Another analysis episode based on questions from the forum / YouTube / Spotify!
This time we looked into a tick-by-tick time in zone analysis of the same cohort from last week’s episode (140k rides from 3,600 athletes). We looked at which athletes achieved 3, 4, or 5w/kg as a recent PR and then looked at power for every second of training from the prior 12 weeks to see what the training intensity distribution is for these athletes.
It’s a uniquely large sample size and has athletes following TrainerRoad plans and athletes not following TrainerRoad plans, so it gives a really practical look into what athletes are actually doing to reach these fitness milestones.
We also compared what structured vs. unstructured training looked like from a TiZ perspective, and looked at gender differences (or the lack thereof, in this case).
Then we looked into erg mode vs. resistance mode, a broad but informative look into training habits and durability (much more to come here), and finished off with a quick recap of Stetina’s Paydirt and my Leadville 100 prep.
Hope you enjoy, and if you want some of these analyses in more bite-sized chunks, follow us on instagram!
// TOPICS COVERED
(00:00) Welcome!
(02:44) Training Time Distribution Across Power Zones
(06:10) Comparing Structured vs Unstructured Workouts
(09:03) The Impact of Training Volume on Intensity
(11:49) Erg Mode vs Resistance Mode in Training
(15:05) Exploring Durability in Cycling Training
(26:56) Balancing Intensity and Endurance in Training
(28:50) Understanding Durability and Power Output
(38:09) Training for Leadville: Strategies and Insights
(50:34) Preparing for Elevation: Nutrition and Equipment Choices
While I understand the desire to separate elites/pros from average cyclists, I feel like having the option to do elite/pro training would be a nice addition to training plan volume and weekly workouts.
Currently I am stuck at 5.2w/kg doing 12-14 hour weeks, despite not being time crunched. Having the option to have a training plan that is tailored for people on the far right end of the bell curve would be nice so I don’t have to go rogue and change duration of the workouts or add more endurance before and after planned rides (yes I have dynamic endurance maxed out on my plan).
A trend I have been noticing in TrainerRoad feels almost like unintentional gatekeeping to protect us from ourselves, like a concerned parent, “here is what the pros do, but you shouldn’t do that because you are frail mortal”, despite having success with 15-20 hour training weeks, doing multi-day events that are 4+ hours per ride, and weekend rides with 10,000+ft of gain in 5-6 hours
P.S. the 3000kj durability test will forever be my favorite, and most painful, method of testing durability
This was a great podcast and a nice change from always having the pros talk about their training. Jonathan mentioned this data and his interpretation is at a high level but would be great if they took a deeper dive into each data set. Maybe invite a researcher or sports scientist to the podcast to provide their interpretation and hypothesis.
Love the durability stuff. Feature request - give us the ability to see those metric individually. Basically, take the current current power curve for NP and give us the option to filter it for “after x tss”. Would love to be able to see how that is building over the season and compare to prior years just like we can with AP and NP now.
@Jonathan - on tire choice at Leadville - Totally agree on not assuming the published test results are as dramatic as some tests show. There are just a lot of variables to control and tough to replicate conditions day to day. I think those tests have value and are a good starting point, but I’m a “trust, but verify” kind of guy. I dug into using the chung method a bit earlier this year and found that it lines up pretty well with roll down testing (both approaches showed consistent winners when comparing 2 tires). For me, the trickiest part was finding low/no wind conditions to make the comparisons valid. I don’t put much stock into the accuracy of the specific CRR numbers I was getting from the test, but I do believe the test gives a good idea about the spread of CRR/watts between 2 tires that are tested back to back. And for what it’s worth, the testing from John Karrash was within a watt of what I can up with for the tires I was comparing (but the spread was only ~4 watts, so 1w was a significant percentage of the difference). My testing was limited to road testing (for gravel tires) and that probably helped with controlling the variables. Anyways, maybe worth spending a little time on a few tests if you have the time. I have zero experience with the Aspen ST’s, but I know some of the tests show them pretty slow on pavement/smooth surfaces. Maybe doesn’t matter if they give you a bunch of confidence/speed elsewhere, but I know you are looking at every angle.
