(Just kidding - I love math humor)
4 Likes
Good stuff.
Without seeing the actual data that looks like a shift from a Threshold TiD to Pyramidal TiD depends on the stuff between LT1 and LT2, but it doesnt matter, you’ve trainned significantly more and arent shattered, thats a win.
Congratulations on the 16 hours that is a big week
4 Likes
Or, do as many hard interval trainings per week as you can digest productively. Usually not more than two. Then fill the rest with as many long z1 rides as possible WITHOUT harming the quality of the hit sessions.
1 Like
What you leave out is that this scheme, 80/20 polarized for session goal, probably applies to people with >10 hrs to train. If you just have 6 hours, more hard sessions are going to be manageable with proper nutrition and sleep. And to my knowledge this hasn’t been studied, at what number of hours per week does polarized make sense when done long term.
Also, mechanical efficiency gains are very important when precise technique leads to much higher speed: Nordic skiing, swimming, rowing. The way to get fast at the elite level in those sports is through many hours of technique practice, and to manage The fatigue with many hours, eg double sessions, you have to go easy a lot. Mechanical efficiency gains are less important for cyclists.
3 Likes
Indeed. And while I’m certain it was helpful for some people to get a bit educated on how to read a paper in general and how to the evaluate the significance of results of a typical trainings physiology papers, where also the best of them have to cope with short intervention periods and abysmally few testing individuals you at least have to question the motivation here. Or rather I guess - we all know the motivation. But - investing so much time by the team into that tangent and thereby at best circumventing the underlying issue which others, me and now again very truly you have laid out has one wondering.
No wonder indeed that they fell so easily into that first incarnation of the PI-Graph which was not only completely wrong but also used a very questionable and suggestive way of arranging the lines on the graph. Textbook “how to lie with statistics” wise.
And which as we have established completely and utterly missed the point. So much so that it indeed left a sour point.
And completely unneccessary. The complete TR package is a nice value proposition. The team is a bunch of likeable and knowledgeable persons. Hell, as I watched the last part of that video where Nate mentioned “How much he would talk about menstruation when he had it” I had to grin from ear to ear. And Amber hit right on with menstruation on sea level as compared to Reno… ^^
But well… they should edit that video again, give it another title and just put it into an independent section: Fundamentals: how to read sports science papers.
3 Likes
I’m replying to this one just because I’ve seen a few criticisms about sample size here. Amber spent a lot of time with the intention of helping people read papers, so this is part of that theme.
First off, there is an over reliance on p-values when interpreting studies. Generally, all this says is: assuming the null hypothesis is true (there are no differences in the distributions for groups being compared), then there is a less than 5% chance of arriving at these results by random chance. The traditional alpha level is 0.05 for a single comparison study.
Due to the nature of a power calculation, having a small sample size requires a large effect size in order to reach the alpha level. In other words, the measured difference between the groups needs to be big. That’s the trade off.
With big data studies, you have very large sample sizes and this means the effect size can be very small in order to be “statistically significant.” The clinical significance may not be there. Have you seen hyperlinks or magazine articles saying “Dark chocolate (or any other food/supplement) reduces cardiovascular risk)”?
Bigger sample sizes don’t make something more valid on statistical grounds, it’s all just math.
For trial design, you usually start off by defining a clinically meaningful difference and then find out how many people you would need in the groups to detect that degree of difference. Recruiting for studies is usually pretty hard due to the demands on subjects’ time, so you just aim to hit that floor.
External validity is probably limited when you have a small group, but this is not necessarily fixed by having larger sample size. For example, if you study 10 U23 white males or 1000 U23 white males, you will still be left wondering if the findings would apply to other ages, races, women.
If you look at the phase 3 clinical trials involved in getting drugs on the market, you might be surprised how small the studies are.
3 Likes
Just a general question about the sessions idea.
Because a lot of Seiler’s original work came from (at least as I understand it) observational data from sports where multiple sessions in a day is common (12-16 sessions/week). In cycling it is not common to do two-a-days or three-a-days
So, in the case of cycling, how do you define a session? If I go out for a three hour ride and in the first hour I do my Z3 intervals and then the last two hours are Z1 - is that one session? What if I stop at a coffee shop for 15 minutes after 90 minutes? Stop for 30 minutes?
I’m genuinely curious - because I think so much of the ‘fatigue management’ aspect of Polarized training is incredibly valuable. Personally, I thrive on two hard sessions a week, and that is my main takeaway from my own body’s response to TR plans and other relevant training methodologies.
Certainly from a time perspective I train very polarized, but many of my workouts look like an hour or so of intervals followed by 90-120 minutes of work below VT1. Doesn’t really change anything for how I approach things, but I am definitely not Polarized if that all counts as one session and you are using sessions as your meaningful metric
It is one Z3 session.
If you refuel, hydrate, and your metabolic, epoc levels return to somewhere near base, (probably going to take around 3 hours) then I guess there is reasoned argument that a Z1 session later in the same day is a second session of Z1. But imo it is kind of missing the point, the point is how much / many Z3 sessions your are doing over a period of time i.e many less than Z1 sessions.
