How many predicted watts are you losing from missing one workout?

Hi All,

Love the program and the workouts. I’m here just chasing a stronger FTP. Naturally I’ve been trying to optimize the plan and I’ll mess with the training approach and other small things.

My last detection was at 276 watts on 2/18. The next prediction on a balanced approach (which is the highest prediction for me) will be 294 on 3/18. I have a “last minute” plan that came up meaning I will have to miss my over-unders on 3/14. When I plug this time off into the calendar, the prediction dropped from 294 to 289. Pretty big drop for missing one workout if you ask me.

I was just curious if anyone else has seen this type of change in the prediction? Am I really losing out on 5 watts for missing one workout?

Thanks and happy training!

It not missing one workout , it is if you repeat that pattern going forward.

All I have changed was one workout in the schedule and it was enough to knock the prediction down. I still plan to do all the other workout surrounding that day and up to the detection/prediction day. The day I have to miss is only 4 days before the next AI FTP detection.

You’re not really ‘losing out’ on anything - it’s a prediction with 3 weeks left until it crystallises. The odds are high it changes again before then.

I’d focus on the fact that it’s still predicting a 13w gain - nearly 5% is not to be sniffed at.

6 Likes

That is a fair point. I’ll be curious to see what it says at the next round of predictions. Being as consistent when I can is all I can do. Gains might be slower but I’ll still be happy to see them.

3 Likes

One could say opposite that watts should go up after a rest. Thats my AI guess anyway. TR AI guess goes the other way.

I guess it depends on how much rest? My hard days are Tuesday/ Thursday/ Saturday with my long z2 day being Sunday. Monday is off the bike.

If I can still get as long of a ride as I can on Sunday (typically 2.5-3 hrs) it hopefully won’t impact the prediction too hard.

My experience last year was that TR can’t determine what you do/when even if it was very regimented. AI maybe different or over sensitive, I don’t know. Little things make a difference currently - going for a swim lowers your FTP and so does missing a sesh apparently.

I think I’m having a similar experience with strength training lowering FTP.

The only evidence I have is that a friend who doesn’t strength train seems to have a higher FTP prediction than I do.

1 Like

I removed a 30 minute easy workout from my plan and my AIFTP prediction went up by 5watts, which begs the question, why put it there in the first place? The whole thing seems so volatile I’m not sure what to make of it.

3 Likes

This is how i’ve come to understand what the Ai is doing: This is a gross over simplification, but I think its about right.

The simplest analogy I can come up with is:

Think of a 4-week training block as a simulation. The AI’s job is to select workouts that fit a specific structure (provided by the training plan) and then predict your resulting FTP based on your completion of all of these workouts at a the predicted perceived effort (and HR if it has this data).

Imagine a stack of 100 blocks, the 100 blocks represents a 100w FTP. And lets assume there is 10 workouts in the 4 week simulation window.

The Ai looks at the first workout, based on your 100w stack of blocks, it picks a workout that will stack one 2w block on-top of your 100w stack (in TR this equivalent is moving your 6 week power curve up at some power / time duration). And this workout, if its difficulty was judged correctly, should feel hard and the Ai thinks this is an appropriate ramp rate in the workout difficulty based on your training history etc.

If the athlete completes the first prescribed workout, the stack of blocks is now 100 + 2 = 102w (equivalent to your athlete level increasing)

The Ai then picks the next workout, to stack another 2w block etc etc. (In reality the blocks will not all be weighted equal, some workouts will be worth 1w and some 3w etc but we will keep them as 2w to make things simple)

At the end of the 4 week cycle we therefore would have the 100w of starting blocks, plus 10 x 2w blocks, so we get given a prediction of a 120w FTP at the end of the 4 week window, awesome!.

