Failing Threshold Workouts

Thank you for being on the forum and becoming voice of reason :slight_smile:

1 Like

I haven’t used TR in a long time, so can’t remember if it’s possible to reduce the %FTP like Zwift’s bias. If it’s possible, drop the load a few % to a sustainable level and finish the workout. So instead of doing 294W you’re doing 270W. If that’s too easy, then up it to a level you find sustainable. Threshold training is 90-105%. Some days you feel cr@p and ride at the lower level/range. The days you feel strong, you hit the higher half of the level/range.

2 Likes

some of us are are able to over achieve on a ramp test and get our FTP too high. After an injury and then illness, AI and honest ride evaluation has helped me hold my true FTP down. After 4 months of consistency in riding achievable workouts I’m close to being 100% back. It’s human nature to get an ego about a high FTP #, but consistency in workouts is were my results are coming from. (Learned that from Jonathon on the podcast)

2 Likes

AI Detection suggested a 212 FTP, but about a week and a half prior to that detection I had done a 20 minute test where I could only hold 213 for 20 minutes (@95%, 202 FTP). Based on an actual test, I knew that 212 would not be feasible for me, so I kept my existing FTP of 207 (this had also been a prior AI generated number). I want to slowly progress through my PLs at that 207 and reach workouts that actually give me some confidence that I justify that 207 AI FTP before jumping any higher. I think there is some sort of anomaly in my outside group ride numbers (which are highly anaerobic) that throws off the AI to give me a higher FTP than my threshold capability merits.

1 Like

Smart… very smart.

It doesn’t matter how your FTP is estimated… if you like AI FTP detection and ramp tests, have at it. The problem with them is that very few people will do what you did - that is they won’t actually try to perform at that level because the big number makes them feel good.

And I’m not just picking on TR here. You can do the same thing with WKO5 mFTP and intervals.icu FTP estimates. If you populate the models with a lot of high power, shorter aerobic intervals, they’ll give you artificially high FTPs too. WKO5 has been telling me my FTP is 300W for the last month, and I know that’s probably 10W too high right now based on what I can actually hold.

My advice: set your FTP based on performance of long intervals, and remove as much estimation as you reasonably can. If you’re not good at pacing… practice it and get good at pacing.

4 Likes

Some stretch workouts have been easy for me and some very hard. I’ve done productive, even achievable, workouts that were very difficult to complete. Another thing to consider is how a workout could feel if your FTP has gone up or down but it hasn’t yet been detected or reflected on a recent test.

@kurt.braeckel Yes, that’s the route I’m taking. I actually think AI FTP is a wonderful thing to have developed and to continue to refine. I think it’s a great tool for athletes to use. I’m just going to be thoughtful in my approach of how and when to use it based on my performance and on periodic actual max power tests of at least 20 minutes; especially now given it is in its early stages of development.

I think it can be a great tool for athletes to use in context. The problem is many TR athletes don’t have the context… you clearly do. The number AI FTP spits out at you isn’t magical - as I mentioned WKO5 and intervals.icu among others give estimated and modeled FTPs and have for years. The problem is when people blindly trust those numbers without understanding what’s behind them, why the number is what it is, and what their performance is actually telling them.

I’d wager the average TR user lacks that context. Some will have it, and in that regard, AI FTP is a great tool… it’s just not a novel one.

1 Like

Apologies if this is irrelevant, I am 20 posts behind.

I think there is definitely a risk of pushing your PL’s up just beyond your actual level as you may have done. either with short effort workouts that you’re good at like you did, or I got “close enough” on some high level Thresh workouts l, bumping my PL further but constantly driving myself down. Until I worked out to just choose the lower level workouts at the power I was actually riding, but it takes a while for the PL to come back down.

I’d get off the trainer, you can’t hold as much power on the trainer. If the trainer road ai thing is seeing your outdoor rides it also may be the cause of the over estimated threshold. Do a real threshold test inside for trainer rides and one outside for outside rides. And get a REAL coach, there just isn’t a substitute for a real person coaching you

1 Like

I’ve struggled with this exact situation for a couple of years, @SOWE34. As @tshortt suggested, a lot of us on this forum have a high anaerobic contribution.

I finally realized, after reading this post from Coggan An Introduction to the New iLevels in WKO4 | TrainingPeaks and analyzing my heart rate data, that:

  1. FTP is not “hour power” for everyone (if you break Figure 2 in the post down, you realize the range of durations at LT is ~20-80 minutes)
  2. I need to train sub- and supra-threshold efforts at different power targets if I have any hope of raising my endurance base

The HR data really drove home how much I over-test by ramp and KM.

