Polarized Training Deep Dive and TrainerRoad’s Training Plans – Ask a Cycling Coach 299

[User67890]

Thanks, Chad. I hadn’t noticed the time stamps in the first post. I told you I was impatient

I’m just excited! I’m going to give the new plans a try. Not like there are races this year for me… And not like I’d win if there were.

[Pbase]

So something really jumps out at me seeing the breakout by plan in these charts. We’re talking about MANY MANY plans, but I’ll bet a very large number of users are using SSB Low or Mid most of the time. That means two of the most utilized plans are also two of the least polarized plans, and that skews perception of the platform as a whole.

I listened to the first half hour or so, but had to go back to work. Looking forward to listening to the rest. Amber’s CV is impressive!

[Nate_Pearson]

@Pbase

So something really jumps out at me seeing the breakout by plan in these charts. We’re talking about MANY MANY plans, but I’ll bet a very large number of users are using SSB Low or Mid most of the time. That means two of the most utilized plans are also two of the least polarized plans, and that skews perception of the platform as a whole.

You have to take them into account in terms of a season, and based on the “polarized index” they are still “polarized”.

[enki42]

To be more specific seems like tr could do this measurement during the ramp test

Zone 2 in adaptive training

I figure this should be a new thread to not get lost… Are the power targets in the new adaptive training staying the same as they have been? As a percent of your ftp. So the adaptive part is just the length of intervals and rest breaks and how many of them while leaving the target the same? I ask this cause most zone 2 exercises are really the same then. They all target a percent of ftp and not based on your LT1 which may be a better target to base off of

Anyone using HRV? (Heart Rate Variability)

hi all, I wanted to share a bit of data that might be informative for some as there isn’t much out there on differences and similarities between night and morning measurements Below I show the two tools I use (camera measurement in HRV4Training in the morning) and Oura ring (night data). You can easily extrapolate to using another app / sensor for morning measurements, while it might be a bit harder to extrapolate to other ‘night HRV’ sensors, as many do not provide the average of the night (w…

[martinheadon]

@Jonathan

My initial comment from this chart is that it doesn’t really represent what I think of as “polarised”, as I would tend to define it by workout distribution rather than simply Time in Zone.

Criterium low volume comes out as polarised because all the rest intervals in between the Z5/6 bursts are at Z1. But it’s 3 intense HIIT sessions per week - for me a polarised training model is not just 80/20 in terms of TiZ but also in terms of workouts. E.g. 2 1hr vo2 sessions, a 1hr Z2 ride, and a 4hr Z2 ride.

[ambermalika]

@martinheadon

My initial comment from this chart is that it doesn’t really represent what I think of as “polarised”, as I would tend to define it by workout distribution rather than simply Time in Zone.

Criterium low volume comes out as polarised because all the rest intervals in between the Z5/6 bursts are at Z1. But it’s 3 intense HIIT sessions per week - for me a polarised training model is not just 80/20 in terms of TiZ but also in terms of workouts. E.g. 2 1hr vo2 sessions, a 1hr Z2 ride, and a 4hr Z2 ride.

Yeah this is something we address in this podcast. There is a big range in the literature even among researchers on how to define “big P” Polarized Training Intensity Distributions. Treff et al proposed the Polarization Index. When we look at our plans based on that equation, we see a big range of what we would call “small p” polarization, but our plans fall more into a Pyramidal Training Intensity Distribution, with varying degrees of “small p” polarization … all based on time in zone. I know others use the ‘session goal approach’ or a binary binning of individual days. We don’t claim our plans are Polarized, but this chart serves to illustration the range in degree of small p “polarization” across the many plans that we do offer.

[Nate_Pearson]

@martinheadon

My initial comment from this chart is that it doesn’t really represent what I think of as “polarised”, as I would tend to define it by workout distribution rather than simply Time in Zone.

Criterium low volume comes out as polarised because all the rest intervals in between the Z5/6 bursts are at Z1. But it’s 3 intense HIIT sessions per week - for me a polarised training model is not just 80/20 in terms of TiZ but also in terms of workouts. E.g. 2 1hr vo2 sessions, a 1hr Z2 ride, and a 4hr Z2 ride.

We cover this in depth on the podcast.

This polarized index is used for TiZ and is often used in studies used to criticize TrainerRoad. This is different than the 90/10 TiZ and the 80/20 intensity distribution that Seiler talks about.

[Pbase]

@Nate_Pearson

@Pbase

So something really jumps out at me seeing the breakout by plan in these charts. We’re talking about MANY MANY plans, but I’ll bet a very large number of users are using SSB Low or Mid most of the time. That means two of the most utilized plans are also two of the least polarized plans, and that skews perception of the platform as a whole.

You have to take them into account in terms of a season, and based on the “polarized index” they are still “polarized”.

True, but are most users doing that? Almost everyone I know who uses TR just does Base over and over with some Build mixed in, but then quit half way through and go back to Base. So… repeating those un-polarized plans over and over.

It might just be my group of acquaintances, who tend to be over 40 and working busy full time jobs with kids.

[Nate_Pearson]

@Pbase

True, but are most users doing that? Almost everyone I know who uses TR just does Base over and over with some Build mixed in, but then quit half way through and go back to Base. So… repeating those un-polarized plans over and over.

Yes they are.

[Pbase]

@Nate_Pearson

Yes they are.

Interesting. Thanks. I have to admit, I assumed most users never get to Specialty based on my relatively small group of friends.

