Missing 'Reload' button in TrainNow

[mirlord]

Besides saying +1 to all the commenters above, I have one thing I want to add to the discussion, and I do hope TR staff will see this and take it into account.

I have a “yellow” day today, and Train Now suggested a 6.1 VO2max workout. Very funny. When I see training recommendations like these given to me by the training platform, I have serious doubts about whether I can TRUST the rest of my training. This particular recommendation is outrageous enough that it requires no competence whatsoever to sense that something is amiss. But how accurate are all those plans and “adaptations” then? Should I, going forward, question every single workout that’s been given to me, guessing if it’s just another silly mistake made by AI?

Building trust takes a lot of time. But it’s fragile, and it doesn’t take much to ruin it.

[NotSoTurboTippy]

Yup. I’m seeing a VO2Max PL 5.1 when I’m on PL 1.0 …
I think the AI is hallucinating or at least TrainNow is quite broken at the moment.

[JoeX]

It’s making perfectly reasonable suggestions in TrainNow for me

[Rizzi]

@NotSoTurboTippy

Yup. I’m seeing a VO2Max PL 5.1 when I’m on PL 1.0 …
I think the AI is hallucinating or at least TrainNow is quite broken at the moment.

@Rizzi

@AldridgePrior

8.5 VO2max suggested for me by TrainNow, vs my current 2.0 PL (been doing just Threshold for a while…)

@UWDawgs

I’m a 1.5 VO2 Level and TrainNow just suggested an 8.2 VO2. Again I declined

@dougpdx

but the vast majority of recommended workouts now appear to be “breakthrough” leve

I’m not sure but I’m hoping that the reason for this might be that under the hood all your rides (outside / non workout/ whatever) are taken into account (as mentioned before in the forum and maybe on the podcast) and from these you have some higher internal progression level (WLv2) that is now finally taken into account for working planning before hopefully making the whole WLv2 thing public.

@ZackeryWeimer Could anyone from support check these TrainNow suggestions and if my assumptions are a valid explanation or if there is anything else or a bug…?

[AldridgePrior]

@Rizzi

I’m not sure but I’m hoping that the reason for this might be that under the hood all your rides (outside / non workout/ whatever) are taken into account … and from these you have some higher internal progression level

I’m all for a bit of optimism, but very much doubt the above! I’ve not done any outside rides for a while, and there’s no recent history to justify the “ambitious” suggestions, which I’d not come close to being able to complete…

I think the machine just needs to sober up and get a grip

[NotSoTurboTippy]

Hehe, maybe its had the AI equivalent of too much Christmas Punch or Gluewein

[eddiegrinwald]

There are a few things I need to clear up here.

1. TrainNow’s goal is to intelligently recommend you the right workout at any given moment. If you’re able to simply shuffle through the workouts and then use alternates, where does the workout you ended up selecting fall in the list of “best” workouts for you at that moment in time? This is something we’re thinking about with the new updates. I’m not sure where we’ll end up, to be honest, but just know that I’m sure we’ll settle on something that works well for you all.
2. We understand that you might be surprised by the level of the workouts that you’re being recommended, but we’re aware of this and each of these recommendations is expected. We think that it’s actually really likely that you’ll be able to pass those workouts. Trust us. We ask that you give them a go! If you need to for the time being, try to ignore the workout level of the workouts you’re being recommended…
3. There are actually a few small bugs going on currently and the current workaround while we get things sorted out is to run TrainNow on the web to select and schedule your workout, and then open up the app to complete it. This ensures that you’ll get a good recommendation for the time being. We know this is an extra step, but we’re working on getting this sorted out as we speak.

Thanks for the feedback y’all! I’ll make sure to forward this all to the team.

[mcneese.chad]

Good that it’s working for some people.

But I’m getting a rest day recommendation again today despite last Friday being my previous workout.

[mcneese.chad]

That’s great and all to look at improvements, but why was this not handled via beta and/or early access vs what appears to be universal rollout of a buggy and underbaked change?

One of the most simple and reliable tools was changed without our choice or notification AND includes bugs… that’s frustrating.

I deliberately skipped the new calendar early access because I didn’t want the hassle of dealing with bugs. I get that TN may be a smaller tool used by fewer people, but the way this is rolling out is less than the usual experience we see from TR.

[Nate_Pearson]

There’s a caching bug that our engineers are fixing now; that’s why you got the rest day.

We’re slowly rolling out a HUGE change to how workouts are selected on TrainerRoad. It’s a new machine-learning model that’s been in development for two years.

In our internal testing, it’s WAY better at predicting what a good workout is for an athlete.

“Good” is a combination of workout performance, failure rates, and long-term FTP improvements. IE, if you do a VO2 max workout and it feels too easy, or you fail; then that’s not a good workout.

When we fully launch this, we’ll explain these metrics and show you how we’re measuring them.

We turned this new ML model on for TrainNow 7 days ago. We did this because TrainNow is a very small portion of TrainerRoad’s total rides (<2%). The machine learning model is also pretty resource-intensive, so we wanted to ensure we did a small test before slowly rolling it out to a larger audience.

Just this small test costs us $5k/month. To ensure we don’t break the bank, we implemented some caching to reduce costs. That caching introduced the bug where you got your rest day, and some workouts weren’t the correct workouts to recommend to people.

Because of that bug, we turned off the new ML model this morning, and the regular TrainNow is running. An engineer has fixed the cache issue, and their code is being tested. We hope to turn on the new machine-learning model later today.

For the shuffle button: The new machine learning model picks one “very best” workout based on workout zone, profile, and time. If we had a shuffle button, it would just continue to pick the best workout for you and you’d get the same one over and over.

