The “specialist” who refused to be categorized

What does it take to be successful in both cross country and marathon racing? (photo: Michal Cerveny)

By Thomas Bonne

When the Red Bull livestream flows through your favourite device with Rob Warner and Bart Brentjens during the UCI World Cup you might hear them joking about the time Annika complained to Bart about calling her a marathon specialist. Rob Warner probably didn’t know how accurate he was when he said that Annika is fast in XC but very good in marathon because she just doesn’t go slower despite the distance. That is exactly her stronghold and why some have seen her as a marathon specialist. In reality she’s never done any training to be good specifically for marathons.

Most people probably think you need to focus on one discipline/distance to become a good cyclist. And truth is, to a certain extend you must. There are exceptions to rule however, and I believe that Annika has proved herself as a rider capable of more than doing marathons. In fact – by winning the first World Cup of 2018 and the following week Cape Epic I will argue that she has proved to be one of the most versatile endurance mountainbikers. In this piece, we’ll give you an insight into how Annika’s training has been the past few years. You can then draw your own conclusions. A quick warning before we start: This post will give you no secret tips on why Annika is fast on a bike, it won’t deliver any “magic bullet” or short cuts to success. The training is very much based on plain and boring time on the bike and we’re not re-inventing the wheel (!) in any way!


Throughout the years Annika has increased her training volume quite a bit. Unfortunately the intensity distribution in her training wasn’t registered or monitored very accurately for the “early” years. Thus, for the years 2010-2014 we only have volume to present. However, looking at volume there is a gradual increase in annual hours from 448 in 2010 to approximately 800 hours in 2015, 2016 and 2017. The year with most training hours so far has been 2015. The year 2016 was a successful one and when looking at the training compared to other years one might be tempted to think that it was beneficial to reduce the annual volume. In this case it should be noted that training almost ceased after Rio due to Annika starting full time studies after year completely off studying. When comparing volume from January 1st to August 31st 2016 to the same period for 2015 and 2017 the volume tops the two latter mentioned seasons.

Figure 1. Progression in annual hours from 2010 to 2017

The past three seasons we have been monitoring training with power meters and can present data with a reasonable accurate intensity distribution. For clarification, data here is presented using a six-zone intensity distribution calculated from Critical Power (CP). As much of a fan I am of performance testing, it just doesn’t appeal that much to some athletes. Annika for sure being one of them! The concept of CP is highly applicable to an everyday training session and provides a sound physiological model. I still like to measure the VO2max of Annika from time to time but we have exchanged all lactate and VO2-testing for the CP-model. Now, we’re using a 3min and 12min mean maximal power test as proposed by Len Parker Simpson. You can find the paper for that here. The test itself is actually simple and easy to do and you can use several online sources to quickly calculate your CP and intensity zones. If interested in seeing how we do it drop a comment below.

For analysis we use the free software Golden Cheetah available for download here. If this makes absolutely no sense I recommend watching a few of the tutorials on the Golden Cheetah website. Start with this one here.

As you can see, the intensity distribution is very similar across the past three seasons. 2016 did contain a fraction more time in the upper intensity zones. This is very likely due to a higher number of races despite a slightly lower total volume. Evidently, a lot of time is spent with a low intensity. Approximately 600 hours (or ~75% of total cycling time). What this graph does not contain is non-cycling related training. Unfortunately, we don’t have these numbers for the years 2008-2014 which would have been very interesting in order to see if the increased training volume is made up of time spent primarily in Z1-Z2 or just across the entire range.

Figure 2. Annual intensity distribution from 2015 to 2017


Table 1. Annual percentage of time spent in zones

Up until 2017 Annika didn’t really devote much time to strength training of core training. She had a go at it but never with much success. She felt it left her exhausted and incapable of executing training on the bike. In 2017 we adopted a different strategy and tried implementing strength training again. This time we approached it a bit differently and added resistance/load only up to a point where it didn’t compromise the training on the bike. It has often been “two steps forward and one step back” but we’re slowly getting there.

For now the strength training session regime has been: Two primary exercises (squat and leg-press) two-three times/week in the off season and during season “ad hoc” with periods of one session per week, while other periods are too busy with racing to do strength training. Whenever a window opens up between races we pick the strength training up again.

