Speedwork for triathletes

(Note: Originally, I had written this article for Gordo Byrn’s website).

In endurance sports, three variables are of prime importance to performance: Vo2max, which is largely genetically determined, anaerobic threshold and economy. Although there is no real consensus on the definition of anaerobic threshold, we will admit it is equivalent to the OBLA, the onset of blood lactate accumulation or the intensity of exercise at which the production of lactate equals its clearance. Economy is generally determined as the percentage of the Vo2max sustained at a particular submaximal intensity. For example, in running, economy could be defined as the quantity of oxygen consumed per minute and per kg of body weight at 14km/h (or 16km/h for more efficient runners).

Vo2max is slightly more complicated as Dr. Veronique Billat has shown. She is a French physiologist whose aim is not to explain theoretically why such or such method of training is more efficient, but to design new methods from theoretical results! She has applied her results to herself rather successfully with a 78′ PB for a half marathon…

The main idea was to separate vo2max in two variables: vVo2max and tlimVo2max. vVo2max is defined as the velocity at Vo2max, and tlimVo2max is the maximal time an athlete can sustain at vVo2max. But what is the point if Vo2max is genetically determined?

Well, indeed it is, but vVO2max and tlimVo2max can be greatly improved! You can have a 65mlO2/min/kg Vo2max and 19km/h as vVo2max when you first test, and a year later, with a Vo2max of 66, could be reaching 20km/h! So basically forget these big numbers of Vo2max. What is really important are vVo2max and tlimVo2max and to find these, you do not need to go to a lab and get hooked to some funny device! We will see further how to find these values in swimming, cycling and running.

After that bit of theory, what you want to know is: how is that going to make me a better athlete? What do I need to improve? Simple: everything! You need to improve your vVo2max, tlimVo2max, running economy and lactate threshold (LT). Though, you may be thinking, that with three sports, and four variables to improve, plus for many a heavy working and family schedule, you would not have the time to work on all these values.

Well, yes, if you can work on all of them in a single workout! What Billat has shown is that a specific work, at specific intensity, for a specific time (understand INDIVIDUAL), you could improve simultaneously these four variables. She has found improvements of 3% of vVo2max, tlimVo2max, LT, and 6% of (running economy) for sub-elite runners (with PB around 70′ for half marathons)!!! These improvements are simply huge and therefore Billat’s work really needed some attention. Besides, this improvement was observed after a four-week period.

We will first describe how to determine vVo2max, tlimVo2max and also a short cut for very busy athletes, or athletes who already have a pretty good idea of their vVo2max. These tests are easy to do and are field tests. You will not have to go to a lab and spend some big bucks! Then, we will described the workouts you can design, and eventually we will show how to integrate these sessions in your training phase depending on the sort of triathlon you are competing in.

Determining vVo2max and tlimVo2max

We will explain the field test in running, and will then explain how to adjust it to cycling and swimming, as was done by Billat. One important remark – you will have to test again to evaluate improvements. When doing field tests, it is very important the tests are being done in similar conditions: do not test in a 50m pool, then in a 25m pool, or on a tartan track, then a grass track, or test on a cycling track with heavy winds blowing that day. Test always at the same location (when possible), at the same time, and keep the same warm-up, recovery, training day before (easy) and if possible eat the same things the 24 hrs before. Okay, to the test now.

  • Go to the track, and have a good warm up, around 20′. You will then proceed with a step test.
  • Start at 10km/h to 12km/h depending on your running level. Assume you start at 10km/h
  • Run 2′ at 10km/h, and increase the pace by 2km/h every 2′ until you reach your estimated LT (your speed for a 10km, to simplify), then increase by 1km/h every 1′, until you can not accelerate anymore (it usually means you have to slow down). This is your vVo2max, or at least an accurate estimate of it. Ideally, place some cones on the track every 50m and for each 2′ step or 1′ step, calculate the corresponding time you need to run for 50m. Then have a thorough recovery.
  • Have at least 48hrs to recover, then go back to the track and after a good warm-up, identical to your warm up for the firsts step, place your cones every 50m, and run as long as you can at your vVo2max. Once you can’t reach a cone at the estimated time (let’s say you are more than 10m far from it), you are done. This will represent your tlimVo2max.