My last FTP measured by AI FTP (before a vacation) was 325. My HR and RPE suggest training at ~280 makes more sense in the sub-threshold zones (and I include threshold workouts in this range) while ~340 makes more sense for supra-threshold zones like VO2Max. Using a single FTP number to train at both ends meant I wasn’t getting the intended physiological adaptations from base training and was exclusively improving duration at high power outputs.

I’ve decided to redo SSB at LV with long Z2 rides added in to lift my floor and meaningfully focus on my base.

Not sure if any of this resonates with you, or is helpful.

I have no outcomes data to share, but I’m happy to report on progress in a few months for anyone else struggling with this issue if interested.

6 Likes

I started to purposely underestimating my ftp (used as a training metric) about 5W - I have never been so consistent and confident in my workouts in my life. I am never gonna over-estimate FTP ever again - I feel so liberated.

7 Likes

:point_up_2: This is where AT goes wrong sometimes - essentially it would have correctly lowered your progression level after your FTP increase but the system allows users to just bump it straight back up by doing one stretch workout :person_shrugging:

From that point on all the workouts are going to be slightly harder than they would have been until AT brings them back on track.

I’m absolutely loving AT and AI FTP detection - I decided to go into it with an open mind and have been blindly agreeing to every AI FTP increase and adaption and have seen great results.

Agree that the AI FTP number doesnt really reflect what people would call my “real FTP” (whatever that means) but my training has felt spot on.

1 Like

Best post on this thread. When I (rarely) FTP test I round down to the nearest 5w. Ok, it could only be a couple of watts reduction, but it just gives a little headroom at all times. As @C_Nay says - outcome is consistency, which breeds confidence. Threshold workouts shouldn’t be terrifying - should be fairly hard work, but nothing to worry about.

3 Likes

I struggle with the same thing you struggle with. I would rather do a threshold workout at 105% then the over unders. Like stromlo+6 felt moderate even though it was a stretch by 1.5 and last work out of my 3 week build. When I fired intervals back up after my rest week I did picket guard +1 and felt it was very very hard. I really need that clear delineation on my rest after time above threshold. I would try other workouts around the same score and see how you do on those workouts

My experience has been that AI has occasional hiccups, but fairly quickly gets back to prescribing the workouts that are appropriate for me. So my suggestion would be to stay the course and do the workouts that AI feeds you, letting it “come to you” rather than the other way around.

Two things to keep in mind:

  1. You can probably throw AI off temporarily by nailing a stretch workout that happens to play to your strengths, as others have mentioned. You can see what workout is responsible for your current progression level by hovering over the red bar (in the website; not on a phone yet; I think it’s enabled on tablets). If it was a workout picked by you rather than AI, that may indicate you inadvertently goosed your levels.

  2. Rather than changing your FTP or manually reducing the workout intensity, click on “Alternates” for the prescribed workout and look for something that seems within reach (change the drop-down filter to “achievable” perhaps).

If you really want to get a better read on your sustainable power and don’t want to wait for AI to come around, do Lamark. It’s 4x10minutes at FTP, but if you start off conservatively and increase a little each interval, you’ll get a pretty good idea of what kind of sustained workouts are currently within reach for you.

Before AI FTP D I did something similar last season rather than doing 20min tests all through the season (I suck at Ramp tests) and only made gradual changes to my FTP. Time wise for my TTing career it produced the best/most efficient results. I’ve not got on the same courses in similar conditions to tell if AI FTP D is faster or not for me; my instinct suggests not though :-/

Ditto, I asked the “Is it better to train with an FTP estimate that’s too low vs too high?” and personally feel that low is the better of those options. That has been my experience (FTP on the low side via testing early, and now with AIFTPD) and I have left it that way vs manually adjusting upwards. My workouts and general progression are more consistent and attainable vs times in the past where my FTP was “optimistic”.

Great if you can get an “accurate” FTP, but if you are in doubt, I think lower is safer / better. :smiley:

7 Likes

I agree. I would rather nail all workouts then fail one a week

2 Likes

Other than the uninitiated (and there is nothing wrong with being the uninitiated), I think is generally accepted now. That’s a good thing.

So……:grin:

Interested in thoughts on how much lower? Generally I hear 5-10w as a common response. I contend that 5-10w isn’t actually lower (measurement errors, etc). It’s the same.

I recently heard a prominent coach suggest much lower than 5-10w as “close enough, still got very fit”.

We’re doing the Limbo Rock.

2 Likes