[VeloJake]

Pretty sure that was me that commented first on no Chad, and it wasn’t a complaint, just wasn’t what I was anticipating. That being said, I think it was a really good podcast.

[toyman]

Sorry to belabor my point, but I find this PI somewhat interesting as a way to classify TID and provide some objectivity to ‘how polarized’ something is.
Doing a quick look at SSB LV1, which on your graph has a PI of just under 2.1, I tallied the following TID, by just counting z2/z3, and considering the rest z1.
I came up with:
1225 minutes total in plan
56 minutes zone 3 (4.6%)
492 minutes zone 2 (40.1%)
677 minutes zone 1 (55.3%)
This is likely slightly off, as I just guessed that 10 minutes of Baxter was z2.
I’m confused how this could be considered anything close to polarized.
Looking at the paper and doing the math, I got the following (math done in python)

pi = math.log10(.553/.401*.046*100)
print pi
0.802338590366

This is a very different result. If I use percent, rather than fraction I get something closer, but still quite different.

pi = math.log10(55.3/40.104.6100)
print pi
2.80233859037

I think that the first formulation is correct, based on my reading of the paper, and in particular the line:

For the polarized structure, we agree on the following necessary conditions: a: Zone 1 + Zone 2 + Zone 3 = 1

Since neither of my calculations are in line with what you reported as the PI for SSB LV1, I’d like to understand what I am doing wrong, and be able to reproduce your calculations. Tweaking my TID slightly doesn’t make that much of a difference, and I don’t think my quick tally is that far off. If you could use SSB LV1 as an example and show the TID as well as how you did the calculation that would be very helpful.

[Joelrivera]

I know that the tt plans are harder than others… Because of running and swimming, but no new plans for tris?

[martinheadon]

@Nate_Pearson

This polarized index is used for TiZ and is often used in studies used to criticize TrainerRoad. This is different than the 90/10 TiZ and the 80/20 intensity distribution that Seiler talks about.

OK got you. So in fact, a capital P Polarised plan would in fact be even more small p polarised in terms of TiZ?

I’m trying a block of vo2 polarised, and for TiZ I’m currently at 71% Z1/2, 10% Z3/4, 19% Z5+, and that’s just from vo2 sessions with a bit of Z2 added on - I haven’t even done my long slow distance Z2 weekend rides yet…

[VeloJake]

Am I the only one who gets annoyed by PI being anything other than 3.14159…

[Nate_Pearson]

@martinheadon

OK got you. So in fact, a capital P Polarised plan would in fact be even more small p polarised in terms of TiZ?

Correct! You got it!

[bbarrera]

@VeloJake

Am I the only one who gets annoyed by PI being anything other than 3.14159…

22/7 is good enough at times, and therefore not annoying. On the other hand 355/113 is too complicated and therefore annoying.

[Nate_Pearson]

You have to count > 50% FTP for Z1, don’t count recovery time.

[paul_rides]

This is a minor nuance to the Polarization-Index (PI) calculation that I reckon a lot of people will miss and lead to major classification errors.

For example, when reviewing ride data in intervals.icu (which, however, I do love using), Zone 1 (for the 3-zone model) is classified as Coggan-Zones 1 and 2.
This skews the PI calculation as according to the scientific papers on PI calculations, Coggan-Zone 1 should not be included. Or, more precisely, “Zone 1 incorporates low-intensity exercise greater than or equal to 50% of maximal oxygen uptake”*

*ref: Frontiers | The Polarization-Index: A Simple Calculation to Distinguish Polarized From Non-polarized Training Intensity Distributions

[toyman]

OK, I just went to the paper for the formula and description, and didn’t read the details about the zones. So zone 1 is > 50% vO2max and < LT1/VT1, zones 2/3 are what I expected.

This change would reduce the amount of z1 in my calculations, which explains my higher number. Making some approximate adjustments to reduce the z1 time so that I get a 2.1 is PI, I get:

pi = math.log10(14/79.07.0100)
print pi
2.0935989844

With the formula this way, this is saying that a plan with 14% z1, 79% zone 2, and 7% z3 has a PI of about 2.1, and the paper advocating this formula says that this should be considered ‘polarized’. I don’t think this makes sense.
Using ratios instead of percent directly, I get:

pi = math.log10(.14/.79*.070*100)
print pi
0.0935989844021

A low PI as above (0.09) make much more sense to describe a plan that is majority zone 2 (when ignoring recovery).

I’m wondering if your math is right - in order to get numbers that are close to what you have in your chart, I have to use percents, rather than ratios in the formula. I think that ratios should be used, and that using percents results in nonsensical PI values like the one above.
It makes no sense for the paper to be describing ‘polarized’ TID having PIs of 2 or greater, and also
says

For this reason, we would like to present an elaborated concept of our previously published polarization-index (PI) (Treff et al., 2017), which is based on the assumption of two necessary conditions for a polarized TID. (i) a polarized structure, where Zone 1 > Zone 3 and Zone 3 > Zone 2 (and consequently Zone 1 > Zone 2) and (ii) a relatively small proportion of Zone 2. The PI aims to distinguish between polarized and non-polarized TID and to quantify the level of a polarized TID.

I don’t see how the paper can say this (z3 > z2), then spit out a PI of > 2 for a plan that is 79% zone 2 and 7% zone 3.
There may be something else I have overlooked, and a worked out example for this plan and how you got the numbers in your chart would be very helpful to clarify this.