Polarised training: my personal experience 4

I was already 3 weeks into a polarised training programme last November when I posted a blog article entitled Threshold or Polarised: why should you care? This was much too soon to draw any personal conclusions, but I promised at the time I would post a new blog with the results a few months later. So here it is!

The bottom line

There’s no question that Polarised training works very well. Plenty of research backs this up, demonstrating higher levels of adaptation than Threshold training. I summarised this research in  my first article and you can find much more through an internet search.

The more interesting question is: should you be training this way? I decided to try it myself and share here my experience and results. The conclusion? It seems to work for me. I have learned from the first block and I am now in a second block of polarised training that will hopefully result in significant gains.

However, it isn’t right for everybody. Read on for the full story and some caveats.

What is polarised training?

First a quick reminder. Polarised training is a method that is followed by the majority of professional endurance athletes. It consists of training mostly at low-moderate intensity and periodically at high intensity, with very little in between.

Using Coggan’s seven zone system and FTP (Functional Threshold Power) as the reference points, this means the pros train mostly in Zones 1 and 2, very little in Zone 3 and lower Zone 4, and sometimes above FTP (upper Zone 4 and higher).

This contrasts sharply with the Threshold training followed by most amateur riders, which is targeted mostly at medium-hard intensities and especially the so-called “sweet-spot”. This is upper Zone 3, lower Zone 4, exactly the intensity level that is avoided by the professionals during their training.

The chart to the right contrasts the two approaches, showing a typical split of time-in-zone over a 4-6 week block of training, in blue for threshold training and in red for polarised training.


There is a substantial body of research, both descriptive and experimental, that supports the claim that polarised training is more effective than traditional threshold and sweet-spot training, even at training volumes as low as 6-7 hours per week.

The theory behind it is simple: you get all the adaptations you need to develop aerobic capacity by riding at low intensity, and to develop anaerobic capacity you need to train at high intensity. Riding at medium intensity provides no additional benefits and has the significant disadvantage of increasing fatigue at a faster rate, with the result that the high intensity sessions are not hard enough.

Read my previous blog for more on this, including the references.

My personal experience

I followed a polarised programme from just after my final event of the 2017 season, the Haute Route Ventoux, until February 8, 2018, a total of 4 months.

During these 4 months I trained 78 times for a total of 152 hours, averaging 11.7 hours per week. The longest week was 13h37’ and the shortest 4h29’. Training included 29 road rides, 10 sessions at the velodrome and 39 turbo sessions. Throughout the entire period I avoided cycling in Zones 3 and 4 (on the Coggan 7-zone system) as much as possible. On the road and at the velodrome I tried to maintain a steady endurance pace towards the upper end of Zone 2, whereas the turbo sessions consisted of structured intervals in Zones 5, 6 or 7 with rest periods in Zone 1.

My actual times-in-zone during the 10 months prior to starting the polarised programme and during the 4-month polarised period were as follows:

Zone Non-polarised
01/01/17 – 09/10/17
Polarised period
10/10/17 – 08/02/18
Net change
1 44% 50% +14%
2 22% 29% +32%
3 17% 12% -29%
4 11% 3% -73%
5 4% 4%
6 2% 2%
7 <1% <1%


I did a 5 minute maximum power test on January 24, setting a new record of 341W. The previous record was 331W, set on September 15, 2017, a few days after completing three 7-day Haute Route totalling approximately 2,500km and 60,000m of climbing over a 4-week period. I was coaching others, and therefore riding mostly in Zones 1 and 2. The super-compensation produced by this amount of riding should be maximal, so the +3% improvement seen 3 months later seems quite significant. One would normally expect a decline in peak power through the winter.

For comparison, my 5 minute maximum in January last year was 307W, so the year-on-year improvement is +38W (12%).

I then did a 20 minute power test the following day, January 25, and also set a new record, at 291W. This was nevertheless only a marginal improvement over my previous 20’ best (287W, set on February 20, 2017).

Since then I have set personal bests at 1’30” and 3 minutes.