I used Golden Chetah for my comparison and it’s definitely a bit clunky, but it’s not rocket science either. It is important to get a speed sensor for the wheel, GPS isn’t accurate enough. There is some stuff out on youtube and other places describing the process, but I never found a great resource that ties it all together. You basically just load the rides in golden cheetah and use aero lab play with the CRR and CDA values to get the virtual elevation to align. If you set the CDA the same for both tires, the different between the 2 becomes the CRR. That assumes the CDA is identical between tires (not a perfect assumption), but the net is that any small differences in CDA basically end up in the CRR spread (not technically correct, but still tells you which tire is faster). Here’s a video that gives a taste of how it works.
I ran dub 2.4’s at Leadville last year and I’m considering running a thunderburt 2.35 rear this year. I haven’t tested either of those tires, but plan to do some A/B testing after Unbound. And I have a trip to Leadville in mid June and will see if the handling/traction is a concern (maybe even try a TB up front, but probably not). I’m not very fast on the MTB (not pushing the limits of my tires) and not contending for a podium, but I’ll still take a faster rolling tire if it doesn’t hurt my confidence. I don’t really worry about flats at leadville (knock on wood), the only 2 flats I’ve had in thousands of miles of riding that course were pinch flats on powerline banging a rim (total user error).
Very interesting. My riding buddy and are within about 5 lbs of each other. We both ride epic 8’s and are riding them for Cascade gravel (80 miles) and high Cascade 100.
We do our weekend long rides together, generally 3.5 > 5.5 hrs of gravel on the gravel bikes, or single track with MTB’s.
Nearly every one of our rides are within 5-10 NP of each other. He just installed 2.35 TB’s and I just installed 2.4 Dub’s. While not the most scientific measure, I’m curious if our numbers will remain so close. Obviously a ton of variables, but running the tires you just mentioned.
hey @Jonathan is there a way we can read these pages you displayed in the analysis ourselves? ive tried looking but cant find it, i was only half paying attention (wfh day) when you were going through it and would love to look at it more closely myself.
Loving these last couple data mining podcasts. Looking forward to being able to read through the write-ups (without resorting to screen grabs) when you get them posted.
Not arguing but how far you want to push your w/kg?. I would not say that 5.2 w/kg is stuck that’s furthest that 1% of the population can get. This is elite stats. I am not familiar with ftps of professional riders, but that should be close.
great question! I don’t actually have an improvement goal in mind (although 6wkg would be nice ), but sometimes I wonder if having my training condensed into 12-14 hours hinders the ability to get above where I am currently at. Truth be told, I feel like the amount of intensity I do is at it’s peak so I would just be adding endurance. My reasoning is more of a “what if” scenario, than a definite “x more hours will make raise my FTP”
As you probably know, the only way to figure out what (if any) benefits you get from more hours/training is to do the work and see where it leads. Everyone has different physiology. But it’s very doubtful you are super close to your genetic potential at under 15 hours per week. If you did a season or 2 at 20+ hours/wk, you would likely see benefits IMO. Maybe dramatic, maybe slight. It’s always going to be diminishing returns as you add volume, every bit of additional fitness gets exponentially harder to gain.
My n=1 as a masters athlete who has gone from 10-15 up to 20+ hours/wk the last few years - My FTP has barely moved, but my durability has gone up significantly and my race performance/results are dramatically improved as well. I’m sitting at ~4.25w/kg right now, probably highest it’s been, but mostly due to dropping some extra weight this year (ie - no FTP increase over last year). As a 57 year old, I think I’m probably near my aerobic ceiling and doing well to hold onto my FTP and vo2max as my genetic potential drops every year. But my power pr’s on long efforts are still getting better every year, so I’m able to sustain longer into races. If I was doing crits or short XC races, I’m not sure the extra volume moves the needle that much, but my results even in “shortish” 3 hour races have improved somewhat. And for anything 5+ hours, the durability is a huge difference maker and I’m beating guys with much higher FTP’s who tend to fade late in races. Bottom line for me - I’m retired and enjoy training, so the volume is just me doing what I like to do during the race season. The race results are a bonus.
This brings up a question. It seems like "time crunched’ is the default. What about “got all of the time in the world”? How would TR structure the training?
Has there been a problem uploading this episode to Spotify? It isn’t showing in my feed, even after logging out and back in. It is online in Apple Podcasts though… very odd.