So I decided two weeks ago to turn the shuffle button off, not thinking people would care. But of course the person who has the biggest hawk eyes on TrainerRoad noticed :-D.

We will bring the shuffle button back, but have it sort through workout profiles for that zone. It will sort through 1-4 workouts, depending on the zone. That adds a little strain to the system, so it might not come back today, but I think we can figure something out for next week.

It’s a bummer that we didn’t get good data for these 7 days. Still, the ML model with the cache issue performed equal or better than the previous version of TrainNow; so that’s good.

But in our internal testing, it blew performance on Adaptive Training, TrainNow, and athlete-selected workouts away. That’s why we were confident to turn this improvement on for TrainNow athletes, but of course, it was a bug and not the feature that bit us :(.

When it’s back on, you might see bigger jumps in levels than you’re used to. That’s one of the cool features of the model. We’re seeing people “lock in” to the correct workout level way faster than Adaptive Training.

It also handles time off better than Adaptive Training. When you have gaps in training it’s better at predicting the correct workout level.

It might even suggest different workouts for you based on the time of day you make the prediction (which is cool).

There’s way more applications for this that we’re working on and we weren’t planning on sharing anything yet but no one will probably read this anyways besides Chad a few others, so .

The next rollout will be on the desktop app in the coming weeks, and you’ll have more control over whether you’d like to do the workout or not.

[MeMeToo]

When it’s turned back on will it also take “custom” workouts into account when adjusting levels to this precise detail?

[stevious]

Can I ask about my own very specific use of TrainNow and whether it’ll be possible with the new system?

In the summer I tend not to follow a plan but I do sometimes decide ‘Today I feel like INTERVALS’. Once I’ve figured out which terrain I can get to in the time I have I fire up TrainNow and refresh it until I get something I can fit onto the hills I can reach.

So I’m trying to optimise for terrain first and then let TrainNow optimise for enjoyment by giving me something at the right level of challenge. It saves me thinking about PLs whatnot.

Will Super TrainNow Turbo Special Edition let me do this?

Thanks!

[Rizzi]

@Nate_Pearson

We’re slowly rolling out a HUGE change to how workouts are selected on TrainerRoad. It’s a new machine-learning model that’s been in development for two years.

Bring it on!

Personally I’d like to use this as kind of dynamic day by day (or one week ahead) plan (maybe with a configuration like “build my FTP”) instead of using Plan Builder and getting served a long term plan. (Maybe think of Join or Xert)

Ideally the new model will also consider training fatigue of all other sports (triathlon), but at least there is RLGL as a fail safe.

[Jolyzara]

@Nate_Pearson

We will bring the shuffle button back, but have it sort through workout profiles for that zone. It will sort through 1-4 workouts, depending on the zone. That adds a little strain to the system, so it might not come back today, but I think we can figure something out for next week.

Not sure what others are seeing but my “train now” link is doing this for me currently. One workout option with a shuffle that rolls through 5-6 similar workouts. (In my case is did a ride already today so it gives me nothing bit recovery options or a rest day). Also to note it does offer the option to change duration of the workout.

One note though, and I sent a message to support, if you click on a rest day there is now way to remove that from your calendar. In my case I already had done a ride 8n the morning. I clicked rest day in train now to see what would happen and it populated by calendar. But it’s stuck there despite having activity previously and no way to delete.

On another note I think this is an awesome tool. The idea of picking the best workout for today is really interesting. Like the traditional train now does it only look at today or does it look to the future such as upcoming races? If you are on a plan does it consider that tomorrow might be a hard workout? Or is it only looking at right now?

[JoeX]

@Nate_Pearson

but no one will probably read this anyways besides Chad a few others, so

Your ‘secret’ is safe with us, and the internet.

[Nate_Pearson]

@MeMeToo

When it’s turned back on will it also take “custom” workouts into account when adjusting levels to this precise detail?

It might; but the list of workout that it would recommend for you would be smaller than the list we have showing today.

[Nate_Pearson]

@Rizzi

Personally I’d like to use this as kind of dynamic day by day (or one week ahead) plan (maybe with a configuration like “build my FTP”) instead of using Plan Builder and getting served a long term plan. (Maybe think of Join or Xert)

That’s where we’re going.

The future workouts change anyways, so we’re going to hide them and just show the the workout profile (like VO2 Max on/offs) rather than a specific workout.

Then once you finish your workout before that we’ll run the ML, and reveal the next workout.

[Nate_Pearson]

@Jolyzara

One note though, and I sent a message to support, if you click on a rest day there is now way to remove that from your calendar. In my case I already had done a ride 8n the morning. I clicked rest day in train now to see what would happen and it populated by calendar. But it’s stuck there despite having activity previously and no way to delete.

On another note I think this is an awesome tool. The idea of picking the best workout for today is really interesting. Like the traditional train now does it only look at today or does it look to the future such as upcoming races? If you are on a plan does it consider that tomorrow might be a hard workout? Or is it only looking at right now?

Thanks for the bug reporting; well get that fixed.

Traditional train now looks backwards at what you’ve done recently, but not forward. It’s mainly used by people who don’t have plans on their calendar.

[mirlord]

@Nate_Pearson

The future workouts change anyways, so we’re going to hide them and just show the the workout profile (like VO2 Max on/offs) rather than a specific workout.

NIIIICE!!! That just makes soooo much sense. Thanks for going in that direction, seriously.

[paulhilliar]

I do like to be able to look ahead and see the suffering that the plan has in store. And often do workouts out of order.
Hopefully you meant that you don’t get to see the specific workout for that energy zone until you completed the previous workout for that energy zone.