The first and obvious statement to make is that these results are anecdotal and in no way scientific, since they concern me alone and there was no way to set up a control. Clearly the approach was not detrimental to my current performance, but was it superior to my previous approach of riding as much as possible in the sweet-spot? Unfortunately, there is no way to tell.

However, I did learn a few things along the way that may be worth sharing with you.

Was my training intense enough?

I did not achieve the target of 10% high intensity (defined as Zones 5 and above), still less the 20% achieved by the cyclists in the study by Neal et al mentioned in my previous blog post. Over the four month period I spent just 6% of my total training time at high intensity, although roughly half the sessions included high-intensity intervals.

These were typical of my interval sessions:

Zone Typical session design for 60’ Time at high intensity % time at high intensity
5 14’ warm-up then 4 x [4’Z5 – 5’Z1], 10’ cool down 16 mins 27%
6 16’ warm-up then 2 x 6 x [1’Z6 – 1’Z1]-10’ then 10’ cool down 12 mins 20%
7 20’ warm-up then 3 x [20”Z7 – 5’Z1] and 3 x [10”Z7 – 5’Z1] then 8’ cool down 1 min 30″ 2.5%

As a comparison, the interval sessions followed by the cyclists in Neal’s study were 6 x [4’Z5 – 2’Z1], with 15-20’ of warm-up and cool-down. Taking 15’ this means their time at high intensity was 24’ out of a total 51’ and thus the percentage figure was 47%. This reduced to 20% over the complete week when the Z1 sessions were added to the calculation: overall they did 3 high intensity sessions (cumulating 72’ at high intensity and 96’ at low intensity) and an additional 3.5-4 hours at low intensity to reach the 20/80 balance.

Averaging almost 12 hours a week, I was training for nearly twice as long as those in the study. The additional time, however, was mostly at low intensity, so overall my training lacked sufficient work at high intensity. Being 20 years older may have something to do with this!

Furthermore, I made things hard for myself by progressively increasing the intensity while keeping the interval duration constant. As we will see below, it might have been better to do the opposite, i.e. increase the duration or the number of intervals but keep the intensity constant.

To summarise, the improvements noted were obtained after substantially reducing the time spent in Zones 3 and 4, while making no significant change to the time spent in Zones 5 and above. It should have been possible to spend more time at the higher intensities.

Practical issues

The polarised training method suits people who are highly motivated and mostly train alone. If you rely upon group rides for your motivation and don’t like doing intervals, it is not for you!

Following a polarised programme requires discipline and sticking to the plan. If you are self-coached, it is important to spend the time to understand what you are doing; to design an appropriate programme and to measure, monitor and adapt as time passes. Depending on how you value your time, investing in a coach may make good sense.

There’s no such thing as a guaranteed response to training: every individual is different. There were substantial individual variations in the study mentioned above, with some showing almost no improvement (with either polarised or threshold training). The results for the 12 cyclists in a 40km time-trial were an average gain of 2.3 minutes, with a maximum of 4.5 minutes, and a minimum of 0.1 minutes. An experienced coach should be able to identify the reasons for the lack of response.

80% “easy”

As Stephen Seiler has pointed out on numerous occasions, easy doesn’t mean brainless, nor is it literally easy! The term is relative. To be effective, all training must be purposeful: discipline is critical. Each ride should have a purpose: to focus on technique, on cadence, on position, on breathing, or on some other aspect of your riding. The pace should be mid-Zone 2: any higher and you will find yourself frequently crossing the first lactate threshold. Since this is painless you won’t notice, but the fatigue will build.

20% “hard”

In this context, hard means “very hard”. These sessions should hurt, a lot. A key goal is pushing out your ability to tolerate pain and to keep going when your body is yelling at you to stop. The more you do this, the more you can tolerate. As you progress, the emphasis should be on making the intervals longer and doing more of them, rather than increasing the intensity. According to Seiler’s studies of Olympic athletes, the goal is ultimately to do 40, 50 even 60 minutes of work above FTP. To make this achievable the intensity needs to be in the range 105-110% of FTP, and the interval length between 4 and 10 minutes.
A good starting point might be 4 x 4’ at 105% of FTP with 2’ rest, stretching this out over a period of weeks to 10 x 4’ (with 2’ rest) or to 4 x 10’ (with 5’ rest).

