Polarized Training with Stephen Seiler, PhD | EP#177 That Triathlon Show

I also think conversely, most people are overdoing it with a mid zone plan if their ftp is not 1 hour power. Sweet spot should not cause any major failures, at least not with the durations of the intervals in most plans.


Yes, it is a common misconceptions. Of course, hard workouts should be hard, even very hard, but I don’t think they should be so hard as to risk workout failure, and for the majority of hard workouts, I want to feel as though I have a little left in the tank. I almost never want to race a workout and “test” or “prove” myself in workouts. That’s what races are for. I use a 10-point RPE-scale (as in training peaks) to rate session-RPE after each and every workout, and all my hard workouts end up being in the 8-9 range.

I used to do a lot of swimming and track workouts with a group last year, and in those workouts you’re really racing them. It was great for short-term improvements, terrible for long-term improvements and sustainability, and mentally very taxing as well.

I love this spreadsheet @macneese.chad, you did a fantastic job with it! I’ve been skimming through your POL experience thread but it’s a bit overwhelming. Do you mind giving a brief overview of what you did/are doing and the results?

Like @bbarrera and @mcneese.chad said, that’s not quite correct. The idea isn’t to be right at that LT1 mark all the time, and especially so if you’re just estimating where it is anyway (and might easily make a mistake in that estimation). I also agree that using HR as a guide that is just as or more important than power is the way to go for easy rides. And using HR and power caps rather than targets.

This to me is a printable! One of the most important points of the interview.

This sounds like you’re definitely going too hard for easy running. I have heard no conservative tone at all in the 75% estimate, and it matches well with the data that I have collected as well (a fraction of what he has of course). So I would be very confident in recommending 75% as a cap (!!) for your easy runs. Which is an important distinction from your average being around 75%.

It doesn’t ruin any effects of that particular workout, but it causes a larger stress response and longer recovery times. We talk about a study that found that training in Z2 (3-zone model) causes a similar stress response as training in Z3.

An important point is, if you estimate that your LT1 is at say 130 bpm. Then why not be conservative and stay at 120-125 bpm. You’re getting the same training effect without risking that you get into the grey zone for long parts of the workout. It doesn’t matter if this happens once or twice of course, but when it happens tens or hundreds of times over the course of a training block or a year, that’s when you’re probably leaving improvements on the table.

And one more thing - it’s ok to go super easy too. I personally often do my easy rides at 65% or even lower of HRmax. Because I also let RPE guide the workout, and if it says easy, I want it to feel easy. Now, I don’t train like a typical age-grouper, as I train around 20h per week (triathlon), so I carry much more fatigue into any workout than most. So I’m not saying that <65% is any better than 70% or 75% of HRmax. I’m just saying that when feeling tired, keep the easy days truly easy. There’s nothing magical about 75% and nobody is saying that all easy training should be don at that intensity. It’s just a cap, and low-intensity spans a broad spectrum of power/HR and even RPE.


Thanks Mikael, that makes a lot of sense. Can I ask about a phrase that was repeated in the podcast, Seiler kept talking about ‘collecting minutes’ when doing the low intensity stuff. Does that mean that doing 2 by 2 hour rides a week is the same as doing 1 by 4 hours? We should just roll around collecting as much as possible?

Excellent episode, Mikael. I was glad to hear the question/response surrounding placing a hard ride and hard run on the same day to protect/ensure the easy days stay easy. Like you, I’m also around 20 hours/wk and I’ve begun using this strategy over the past 4 months as well, not for a physiological reason, but psychological. It’s significantly easier for me to wrap my head around doing two hard workouts on a Tuesday and doing (polarized) Z1 on all Wednesday workouts versus riding hard on Tuesday, running hard on Wednesday, then riding hard on Thursday, etc. Thanks!

1 Like

Agree. I think high end of TR sweet spot is in the lower end of Seiler zone 3. This is also consistent with what I recall from another Seiler podcast where one of the interval formats he recommends is a 4x16. This is a sweet spot workout in TR.


for intensity his 2013 study showed 4x8 intervals resulted in more improvements than 4x4 (it also included the 4x16 mentioned above). The athletes in that study were instructed to ride at their maximum for each interval, so there wasn’t a prescribed power. The 4x8 resulted in twice improvement of vo2max and power at 4 mmol/L lactate.

As participants were instructed to do max efforts (no Erg, no power target), these are the average % HRmax recorded for each interval type:

  • 88% HRmax average for 4x16
  • 90% HRmax average for 4x8
  • 94% HRmax average for 4x4

which for me personally, maps 4x16 to TR Sweet Spot, and 4x8 to TR’s threshold, and 4x4 to vo2max (30/30s, and 1-3 min). If that study holds true for me, that means focusing on doing threshold workouts. And to do quality threshold, I need a good base of sweet spot. My personal interpretation is: 1) establish a base of sweet spot, and then 2) build with threshold intervals.


