Take a look at levels in training plans. It looks like they progress you from about 4.5 to about 6.0. So I would say that if you can complete key workouts at 6.0 or above, your FTP has likely increased.
2 Likes
I wont be doing another FTP test until Iām upwards of 8 in threshold / sweet spot.
Adaptive Training messes with your levels too much when you do an FTP test which has screwed up my progress somewhat.
There is huge difference between the two. SST lvl 8 is quite easy (like 3x25@90% in Tray Mountain +1) and you should be able to do this without any problem with proper FTP. Threshold 8 - those are properly hard workouts and your FTP is probabably up or you have impressive TTE (example of lvl 8 workout for comparison - 4x20@100% FTP)
1 Like
Sure, I was at 10 SS last time and 7 Thresholdā¦ but my point was more that those are the only two zones I care about and what Iād base my progression on.
1 Like
I think ramp and FTP tests will stay with us for a while simply as an additional data point for AT, to see if it is on track. IMHO it is an important data point for the feedback loop.
2 Likes
You pose many valid questions here, that are on athletesā minds, I for one find this crucial when putting my trust into ANY complex training method, vs going on feel.
I have certainly never found (ramp test based) FTP to be leading in terms of my specific āskillsetā for my XCM races: a race with short but powerful climbs resulting in peaks and often less than a minute or so of total recovery, thatās a totally different ballgame than a steady-effort race with 20-40 minutes 5% climbs and a fairly long descent following. So not only do we need to be aware of our strengths and weaknesses in the different power zones, but need to adapt these in a tailor-made fashion to our race plan for a season or even several seasons out, almost like a career path.
For good measure I have compared my FTP values of this year to 3 seasons ago and while they were pretty similar due to the higher volume then (and longer race types), my overall achievements on identical training courses showed that in fact I now performed better with a lower FTP, probably due to the tailor-made TR trainings vs less specific training 3 seasons ago.
There is a point to be made here however: specificity vs robustness. The more specific the analysis of your training and resulting recommendations become, the more dispersed the focus of your training becomes and at some point, the robustness falls apart. I see this in all business intelligence nowadays and this is a looming threat for TR as well: if you start digging down deeper and deeper on an individual level, deeper and deeper into a personalized approach, dissecting your abilities and progression levels, at a certain point there is no value anymore in collecting data for a meaningful statistical use, since all data becomes anecdotal and impossible to find correlation due to the immense data load.
So there needs to be a balance between a robust system with a few key cornerstones, to make sure there are solid guidelines that can be determined and contribute to the vast majority of the Bell curve of athletes.
At the same time, the methodology needs to dig down deep enough so that the individual athlete benefits as much as possible from the aggregate data as well, because as soon as youāre slightly off the median mark, the training methods apply less and less to your specific human condition, be it genetics, response, recovery, etc.
I personally favour a 4DP typ of method for that very reason. It is a sweetspot with not too much and not too little.
A simple example of too little data with just an FTP: until AT the recovery time was pretty much tied to your FTP. Undoubtedly however a person my age (50) has a totally different recovery curve than a 20 year old. Still, the system never factored that in, therefore with just a universal FTP test fatigue and possible overtraining are a much higher risk factor for the elder age group, while the younger ones are missing opportunity to develop. So clearly more specificity was required here from the start and probably technology was the real limitation so far.
On the other end of the spectrum, we can see a situation arising whereby AT features dig down deeper and deeper to ensure a tailor-made experience for my needs, going as far as to allow me to put in my calorie count on an hourly level (so factoring in sleep/office/training/leisure differences), sleep patterns, time spent on activity outside training (plus intensity), stress levels, etc.etc. Whilst this will result in tons of useful data, it will certainly not result in any learning for the whole community, since replicating the exact same patterns and lifestyle is not only impossible, but with the sheer amount of different data the interdependencies are impossible to conclude, often even for the single person whose data it is. In other words: did I train well today because I slept more than usual, or because I ate less time before the training? Am I on an improving path or actually getting overtrained? The more data, the more questions on how to interpret results.
