Maybe an AI bug?

5 weeks ago I had surgery on a herniated disc, of course I’m itching to get back on the bike, there has been zero riding but alot of walking.

I wanted to see what the system would give me, so I manually dropped an AI ‘SweetSpot’ onto my calendar - it switched it to a VO2Max….

I dropped this:

I was under the impression, if you drop a ‘zone’ workout, the system will keep the zone and assign something within that? In this case it switched it to VO2Max i.e. out of zone. I’m assuming that isn’t expected? (or is it?)

It did this:

I’ve seen it switch workouts between sweetspot and threshold but never sweetspot and VO2max.

If you click on the card at the top of the workouts where it says “VO2” you should be presented with a reason for the change.

Ah yes, it shows this:

given though I’ve had almost 3 months off due to my surgery - I’m surprised it thinks a good way back, ramping up - slowing is to hit the very first workout with a VO2

Interesting. You’re in the midst of a plan even though you’ve not ridden in 5 weeks, right? IDK, but I’m guessing that based on the plan objective, it figures a sweet spot workout would not conform to the objective and you need a VO2max workout.

It seems roughly related to TR AI reducing my FTP prediction whenever I do anything other than the planned workout because what’s in the plan is optimum and anything else, harder or easier, is less so.

yeah, defo seems like a feature rather than a bug.

Much more simple. no plan. Just dropped on an AI zone workout ‘sweetspot’.

In that case it makes no sense. How can a workout not fit your FTP and goals if you have no goals? Goals are only defined in the context of a plan, right?

You’re right it makes no sense. There is an FTP set. AI FTP set a low one at the last update due to recognising the lack of training.

The system should not be giving someone a VO2Max first ride back after 3 months off. Second if you set a SweetSpot AI zone workout it certainly should not be upgrading it to a VO2Max. The behaviour of selecting a specific zone then it picking a workout within that zone was described as a feature so I don’t expect it to decide to ditch SweetSpot and prescribe something out of zone.

I never knew that TR could even do that but base on the data is has so far I agree it should at the minimum get some workouts under your belt before throw a VO2 workout at you.
I guess this is TR’s way of getting data without giving you a type of FTP test.
I wonder if TrainNow would do something similar. :thinking:

I agree that this isn’t a bug.

I do think it would be nice if we were to ease you back into training after time off (especially after an injury), so I’ll pass that along to the team.

I’m curious, @NotSoTurboTippy, what training phase are you in?

It might have made sense to edit your plan after 5 weeks off to get you back into some base training. That way you wouldn’t have come back to really hard workouts like this one..

Hi @eddie in theory I would be starting a base phase (if I had setup a plan). I was just interested to see what dropping on an AI selected Sweet Spot would select, I didn’t expect it to change it to a VO2Max, thats counter to the idea of selecting a sweet spot and expecting the system to select one.

Its been ~3months off since around Christmas. i’m being very cautious because of my back surgery, so I need to ease in and build up.

Ahh. I didn’t know that’s what happened here..

Let me look into this just to make sure things are working properly.. :face_with_monocle:

Do you mind if I share what I find with your account @NotSoTurboTippy?

1 Like

No I don’t mind, go ahead might help others too.

I believe this is an issue with your training data. It looks like you haven’t done a bike ride since late December, and we switched your SS workout to a VO2 one because we couldn’t find a SS workout that we had confidence you could pass.

I’d use AI FTP Detection when you log in next time. Once we give you a new FTP, we should be able to get you the right workouts. :+1:

I know that this wasn’t at all intuitive, so I’ve reported this to the team as something we can improve upon. Even some copy explaining what’s going on in these situations would be helpful.

2 Likes