In my comment quoted below change the hour of Z1 to 2-3 hours add a coffee stop the same still holds true.
1 Like
Fair enough - I wonder if/how they tracked this in the literature around multiple sessions within a day. Nothing I’ve read in the papers showed that they tracked duration of recovery time before sessions, but it is possible I missed it
This I completely agree with, but - is this where the 80/20 split becomes meaningful? Because even then most people would be looking at 1 hard session a week
Having caught up on a few other posts - I do think that comparing TID at an 80% ratio is ludicrous. Honestly by a time perspective, and maybe only talking lower case polarization, if you’re doing 95/5 or 90/10 you’re probably much closer to what makes sense
Ultimately, I think these things are where people get too caught up on following the specific ratios and ‘rules’ of Polarization. Everything I’ve heard from Seiler runs contrary to this, but many people are saying you must do 80/20 or some other ratio or else you’re not a true believer. Seiler doesn’t seem to advocate for this, and really seems more focused on decreasing intensity frequency so you can increase intensity quality. To me that is the primary goal of ‘polarization’ in any form - improve the quality of your hard days by reducing their frequency and thus your fatigue
For each of us these ratios will be different. How many hard days, how many easy, how long should they be, does X% of VT1 count as hard or easy? These are the things we all have to figure out (or have a coach or a computer figure out) about our own bodies
Good thought, thanks. However, I’ve sworn off social media so I’d have to make an account just for that purpose 
2 Likes
Yes.
I see it as a rough distribution, no need to get caught up in the finer details, yes 80/20 is the popular one quoted, but…
to be polarised you just need more Z3 than Z2 and more Z1 than Z3, that is it. (imo days or sessions.)
Then you’ve got the TiZ, Sessions, days debate. Personally I fall in to the session or days camp, I have explained why my thinking is the way it is, imo opinion one needs to think about what the idea is and what your are trying to achieve no cheat or adapt things to hit a ‘perfect’ ratio (no such thing.)
1 Like
Could your logic not be expanded to defeat itself?
Start the first week 1 with “hard” sessions then proceed to have 4 weeks “easy” sessions this would equate to 80:20
From what I understood with the pod was that they were using the logic that the research papers presented with TiZ. Some of which are the same papers that Dylan cited.
1 Like
He says on the TR Forums 
3 Likes
Does a forum count? Shoot
4 Likes
Was it a slip of the tongue or did Nate say he wanted to develop a Nutritional product for TrainerRoad? It was at the end of the podcast.
TR Gels for indoor training?
@ambermalika @Nate_Pearson This is a great discussion and I genuinely wish I had time to go through, what is now over 400 comments. Unfortunately I do not. If what I have to say has been mentioned previously I apologize.
I listened to the podcast and I thought it was well presented and laid the facts in as clear a manner as possible given the topic. I do feel however that it missed the mark a little, for me anyway. The podcast does lead me to think I have some more insight as to why much of the criticism is cast toward the SSHV programs. Keep in mind all of this may be moot with the new AI Adaptive Training.
For many of us, we have 8+ hours per week to dedicate toward training (400-500 hours/year). It would make intuitive sense then to gravitate toward a plan that best utilizes those hours. Unfortunately this is not necessarily the best idea given the intensity of the SSHV plans. Based on having listened to over 95% of your podcasts, I would guess that your recommendation to someone who has the time but not the capacity to do SSHV would be to do a lower volume plan and add in some lower intensity rides - please correct me if I am wrong. The inherent flaw with this is the mid and low volume plans already have quite a bit of intensity and trying to recover from that intensity while adding more stress (albeit at a low intensity) reduces the likelihood of being able to perform the HIT workouts well. It takes a lot of self knowledge to know when to go at a higher intensity or when to back off. For those of us looking to “just show up and ride” that level of self awareness may not be there and I think this is where problems begin to arise. I recognize that SS may not be technically high intensity (Z3 in a three zone model), as Jonathon has described in the podcast, energy systems are a fader not on/off. If an athlete is working at 95% of their FTP, that fader is pretty red and not much blue. Further, on any given day 95% may be over threshold even 85% can easily be in the threshold range if the day is not going well. Further, depending on the anaerobic contribution to an athlete’s FTP even in SS it is possibly that you are working at a very high anaerobic energy contribution simply because your physiology contributes so much anaerobic contribution to the max power on a ramp test. To be clear, I am not criticizing the ramp test, I think it is a good tool. I am saying that it has limitations and athletes need to know what those limits are and how to spot them. Personally, I like the idea of a confirmation workout of say 2*20 minutes at threshold (+/- 3% say) with a relatively short recovery, a day or two following a ramp test is a good idea to confirm the FTP value. This should be a very hard but doable workout assuming the athlete is reasonably fit and recovered. If not then the FTP is set too high. It’s a little simplified but you get the idea.