Yeh, this has also been a sort of suspicion of mine:

Basically, most people on here (me included) probably have FTPs that are overestimated compared to what our true MLSS/FTP1hr would be. Let’s say from 3-7% overestimated.

The threshold workouts put us squarely in the 90% of peak range he’s talking about for 8min intervals (for me cycling peakHR is 177 and my 8min @ 108% goes from about 158bpm rising to around 162-3 for the last interval at the end). Sweetspot work would nail us right around the 16min interval areas he was looking at. Trainerroad gives typically more generous rest intervals than he recommends (but usually makes us do more intervals as well).

So I think beyond this whole mess with base, build and specialty, and maybe with the exception of the high volume SS plan (which is probably too “monotonous”), Seiler’s research probably doesn’t have too much beef with typical TR prescriptions. He might prefer the old TR plans (with the long weekend rides) to the new ones, but whatever.

My take home from all of this and what I have personally noticed is that the TSS that I get from a long ride when I do have the time is much easier to recover from than similar TSS that I’ve done at sweet spot. So, when I have the time, I’m going in for the long ride as opposed to relying on time crunch short cuts.


Agree, I’d say one of the ideas he is promoting is that people need to do more low intensity. Before TR my outdoor workouts would go to tempo for rest intervals, and I’m sure that was reducing my progress (I was slowly getting faster). After TR my rest intervals both inside and outside became actual rest intervals (zone 1 of 3 zones), and my long outside rides are now split between intensity (sweet spot / threshold), or around the borderline of zone 1/2.

Personally I feel TR and smart trainer did more to force me to slow down, and TR seems consistent with Seiler’s research now that on this podcast (versus Fast Talk podcast) Seiler is saying cycling is more pyramidal and includes a fair bit of zone 2.


My takeaway is that you could do a number of workouts that are made up of intervals that get you into that 90% of max HR zone. I would say that whether it’s 30/30’s, 2 min, 5 min, 8 min depends on whether or not you’re looking for specificity to an event, climb, type of race, etc.

For example, let’s say you’re targeting a gravel race that has a couple hills that range from 5-7 min in duration, you would likely do intervals around that length at an intensity that keeps you in that 90% of HR max.

If you’re a cyclocrosser you might focus on shorter intervals—that still get you into 90% of HR max—that have less rest between them to build capacity.

I also think that variety is your friend, mix it up.


Not exactly the same, but he definitely would say (as would I) that overall training volume is the number one most important thing for endurance athletes. Whether 2 x 2h is preferrable over 1 x 4h would depend on the rest of the training of the athlete as well as their goals. For more beginner athletes I think frequency trumps duration of individual workouts, whereas for a more advanced athlete it might be preferable to do the 4h ride. It’s impossible to say that one is better than the other as it is context dependent, but for example, I think Seiler gave the example of a 5h per week athlete and how to distribute those hours (or it may have been on another podcast), and a good way to do it would be 1 x 2.5h, 2 x 1h, 1 x 0.5h (those aren’t Seiler’s exact numbers, but something along those lines just to exemplify).

To give a personal example: I typically distribute my low-intensity bike workouts as 1 x 4-5 h (Z2 in a 5-zone model), 1 x 1.5 h (Z2), 2 x 1h (Z1). So I see the long ride in Z2 as a key workout, even though it’s low intensity. How I then distribute the rest of the low intensity isn’t that important, but I know that doing everything in Z2 (5-zone) will be too much for me if I want to perform in the hard workouts, and I am a believer in frequency also for advanced athletes, so I choose to distribute it over 3 workouts, where two of them are pure recovery.

Totally agree! Having a hard workout to “look forward to” every single day is a good way to end up with exercise-induced anxiety. Or at the very least, inducing a big “mental load” every day, and that system too for most of us needs regular recovery.


Thanks Mikael, that’s clear. The guys above are saying now that 4x8 in Seiler’s model is pretty much a 4x8 threshold workout. Is that your understanding too? Or is it supposed to be something harder than threshold? Like the TT sort of threshold/VO2 borderline effort?

@Shrike I would say it’s at or ever so slightly above TrainerRoad FTP, but physiologically it’s definitely a bit harder than Maximum Lactate Steady State. Between MLSS and VO2max. I would say the intensity is slightly lower than TrainerRoad Threshold/VO2 though, because at least I couldn’t do those TR workouts without the generous recoveries, and these 4 x 8’s were done on 2-minute recoveries if I recall correctly.

1 Like

Here’s a close example from the TR library.

Washington. 4x7 at 105% FTP. 3 min test intervals.


1 Like

Weekly sessions from the 2013 “adaptations-to-aerobic-interval-training_interactive-effects-of-exercise-intensity-and-total-work-duration” study:

With participants instructed to do perform maximal sustainable intensity (“isoeffort”) using their own bike on Computrainer Lab trainers.

1 Like
  • I’m happy you like the sheet. It’s something I continue to refine and am open to comments (and corrections) if you see anything you think could be improved.