The athlete then starts training. and the following can occur:

Scenario 1 (the ideal):

They do the first workout. They meet the power at the perceived effort simulated, the simulation is proving correct (so far) and the predicted chain of stacked blocks are still a 2w gain each time and the end prediction holds at 120w FTP

Scenario 2 (struggling with intensity):

They do the first workout, they meet the power target but the perceived effort was harder than expected. The Ai stacks one 2w block (as the power curve moved up), but concludes that stacking 2w blocks each time is too much, so reduces the remaining blocks to be 1w each time and therefore the final predicted stack becomes 100w + 2w + 9x1w = 111w and the prediction therefore drops from 120w to a 111w FTP

Scenario 3 (skipped workout):

The athlete completes the first 5 workouts at the power and perceived effort, but then skips 2 workouts, before completing the last 3. The final stack will be 100 + 5x2 + 2x0 + 3x2 = 116w FTP

Scenario 4 (fatigue):

The athlete completes all workouts, but after the 5th workout, brought in some fatigue that the model had no way of accounting for (went skiing, to the gym, huge Sunday ride, etc). the model then adjusts the final 5 workouts down to being only 1w gain each time to allow the fatigue to dissipate so the final stack is 100w + 5x2 + 5x1 = 115W FTP

Scenario 5 (most likely): A complex mixture of any or all of the above.

3 Likes

You have to think about those two things independently:

The workout selection and placement happens as part of your training plan. This plan has a specific long term goal in mind and aims towards it - possibly way further than the prediction range of the AI.

Meanwhile the adaptation and the FTP prediction are limited to only 4 weeks, which is a reasonable sweet spot for prediction accuracy and effort. If you move workouts around or remove them, you may optimize your short term target (FTP prediction) but loose out towards your long term one.

In a way it’s like the difference between TrainNow (best workout for you right now) vs Training Plan (best workout for your goal later)

How did your other workouts change from before and after you removed the workout?

AI says I lost 2 watts from missing 1 workout. No big deal because I don’t really believe that is outside AI’s margin of error. It is an estimate after all.

Clearly that is nonsense. Something else must be changing in four week window when you miss a workout.

My prediction still jumps about the same from when I posted this. 294 to 287 (even lower now since I’ve added some strength workouts) I’ve come to expect that my prediction will always be too high since it doesn’t know my strength training plan even though it stays pretty consistent.

To add to the mystery of detection and what ai tells us I had a “Dynamic Endurance” Ride for 2.5 hrs set on Sunday. I wanted to have ai program an indoor ride for zwift. I noticed the night before it brought my prediction up but then it changed the work out as I was getting on the trainer. Probably detected too much “fatigue” from a short xc ski I did with the dog earlier that day.

I wasn’t happy seeing the new easier workout and the predication being dropped again so I dug around trying to find a slightly more challenging z2 ride. The predication would go up and down depending on how much TSS I would add. Too little or too high and the prediction stayed on the lower end. I ended up finding a sweet spot where I thought the predication was highest. All to say I feel that the workout was not optimized. Why tell me to do an easier ride if I can push myself a little more to bring that prediction up?

Similiar to @Chop_Stick- if you remove the ride and increase your prediction why would the ai even tell them to do that ride?

I’m eager to get stronger but still slightly relieved that it has dropped some because at its the peak the following workout looked like it would be extremely tough.

AI is optimizing the schedule as designed for the best workout for the day. The overall schedule was made by a human either the athlete or trainer road.

1 Like

To put some perspective from the other side. During this training block my original predicted FTP was originally +4.9%. I spent 2 weeks in spain cycling with my partner, majority Z1/Z2 and few efforts but mostly easy and 650 tss weeks.

Watched the predicted dwindle to 1.6% with FTP detection tomorrow. Back a week and well rested so complete a 3.9 over/under todqy pulled forward from tomorrow. Nailed it and got 4 WL for threshold and rated it hard.

Predicted FTP jumped to 3.7% and that will be the number tomorrow on detection day.

Still can’t quite get me head around how the number changes so much based upon a single workout.

But the workouts are generally spot compared to predicted difficulty. I just wish it would provide the old ramp test AI FTP for comparative purposes.

According to TR the aiFTP was not predicting the result of a ramp test.

Does AI FTP Detection predict my hour power or expected Ramp Test result?

AI FTP Detection isn’t meant to predict the results of any individual effort, because all-out efforts can be affected by nutrition, fatigue, and other subjective factors. Instead, it’s designed to give you a training benchmark (FTP) that represents the full scope of your abilities, so you get the most productive training possible.”

Difficult to say as it seems to change every single time you do anything and I honestly can’t remember all the different workouts it has put on the planner. I think at first it kept things as they were, maybe changed a sweetspot workout but not 100% sure about that. It’s a kiddy kipper this AI Overlord as it’s constantly changing, even overnight on occasions, I’m honestly not sure I want to continue with it.