As far as the debate today, here’s my $1.40…
For any endurance athlete to improve I am reasonably certain that Seiler’s “hierarchy of endurance needs” model is as good as any to look use:
- Frequency and Consistency of workouts
- Volume of work & Duration of workouts
- Training Intensity of individual workouts & Intensity Distribution of overall training
- Periodization
I have bastardized it a little, but that’s how I look at it. One of the things that I like about Seiler is his respect for the knowledge of coaches and athletes and this hierarchy was born from that knowledge and his own understanding of physiology. Based on my understanding and experience, if you can conquer the first three, that’s about 90% of training progress right there. If you can get the forth then that’s another 5%. People will, quite rightly, say that nutrition, rest and recovery are a significant parts of endurance needs. My take is that those two (vital) parts are inherently required to perform the first three parts of the hierarchy. Nutrition, rest and recovery are therefore ‘baked-in’ to the first three. I am an evidence based athlete and coach and I recognize what I have presented has little (direct) foundation in scientific evidence but, as was highlighted in the podcast, there really isn’t a lot of long term studies in comparing training methodologies. To me at least, this hierarchy in combination with the basic training principles (progressive overload, reversibility etc.) form a significant part of the art of coaching.
If we assess the TR plans using this (admittedly, largely unscientific) hierarchy then we see that the frequency and consistency as well as the periodization are well covered. What I do not think are well covered are the volume and duration as well as the intensity and intensity distribution pieces. I looked at the completion rates by workout duration chart that was provided in this discussion, which clearly show that the longer the workout gets the lower the completion rates. While I understand that is why the “weekend long ride” was replaced in many (all?) plans with a shorter SS ride I do not believe this was a service to the customer. TR’s mission is “to make you a faster cyclist”. In order to accomplish this mission all energy systems need to be stimulated in order to see adaptation. There are certain aerobic adaptations, vital to an athlete’s development that can only occur in the 3rd, 4th and 5th hour of a workout, especially as a rider develops over time. A strong encouragement by TR to its customers to do long lower intensity rides whenever time allows is vital to any endurance athlete’s development. This addresses the workout duration. In terms of volume, as I described in the beginning, there needs to be more direction for the athlete with more available time but not the capacity for the amount of intensity that the SSHV prescribes.
As for intensity distribution…this is really what this discussion is about isn’t it? I think it is foolish of everyone (myself included) to make this about threshold versus polarized. Full disclosure, I do feel that, generally speaking, a polarized approach is a safer and more sustainable. That said, if an athlete only ever trains with supra threshold intervals and long easy rides they will get far but they will not have all the “arrows in the quiver”, so to speak. Tempo, SS, Threshold as well as anaerobic capacity all have a place in training but only ever train in these zones is even more foolish that never training in them because the risk of burn-out and non-functional overreaching is so much greater.
Just to kick this dead horse some more, while I understand why TR did the analysis using the polarization index, I think we can agree that this method of assessing a training plan is not great. Unless I am mistaken, this polarized index analysis includes the low intensity recovery valleys between intervals and the warm ups and cool downs…this is not how any coach I know would think about intensity workouts. A hard workout is a hard workout the low intensity within it is not considered. Using this method I could do threshold and VO2max intervals everyday and still have a polarized plan if my warm up and cool down are long enough. The original Seiler research that coined the polarized term, was based on the notion of “buckets” of high medium and low intensity workouts. The PI does not really use that method as far as I understand.
Finally, while it may not seem like it, I am a big fan of TR. I’d like to give a major kudos for having done this podcast and launching the Adaptive Training AI as well as the launch of the polarized plans. I look forward to seeing how it all comes together. I genuinely appreciate @Nate_Pearson 's management style and progressive thinking. I wish I could work at TR. I look forward to see what’s next!
9 Likes
Exactly. A finger pointing at the moon is not the moon. It does indeed seem like some level of polarisation is beneficial in endurance training. But calculating hard vs. easy sessions or TIZ are really just approximations.
In isolation, it was a really good podcast. In itself, I think the observations and caveats they had around most the studies seemed quite fair. And I like the people. But I agree, that it did skirt some of the deeper problems with the TR plans.
If they would do a similar deep dive on pyramidal training, I think they would find even less compelling studies than for polarized.
What I like about Seiler’s polarised training idea, is that I have time for endurance level training. I get a lot of reading done while doing endurance. On weekends I don’t mind doing long low intensity rides either. The tempo level and sweet spot seems to burn me out. So the idea of doing one heavy day and a bunch of endurance days will fit my lifestyle. I also like doing marathon multi day MTB stage races.
4 Likes
I would still like to see the Neal et al. study be a little more thorough.
-
Assuming 240w for a 40k TT as elite is shortsighted
-
With such small sample size having a sleep log would be nice.
-
I know there is mention that all athletes stayed relatively the same weight and had a food diary for 1 week. Why not have them do a food log the entire time.
-
The fact that during testing the athletes only see distance traveled means they are in a mental 1 hour battle.
-
The study also mentions that for athletes doing the POL if their HR met certain criteria the intensity was increased. This in itself will change results.
-
The study had 3 (pre-, post-, 4 week taper) testing phases IMO but presents it as 2 (pre-, post-)
Outside of #5, 6 I feel that with greater sample size you avoid these lurking variables that would have a big impact. I am not going to get into the weeds of TSS
1 Like
Recovery valleys weren’t counted, according to Nate. It’s part of properly calculating the PI for the workouts. See below.
2 Likes