  • As to my POL experiment, I think the basic approach can work well for me. I finished the 4-week test with a good overall feel and no excessive fatigue. It seemed like less tired legs than a typical TR SSB 2 or Build for comparison.
  • I wish that I would have had the time to apply a 6-week test, then a break, and a 6-week TR Build for comparison. But I was short on time and then got sick for two weeks shortly after completing the POL period. So I don’t have any great comparison that was as closely times as I would have liked.
  • But as I am currently in the 6th week of my Sustained Power Build, I can see some differences in fatigue. The first 3 weeks I followed the main plan and did the SS Sunday workouts. I think I ended up with a decent dose of fatigue with that approach.
  • I swapped to the longer Endurance pace workout at the end of week 5, in hopes of getting some relief that is more like the POL approach. I think it likely helped me recover for this current week. I hope to do the long trainer option again this week and next.
  • I think I got good, but manageable stress from those 4+ hour trainer rides. They lead to some notable DOMS that I haven’t seen in other shorter, higher intensity rides for the most part. I have had some (like Le Conte) that crushed me and I was sore for days. It’s interesting to me that I had similar muscle soreness from “easy” rides on the trainer that stretched well past 3 hours.
  • On those long trainer rides, I made every attempt to do the entire workout with little or no coasting. I nailed most with less than 10 seconds of coasting in most, and a few with no coasting. The mental strain got tough on one or two, but it was a different strain than a over-threshold workout like VO2 Max ones.

  • One very notable difference in my training right now is that I am including run training 3 times a week. That was absent during my POL test last October. I am working to get better run results in my duathlon’s this season. I am certain that the extra time on the run is adding to my overall fatigue. I eased into it with 2 short bricks per week for about a month. At 8 weeks into my run training, I am up to 3 runs (2x 20 min bricks at easy pace, and 1 solo LSD run for about an hour) that are aimed at my Sprint distance Du’s.

  • Back to POL, I think the basic concept with my 7-9 hours per week allow for good use of the POL approach. I have taken the positive feel of the extra long easy days and want to do those at least twice per month for my Sunday rides. More if I can afford the extra time.
  • I still like the option of the shorter Sweet Spot option to meet life’s other commitments. They still pack a decent punch, but ideally I hope to do a mix with the LSD ones. Maybe alternate weeks in the best case.

Let me know if I didn’t answer your questions, as I rambled a bit above. Happy to answer any other questions.


Those 2 minute recoveries still kill me. Maybe I’m soft, but I couldn’t come close to resting that short and still getting up to the power I estimated was needed for each set.

However, that was also before I got more familiar with the real goals of VO2 max and the adjustments I can make to get through the sets and get the right demands from my body. I should probably retry that approach again at some point, but a 1:1 work/rest seems more to my liking :stuck_out_tongue:

1 Like

Yes. Although Hour of Power is replaced by 40km Time Trial data. In my case the 40km results produce a CP60 of 235 watts +/- depending on race course and conditions. But over years and many races that number holds up.

FTP testing (20min, 8 min or RAMP) are all fairly consistent too and produce an FTP estimate of 240-250 watts.

We’re talking almost a decade of tests and races so pretty good consistency over time and accounting for fitness and training consistency across seasons.

MLSS testing shows about 230-235w which is probably not accidentally consistent with 40km wattage.

In terms of TR programs, if I use the 240-250 watt estimates I can complete SweetSpot Build I and II as programmed. The longer workouts hurt, and they should, but are doable.

For Sustained Power Build, I had to change from 3:1 to 2:1 work week to rest week to complete the plan. There were also several workouts that I could not complete. Had I used a 230w or 235w FTP suspect that 2:1 would have been 100% success and probably OK at 3:1

Since I’ve been training for TTs over the last many years, when I try short power build or other heavy VO2max type workouts I really suffer. That’s specificity in training revealing itself.

Hope that addresses the question. I’m only one data set so others might have a completely different experience to report.



Yes, same here as a true recreational/enthusiast cyclist. Keep in mind those were self-selected maximal 8-min efforts. They were not done in erg mode. Would be interesting to go out and try pacing outside, I suspect my power would drop on 2nd, 3rd, and 4th intervals with only 2-min recoveries. Not sure how much, but it would drop. Forgot if the study commented on the same.

1 Like

Also of note… 90% HR max is a much easier and repeatable interval than 120% of FTP, even if that FTP is your true 1 hour power for most people. 3 minute self-selected intervals will often put me in to zone 5 of the Norwegian model or greater than 92%

So we’re probably talking about something like 5-10% over threshold at most. (or even some people’s FTP in TR)

1 Like

Kudos for that episode. 2 other takeaways:

  • According to Seiler polarized is also recommended with 5 training hours per week
  • From around 1:13 hours in (devils advocate part) Seiler says that those other approaches (like TR I assumed) would only work short term (giving the example of the most winning Norwegian Olympian cross country skier) which contradicts what Nate pointed out here

So it seems a bit words against words :exploding_head:… Looking forward to hearing your take on this.

1 Like