The above has more or less been standard for Annika regardless of races or time of year the past three years. In many ways she’s a very easy rider to prescribe training for. She can go fast in most type of races as long as she’s in shape. In all honesty there’s no magic bullet here. We simply aim to be generally well prepared and it often works out in our favour. But even if you’re no Annika Langvad I will argue that adopting the same strategy if you’re racing at the same level is a good idea. For two reasons: one (1) being the reliance on the same energy systems and the other (2) being the structure of a pro riders season.


Those claiming that you cannot be fast in both XCO and XCM because it requires different training, I’d like to underline the fact that despite the first being app. 90 min and the latter somewhere between 210 and 300 min they are both relying on a highly developed aerobic energy system. As such the core (NOT body core stability) training can/should be the same. To illustrate this let us look at some numbers from actual racing. We do not have power files from all races but we do have some. Below you see files from an XCO World Cup win, an XCM World title and a 7th place finish at a cyclocross World Cup overlayed. I will argue that they constitute a good basis for comparison of world class riding across those three terrain disciplines. For a nicer visual presentation the data is a smoothed 30 sec average. Now, you can analyse the graph below in many ways (and probably make more of it by having the raw data) but I will still argue that the racing (and hence physiological demands) aren’t that different. To confirm that I’m correct I would of course need data from others riders as this could simply the way Annika races.



There are a few caveats to this comparison though. One being that all three files was collected using different brands of powermeters. We’ve put a lot of energy into using only one brand of powermeters or calibrating the different brands against each other the last years. Still there will be differences from brand to brand. Be aware of this when looking for your next powermeter. Another caveat in this comparison is the running sections in a cyclocross race. In the graph I’ve used all data points (including the non-cycling ones) with a 30 second smoothed average. It will undoubtedly give a lower average as seen in the table below. Comparing the mean power for a 90 min XCO World Cup to a three hour XCM World Championship there is only a four watt difference despite double the length of racing. This perfectly underlines the comment by Rob Warner that Annika doesn’t loose very much power from an XCO to an XCM. And on top of this she is also able to finish strong. Peak power for the XCM race tops at 999 watts and is higher than that attained during a XCO and CX World Cup. In this instant it was attained in sprint finish which of course is close to the best way to get a peak power output. The second highest peak during the XCM race was somewhat lower with 780 watts.


Maxing out at 999 watts for a Marathon World Champion title in Singen (photo: Michal Cerveny)

Cyclecross racing has a lot of dismounts which will undoubtedly create a lower average power. For this race in particular it had a long uphill running section. This makes it somewhat unfair to compare the average to the mountainbike races. However, when removing those running periods with zero watts (but still quite a strain to the body) the average comes closer to that of the XCO and XCM races with 212 watts. I’m pretty sure that on a course with no dismounting the average power of a cyclocross race would for Annika be in line with the other two. Again, the question is of course if this is the nature of the racing of if this data is unique to Annika and her way of riding. Regardless, the numbers are much closer than I would’ve thought. 

Table 2. Maximal and mean power from three disciplines. * indicates average after removing running sections from the data  


Running up stairs in Madison (Photo: Russ Ellis / Cyclingphotos)


In comparison to other sports cyclists have a lot of competitions in a season that stretches almost 9 months of the year nowadays. Only a few have the luxury of complete autonomy when deciding their race schedule. For many riders the races come back to back and as much as a well planned, theoretical, evidence based plan is a nice thought, it’s probably mostly about being as fit as possible and having the mind for racing to get you well through a season.

With the presented data above I think it’s in order to say that one should be cautious about simply copying this and hoping for success. As stated earlier Annika is a very forgiving rider to train. She responds well to almost any training under the notion “no pain, no gain”. And as written in part one it’s likely not how you inflict the pain but rather that you inflict an adequate amount of pain. And this has been a trial and error process.

My personal view is that it has been beneficial for Annika to add volume to her training despite the fact that we often hear that she is fading at the end of the season. I do not agree on that postulate. And here my argumentation moves outside the field of physiology. In 2015 she won the World Cup final and was the fastest woman in the Andorra Worlds Team Relay. Despite a bad XCO Worlds last year she was still very strong at the first CX World Cup the week after. Keep in mind that performance is more than physique and made of many different components. Some that for Annika still requires experience and learning since she started in the sport late. Fortunately, this also means that there are plenty of room for improvement years to come.



12 Responses to “The “specialist” who refused to be categorized”

  1. Tom Bell says:

    A very interesting read and thanks for taking the time to publish this.