All in all, this discussion will play endlessly and we will probably never get to a point that will enable us to see clearly in this sea of data. The deeper the dive, the more we loose of the modelās robustness and finally risk having it fall apart alltogether.
Which brings us back to the matter of trust: AT and all similar methods are at the end of the day, by definition, a black box. The more complex the algorithms and methodologies, the more we simply have to trust the system, for lack of tools to verify its workings. With that comes the risk of allowing mistakes of the system to creep in that either we or the developers can no longer trace, as the human mind has its limits, while the systems get more and more complicated. Since TR has limited resources as well, there is only so much manpower and intelligence they can allocate to make sure the system works well.
So the point you raised is pretty much the āAI dilemmaā of the modern age: the more possibilities, the bigger the risk of loosing grasp on any kind of tangible results, for lack of ability to comprehend the complexity. Many companies underestimate the risks involved as it takes experience of working with complex systems to understand these risks. People in sports have little experience in this field Iām afraid and judging by the TR podcasts, this is pretty much uncharted territory for Nate and crew.
Which is why a sweetspot needs to be struck. In my opinion, 4DP is such an option, but definitely not a simple one-size-fits-all FTP ramp-test, so they did well by moving away from that. I think that AT is an experiment, where not even TR has an idea yet of the complexities they will meet along the way and this learning process will be painful.
Since it is pretty much built into the core concept of training, I think they have taken quite a risk by not submitting it for a beta for at least a year or two. Letās hope they have brought some heavy-weight analyst on board who has guided their methodology and painted a realistic picture of where this may lead them. I have seen billion-dollar companies fail at succesful transitions due to riding a hype train they had no idea about where it leads them to.
1 Like
Go look at the Sweet Spot Progression thread, and I think youāll observe that for many self-coached athletes, following basic endurance training concepts and achieving success is actually quite simple and straightforward. On these forums I believe some overrate the benefits of (adaptive or not) structured plans that look āmore interestingā and/or accept āscience based coaching principlesā without reviewing the science, experimenting, and reviewing their own results in response to widely acknowledged (āscience basedā) training principles. In addition I believe some underrate their ability to adapt/adjust efforts by listening to their body.
We all need something to believe in, whether that is self-coaching using the fundamentals of endurance training, using templated plans you buy (not rent) from respected coaching companies, TRās templated plans and machine-learning adaptation, or hiring a coach. Pick an approach, donāt be afraid to experiment, and donāt be afraid to expand your options and try a different approach.
8 Likes
I just feel so much more secure when my (limited) grasp of self-coaching and gut is in line with what TR tells me to do 
Ya, I hear you. Similarly, misalignment between my gut and TR should have been a warning signal.
Vaguely related to thisā¦ is there any workout / level that would give an indication that your FTP is about (or rather, at least) what you tested at?
So for instance, if I could do a 1.5 hour sweet spot at level 7, or a 1.0 hour threshold at 5.
I appreciate thereās no gold standard there, but Iām just wondering where the line is realistically.
I can, and have, hit level 10 on endurance, tempo, and sweet spot at over-inflated FTP - assuming you define FTP as the number you can hit for at least 30 min. Once you get to about a level 7 threshold, my scanning of the workouts indicates an accurate FTP setting.
But the caveat is that not all threshold are consistent. I did Echo +3, which is rated as a threshold 7.9. Best I can figure is that is rated so high because there are no breaks. But since I am good at sweet spot and sub-threshold work I can hit it. Give me a threshold workout that goes sustained above 95% of my current setting, or worse, above 100% sustained (like 8 - 10 min intervals), and I wilt quickly. There are like level 5 workouts that Iām pretty sure Iād fail. Lamarck (not super high level) is always tough for me - maybe because I need a longer warmup?? There are some 45-55 min workouts in the 40K TT plan right near FTP that are TT simulations that are brutal. Once I can hit those, I know my FTP setting is accurate.