Would be interested to see absolute weekly time spent in high intensity by FTP, rather than percentage. I suspect it’s probably more similar, with the higher FTP riders just doing more low intensity volume, decreasing the percentage.
Curious how the use of predicted FTP impacts this analysis. AIFTP currently thinks my FTP is ~280w (which would mean I did an hour straight at 0.93 yesterday after ~2000kJ… lol) but this is in the context of being in base period and having no recent high end efforts. This means most of my ride yesterday would fall into “z2” in the three model system, when in reality it was z1 in the three model system. You get too many riders in this situation and it inflates the time spent in moderate intensity and makes it look like everyone is riding more pyramidal than they really are. You could do a sensitivity analysis by redoing the zone distribution by “active” FTP (including manually entered ones) and seeing how the distribution changes (if at all).
That’s a very high percentage of time spent in active recovery or coasting. In my own data it’s closer to 5-10%. BUT… I live somewhere where it’s extremely flat and windy. There’s no opportunity to coast or soft pedal. And that’s interesting to me because the impression of local riders who had gone pro or are involved with development/coaching was that prairie riders tended to be stronger than average, which they considered unexpected due to the lack of hills. I wonder if some of that “paradox” is because the flat terrain cuts down on time spent coasting and recovery, meaning an extra 20-30% of your training hours are spent in productive training zones.
I love these types of analysis looking at yur large data set. This is another reason why in the past I asked @Nate_Pearson to add the ability for the app to record r-r timing from HR straps used for HRV metrics. Is DFA Alpha 1 or other HRV metrics useful? If you have that information to look at you could more easily check.
The app could allow pairing other sensors too like temp (Core sensor), Muscle Oxygen (Moxy, train red) so that you’d have the data to look at. Seems like having more data could help figure out how to best use that info by looking at trends in the data. This also goes for other data pople have that could be imported like garmin health API. For example body battery, I’m not saying body battery data is always good, but maybe for some people it trends with what you expect and so for those people that data could be used to plan better guess how much the person could handle.
As to the erg mode vs resistance mode difference I wonder if the data set is large enough to break up the resistance mode group based on how well they improve with training. Are there subsets that improve more than expected and some that improve less? If so how do they differ in the metrics you compared the erg vs resistance groups. i.e. is there a subgroup that is able to do the workouts correctly and those people also improve more than expected? So could maybe show that while resistance mode increases the chance of people not following the workouts correctly who can correctly execute do better than erg mode.
I’m guessing people with proprioceptive impairments (common for those on the autism spectrum) would always do better with erg mode but those with great proprioception could possibly do better without erg mode which could explain why some coaches (see emprical cycling podcast) seem to think resistance mode is better than erg, because it is for a subgroup. The way they coach high end individuals leads to selection bias towards those who can execute workouts correctly and are in that subgroup.
When Nate talked about adding r-r timing recording back in 2022:
Did a ride together Sunday. 40 miles, 5000’ ascent of mostly gravel. 3.5 hrs. He’s around 5-7 lbs heavier. Both on Epic 8’s (mine a little lighter being s-works). He was on 2.35 TB’s and me on 2.4 Dub’s. My average power was 184 & his 192. Pretty standard for the large volume of rides we’ve done together. Nearly always within 5-10 watts (usually closer to 5).
At some point we’ll do a similar ride with him on his diverge with 55 schwalbes and me on the Epic 8. Now that we know the standard deltas, I’m incredibly curious how our power differences will be on the rides with the gravel va mtb. I have inner bar ends and also run a 90mm -6deg stem so my fit is pretty comparable to my gravel bike.
Have to say I love my beautiful mosaic gravel bike, but the 50R and 2.1 TB up front are feeling more limiting on our local very-rough gravel. When we do our comparison it’s going to be hard to justify as many miles on the Mosaic if the power comparison on the terrain is similar, particularly given the comfort and convenience of the full sus/flight attendant. I’m using the Epic for next weekends “gravel race” we pre rode and expecting it to be a cheat code..
), but sometimes I wonder if having my training condensed into 12-14 hours hinders the ability to get above where I am currently at. Truth be told, I feel like the amount of intensity I do is at it’s peak so I would just be adding endurance. My reasoning is more of a “what if” scenario, than a definite “x more hours will make raise my FTP”
Time to see if you can get on a development team and shoot your shot.