Preparing for a race

The obvious gap in a polarised training programme is time spent at race pace honing reactions and bike-handling skills. The professionals have built these skills over many years and use training camps and minor races to sharpen up. This may work for some amateurs but not for all of us. For many amateur cyclists it makes sense to adjust the plan to allow for joining some fast-paced group rides in the month before the first race, for example replacing one of the endurance rides by the group ride and one of the interval sessions by a recovery ride.

What next for me?

I have now begun a second cycle of polarised training, with the intent to increase the percentage time spent at high intensity from 6% to at least 10%.

In practical terms this means I will do fewer Zone 6 interval sessions and more in Zone 5, and I’ll introduce some longer intervals at just above FTP (e.g. 3 x 10’ in upper Zone 4).

In parallel, my endurance rides should become significantly longer, stretching out to 4 and ultimately 6 hours.

Verdict in 8 weeks!

Leave a comment

Your email address will not be published. Required fields are marked *

To reduce spam, please answer this simple maths question: * Time limit is exhausted. Please reload CAPTCHA.

4 thoughts on “Polarised training: my personal experience

  • Ben

    Great read! Thank you! I have been reading and researching about Polarized Training for a while now and have also implemented it as my training plan. What seems to be confusing and conflicting between papers and reviews of the polarized training plan is the 20/80% training rule. Some say 20% of your training sessions should be in Polarized Z3 and 80% of your training sessions in Polarized Z1. Others say 20% of your overall work time should be 20% L3 and 80% L1. The two are very different, but which is correct as people seem to be interpreting the study in two different ways?

    I have been using the model as followed:
    Day: 1 – 1hr30 with 5x8min L3
    2 – 2hr L1 with max 5sec sprint every 20min (high end of Zone 2 power)
    3 – 3hr L1 with max 5sec sprint every 20min (Mid Zone 2 power)
    4 – 1hr30 with 3sets of 8x 40/20s (40sec is max effort)
    5 – 2hr L1 with max 5sec sprint every 20min
    6 – Day off or super easy 1hr spin
    And then repeat
    I may change the L3 workouts now and again though to save too much repetition. If an L1 session needs to be shorter than 2hr then i may tap into my lower Zone 3 power, as using the L1/2/3 polarized model chart allows upper L1 to go into your lower Zone 3 power.
    My sprint power is up by 214w, 5min is up by 47w, 8min up by 30w and my 20min by 18w. After 4 months. And on the plus side, my weight is down by 2kg!! Giving me 4.7w/kg @8min and 3.91w/kg @20min. That is a fairly nice return on averaging 11-12hrs a week.

    • Marvin Faure Post author

      Those are great results Ben, well done. It shows that your plan is working. You will find that after a while you will need to change the plan to provide a different stimulus in order to continue your progression.
      To answer the question on how to calculate the 80/20 split: this is not a magic number, but has been shown to be empirically true for many elite athletes. I have also seen 90/10. In practice, some athletes respond more to low volume/high intensity, whereas for others it is the opposite. This could explain the difference. More on this from Alan Couzens: https://alancouzens.com/blog/vol_int_responder.html

  • Jürgen Kerstna

    what is not clear to me is how to count recovery time between reps. I have searched and some sources (can’t recall which ones) say these should be included in high intensity time. So that 6 x [4’Z5 – 2’Z1] high intensity time is 24′ not 16′
    What are your thoughts?

    • Marvin Faure Post author

      Hi Jürgen,
      I have always taken the warm-up/cool-down and rest periods to be excluded from the high intensity time. This is in line with what I read in Seiler and Tonnessen’s paper “Intensity and Duration in Endurance Training” and makes sense to me. Recovery time is variable and not standardised in intensity so it makes little sense to include it. However, I’m not sure if researchers are always consistent on this point and they do not always clearly define it in their papers, so there may be exceptions.