Yes, this is per training zone.
FTP Prediction is not in early access yet.
Machine learning & adaptive training is all so exciting.
Presumably we are moving towards being able to not only input an A event into our planner but also a target āftpā which, once the ML checks our planned training time etc tell us how achievable this is or if we need more training time or maybe we are 20 years too late ![]()
I am so excited to see where the TR team take this. Whatever happens I feel in safe hands knowing itās a commitment to getting faster rather than a training methodology driving this ![]()
Thatās not a good metric because of the workouts that have extended endurance at the end will change it.
Also as I said on the podcast, TSS (and thus IF) doesnāt match difficulty of a workout. 8x5 min at threshold would have the same IF as 1x40 min at threshold.
IF is just not granular enough for structured workouts.
Weāre still improving that model (and always will be). Right now itās not available for beta users and Iām not sure if weāve defined a metric for when we should release it to everyone.
Nate and crew working OT on the weekend. ![]()
Thanks for all the follow-up and info. ![]()
This will be released to everyone before AT is live for everyone (if that makes sense).
The team is doing stuff now to make this possible.
Nate, that makes sense and can understand that you donāt want to reveal too much of your IP.
Really appreciate you taking the time to respond to my post. Made my day
.
@Nate_Pearson lot of questions on roll out for AT. Obviously appreciate you will roll out in stages and fix any bugs etc⦠But based on the beta list are people going to be waiting days, weeks, months,years?!
Like I say itās a genuine question, would just be good to know your roadmap for rolling it out as Iām looking to start a new plan but if itās days Iāll hold off. If itās months Iāll start.
If youāre truly changing the purpose of the workout, then yes, it will mark it as a struggle.
If youāre taking a little extra time during a long recovery you should be fine.
But as I said above, your response will help train ML in the future. But your response will be human reviewed to make sure we agree. IE if you stop for a minute, every minute, in a 20 minute threshold set but say you āpassedā, I think a TR employee with not agree with that and it wouldnāt be included in the training set.
Correct.
We need survey RPE data before we pop in these dynamic recovery weeks. Right now the model doesnāt know why someone chose to do different workouts on different days or why they stopped early.
With the survey results, we plan on doing what you have suggested above.
Sorry if Iāve missed this, but I usually train Monday Wednesday Friday, but quite often like this week, I left my Friday workout until Saturday. I assume the āskippedā Friday workout wouldnāt cause an issue once completed on the following day or two afterwards?
Iād suggest that you actually modify your calendar by dragging the planned WO from Fri to Sat. Makes sense to point AT in the direction that you arenāt āskippingā but actually ātweakingā your schedule intentionally.
So my predicted FTP is 400 confirmed by @Nate_Pearson ![]()
Big kudos and respect to @Nate_Pearson and the entire TR team for their effort on this launch and engagement on the forum.
uh @Nate_Pearson do you actually understand how cool this all is?
Itās very cool.
What would be AMAZING is if you could set the AI to āget me to 300w ftp in the next 3 monthsā and voila! Training plan ![]()
Judging by the smile and enthusiasm seen the the cast⦠I think he (and the entire TR crew) have a solid idea of Kewlness Factor. ![]()
I bet specific goal inputs will be a great/logical next step in the months to come. Things like a specifc FTP increase, developing a specific energy system, etc.