    As the British National Champion in Marathon (XCM) but primarily an XCO athlete, this was of particular relevance to me and confirmed a lot of my own findings.

    If I can ask a question? I’m assuming the mean watts listed for XCM Worlds and the XCO World Cup showing a difference of 4 watts are average power? I would guess that the difference in normalized power would between the two would be much greater, and if so, may be a better value to compare given it’s much more representative of physiological demand? Would you be able to confirm this?

    Once again, very appreciative of the content and I also try to be open with my own training and numbers.

    Best wishes,
    Tom Bell

    • Christian de Graaff says:

      Indeed a very interesting read. Thanks for publishing. I also like to see the difference in NP i also think the difference is more than 4 watt. Also think that the course plays a big role. If there are long and slow downhill sections with no peddling that the avg watt’s but also the NP drops significantly. Especially in XCO racing where the downhills are physically very demanding and you cant recover like an easier downhill In something like marathon racing. What’s your opinion about this.


      From ??

      • annika says:

        Hi Christian!

        Check ouot the response I gave to Tom. xPower of the two weren’t that different. I think your point about the course AND how the racing plays out (head to head or “single” riding) has alot to say as well. I bet that Nino and Sam’s files are way different when going head to head like in Stellenbosch compared to a race where they are alone in front riding their own rythm. Even on XCO courses.


        • Christian de Graaff says:

          Thanks Thomas for you’re reaction. And also thanks for your imput on the forum.
          I agree with you out of my own experience dat my power profile is different form a solo ride or a man to man fight like Schurter and Gaze had in SA.
          I would like to know how you’re doing the 3 and 12 min tests and how you use them for the training schedule. Maybe I can implement it by myself en the athletes I train. For now I use lactate testing and 20min FTP test. And for a indication of vo2max I use 5min max test.
          I’m very curious about you’re paper. I hope you can post a link when it’s ready. And we can buy it.



    • annika says:

      Hi Tom!

      We don’t use TrainingPeaks but rather the CP concept and GoldenCheetah. I can tell you what the xPower for three races were but keep in mind that there a differences between Coggans and Skibas model. So for XCM, XCO and CX the xPower is 261, 254 and 200 watts, respectively. Not that much different. As you yourself have proven and XCO’er can go fast in a XCM. I think that if you look at the Cape Epic the past years you can make the same conclusion. As someone pointed out on mtbr “Going from a 90 min. race to a 4 hour race really should be pretty darn close from a training and physiological sense. Just a bit of fine tuning I would think.” Check out the discussion here


      /Thomas (better half and “training partner”)

  2. Daniel Sánchez Carvajal says:

    Quisiera saber como aplican el modelo CP (Potencia Critica) para monitorear el entrenamiento aparte de los protocolos de Lactato y VO2max.

    Estoy totalmente interesado en saber todo sobre este modelo pues soy estudiante de fisiología y me interesa muchísimo la aplicabilidad de este modelo en el ciclismo general.

    Muchas gracias y saludos desde COLOMBIA.

  3. Luiz Fernando says:

    Very good article !
    I’d like to know how is the protocol of the 3 and 12′ CP test.
    Thank you very much in advance !

  4. Adrian Espinosa says:

    It’s always interesting read this kind of reports.

    I really was wondering about how she can do all she does.

    Thank you for sharing.

  5. Dave Pusey says:

    Thanks for the great write up, it’s always fascinating to see how others train and it confirms my own thoughts that you can train for both disciplines and still achieve excellent results.

  6. Fabian says:

    THX for this insight in Annikas training – is the 6 Zones model compareable to the 7 zones model, i.e. Zone 4 is FTP or LT2 and above? Z5 VO2 max and Zone 6 anaerobic?

    thx for a short reply

    • annika says:

      Hi Fabian!

      All the systems using different zones has the same goals (I assume) but are slightly different from one another. If using a 7 zone system it will be different from a 6 zone system. Sometimes we do intervals using a 7 zone system and sometimes a 6 zone system. But the analysis is done using a 6 zone system.

      It may be or sound a bit complicated by I like to roughly divide it up into work above or below threshold. The rest is nuances of those (although there’s a bit more to it). I you google you’ll find many explanations on the different zone system. Both Golden Cheetah and TrainingPeaks have excellent online videos

Leave a Reply