That gets into part of my gripe on focusing on it. We use it to set training zones, but you can figure that stuff out without it. Unless you need to sit at FTP for a long time, there isnāt much point worrying about it. Just use the levels in the different zones that matter to you and try to make those numbers go as high as possible. Simple as that.
In this case Iām defining it as whatever I get from a ramp test (the number is largely meaningless). Looking back at power records, I donāt think Iāve EVER hit FTP for 30 minutes, but then Iāve never really had call to try as my sessions tend to top out at 20 minute intervals anyway.
I have a nasty tendency to do too well on the test and then really struggle withā¦ well, basically all the sessions afterwards (at least on the old plans). Iām also looking for markers I can use to see if I need to retest after a break / big event (I do ultra endurance / multi-day stuff).
It doesnāt matter if you define FTP as the ramp result (which isnāt a good assumption). You can tell if your FTP estimate from a ramp is pretty close if youāre making level 5 and above threshold workouts regularly. Once youāre at level 7 or 8 on a few, itās very accurate.
Iād look at ramp result as a FTP potential if you trained it up. The nature of ramp means you might be higher than what it estimates or lower or spot on. You donāt know until you start doing threshold work.
Best measure of fitness is to go out and do different efforts in specific zones you care about and look at a smattering of your performances. Once youāre out of that initial rapid growth phase of fitness, things are pretty stable. I.e., not changing much over a 1-2 week period at least.
Sure, but Iāve been training with power for more than a decade now so Iām not really worried about gains like thatā¦ hell, or any gains at all recently!
I think Iām just going to pick a few benchmark sessions that Iāve done quite a bit and use them as an indicator of where I am vs what I tested it.
Iām catching up on the podcasts, but Coach Chad just covered this super well at like min 52ish of episode 326.
Use the discipline specific workout performances to give a more precise indication of fitness rather than imprecise assessments that are used to anchor against imprecise percentages of that for different energy systems.
Using FTP as a guideline for setting training targets and fitness is a great initial guess. But itās analogous to using a language translator to understand what someone is saying. It works in a pinch, but much better to speak the language.
I was thinking a bit about this, and ultimately I think in order to be successful you need to have a decent ballpark of knowledge regarding FTP to be successful in using progression levels. An underestimated FTP might lead to just doing a lot of easy work, whereas an overinflated FTP will make it really hard to progress in certain energy systems.
Iāll have to go back and listen (I probably wonāt though lol) but I think the team had said level 4 workouts are a place where people should be able to complete workouts. It would be good to better ascertain the minimum progression level one should be able to complete in areas with a pretty properly estimated FTP
Surely thatās only the case initially? If the adaptive training is doing its job then it should progress you up the levels fairly quickly.
On my other point about ābenchmark performancesā. Is there any way of searching for what sessions youāve done most?
Iām not sure, if someoneās FTP is really overinflated, they may be able to do some lower level workouts but may hit a ceiling for progression levels. If someone is doing vo2 workouts but in reality, due to their inflated FTP, theyāre actually in anaerobic zones, they may not be able to progress beyond a certain duration of vo2 intervals. So if, say, someone canāt progress beyond a level 3 in vo2 due to an overinflated FTP, Iām not sure anything other than fixing oneās FTP is going to solve anything
Iām pretty confident AT will work just as well with too low of a setting than too high a setting of FTP. Iām not confident it is so hands off you get the best training from that - assuming you have a good understanding of how to train. But if you left it completely alone, Iām not sure why it wouldnāt continue to increase difficulty. I think a baked in feature of it is a set progression rate goal. The FTP setting + training history sets the progression level starting point, but the rate of progression is fairly constant and only gets modified if you struggle or are (I think) stating all your workouts are easier than the system expects - i.e., pretty much rate everything as easy.
I tend to always be trying harder stuff than the system thinks I can do. So iāve never been patient enough to see if it will correct, making this speculative based on listening to the team discuss it.