a custom “plan” put into the calendar could be adapted? I realize this is probably considerably more work as the system won’t necessarily know the goal of my custom plan and maybe this isn’t really feasible, but I’d like to at least throw out there that there’s some interest.

I usually structure my base progression with a philosophy somewhat matching that found in the “Sweet Spot” thread, i.e. increasingly long SST intervals out to at least 2x25/2x30 before moving on to build. I’ll probably use the built-in LV SPB plan. And then finish with a custom specialty plan that pushes out SST duration well past an hour. The sweet spot durations in the built in plans feel insufficient for the long distance events I target (but I could be wrong! Maybe I’m being a "special snowflake! lol).

In any case, don’t take this as complaining. I love the product y’all are putting together, the community here, and TR’s responsiveness to its users.

Follow up question: Could the planned energy system change as well or will AT just give me a easier/harder workout for the energy system that was originally scheduled for any given day?

Training Zone will not change. If it’s a VO2 Max day, you’ll get a more curated workout within that zone.

I guess I was asking how it would take into account these types of failed workouts due to whatever reason that have already occurred? In my case the ramp test wasn’t accurate due to losing connection.
I seem to remember something from the podcast how the AI would be looking at all past workouts for determining progression levels. But these past workouts would not have any info via surveys on them.
Ty

The system is just looking generally at your past workouts to get a good starting assessment of your Progression Levels on which to build. Don’t stress over device issues or any other struggles in your workout history. By design, AT learns and reacts very quickly to your performance and current capabilities.

K. Thanks for the quick response. Since it was a “ramp test” that failed due to device failure I was concerned it might put more weight on it then just another workout.

This is probably a point worth repeating or making dominant in communications to existing users.

Whilst I took to structured workouts fairly quickly, structured weeks and months took much longer. Plan Builder made me feel a bit uncertain in planning my year, Adaptive training did also, and I’m sure I won’t be the only one.

Now I understand it though, I cant wait for it :slight_smile:

We’re going to score outside unstructured rides and assign level(s) to them. This will move you in your progression.

Coach Chad hugs.

There is not just “one” model. There are many parts that chain together.

The ML team is always either adding a new model that chains in, or adding new features for existing models to see if it increases accuracy.

  1. Sort of. You could go back in the calendar and see the workout levels you were doing. We’ve been discussing historical views so that you can mouse over your TSS calendar and see your highest level in each training zone per week.
  2. We were just talking about this! I originally had a bad idea to make everyone level “4” across the board when they manually change FTP. We saw some bad rides come in on the beta after that so we’ve now made it like if you did the ramp test. If you manually change your FTP you will have a more progressive step in all of your levels.
  3. Yes, A/B/C is just like it is today.

There’s been a lot of people wanting us to get people in faster to the beta. Here are our current issues that we found in this round that we’re fixing.

That means this work is almost done. “PR” means “Pull Request”. In the above image that means Alden has submitted code to fix the “Check for adaptations after reapply”.

Want us to go faster? Have your friends apply at Trainer Road, LLC jobs | Trainer Road, LLC openings | Trainer Road, LLC careers.

Our COO also put more info about the beta process here: Adaptive Training Closed Beta Update - #4

:smiling_face_with_three_hearts:

@Nate_Pearson A few of us asked about this about a million posts up in this thread and I think it got lost. Are there any medium term plans to add prospective data into the featureset such as ‘on a diet’? I could see some of your menstrual cycle data being something like this as well. It is one thing to tell the model a workout went badly because you were 750cal per day restricted for 2 weeks (or a woman on her cycle at x point, or pregnant, or in a house that just had a new baby) and it is another thing to tell the model that before a workout and let it make better recomendations. The tech related to this feels like its in the same wheelhouse of work you already have in place.

This is exactly right.

I like to think of ML like cooking.

Ingredients = data
Chef = ML Engineers

You need a great chef and great ingredients to produce something special. Just because someone else is doing “ML”, doesn’t mean they will come up with the same output.

Just like if I had great ingredients I wouldn’t come up with the same dish that Gordon Ramsey would.

I read an update today that this is production ready now. So most likely this will be in beta very very soon.

If you’re on a plan, it will suggest a TR workout.

But…you could find a workout at the same level in your WOC and do that.

A little bit. The thing we aren’t showing you in TrainNow is if it is Achievable or Progressive. That data comes out soon.

So you might get a little bit harder workout or a little bit easier workout.

Once that gets released with levels you can probably do a bunch of AT stuff manually in your plan.