Remark: don’t think you need some quality workouts this week. I can guarantee you, these two sessions will be enough!

Now, you want to determine your vVo2max and tlimVo2max on the bike and in the water. First, let’s see a good way to estimate it in the pool.

  • Have a good warm-up again, then because, we do not care about the speed in km in the pool, you will determine your vVo2max as a pace per 100m. Start between 2’20” and 2’/100m then increase your pace by swimming 10” faster for every 100m until you have to slow down or you have to stop, or you can not go any faster. This gives you a good estimate of your vVO2max. Then again, have a good recovery. At least 48 hrs later, you can determine your tlimVo2max by swimming as long as you can maintaining your vVo2max.
  • For the bike, it may be slightly more complicated as the weather may affect widely the results. Try to do these tests when the conditions are as quiet as possible. Also, a track seems ideal for this, for obvious security reasons. Another possibility is to test on a trainer that displays the power output such as a computrainer, or a Tacx Grand Excel, or for the more wealthy, to test with SRM cranks that display the power output when riding outside, as the power output is not (as much as the speed) modified by the weather conditions.
  • Again, have good 20 to 30′ warm up, and start at about 20km/h, hold it for 2′, and increase your speed by 2km/h every 2′ until you can not accelerate, or have to slow down. This will be your vVo2max. 48 hrs later, you can go back to the track and determine your tlimVo2max (a flying start is allowed).

Adapting the tests for busy athletes (or athletes who already have a good idea of their vVo2max)

Okay, I know that these tests represent six training sessions (although, as said earlier, they are worthy!), and most athletes do not like to spend too much time testing. An alternative to these tests is to do a Time Trial. Indeed, Billat has shown that on average, runners would have a tlimVo2max of 6′ or thereabout. Therefore the alternative you have is to go to the track and perform a 6’TT.

Your vVo2max will be determined as the speed you have hold during this 6’TT (once again, after a good warm up, and do not forget the recovery). Of course, your tlimVo2max will be assumed to be 6′.

Nonetheless, we believe, it would definitely not be waste of time to test at least once both vVo2max and tlimVo2max, specially knowing that tlimVo2max is linked to your lactate threshold (see further). Once you have tested both, you can perform a TT (at tlimVo2max) to determine improvement in vVo2max.

To make things slightly more complicated, Billat has tested the relation between tlimVo2max in different sports (swimming, cycling, running and kayak). It appears, that generally swimmers and cyclists will not be able to hold their vVo2max as long as runners or kayakers. The explanation is pretty complicated though (check the references).

So, if you rely on a TT to determine your vVo2max in swimming, you should perform a 4’30” TT which is the time swimmers can hold their vVo2max on average. To be more convenient, a test of 300m for moderately good swimmers to 400m for advanced swimmers would be better to get an estimate of your vVo2max.

In cycling, the time spent at vVo2max is also around 4′ to 4’30” and therefore, a 4’TT would provide an accurate estimate of your vVo2max.

Once again, although, this will provide reasonable estimates, we believe if not testing entirely at least once, you would miss a very important point in Billat’s work.

Now, you have six variables: vVo2max and tlimVo2max for swimming, cycling and running. So the next step is to know what you should do with these figures. What training should you do to improve these numbers?

Improving your vVo2max and tlimVo2max (and other variables)

Training sessions at vVo2max have been “traditional” sessions among runners, cyclists or swimmers for years now. Although initially, a whole bunch of athletes was doing the same session at the same pace, it did not took long before coaches realized that in order to be individual, each athlete should train at a different at Vo2max.

However, this was the only change. No one ever questioned seriously if everyone should do the same distance or time at vVo2max. This is mainly because no one ever wondered how long could someone sustain vVo2max, essentially because there was not one single answer as tlimVo2max is as individual as vVo2max.

Billat has found that even among elite runners (with vVO2max between 22 and 24.5km/h) where we would have expected some homogeneity, the difference in tlimVo2max was important, going from 4’30” to slightly more than 10′. A difference of nearly 130%! The natural question to come was whether tlim and vVo2max were related or not. Well, generally, someone with a very high vVo2max, for example a 5000m track runner, as a 5000m is run at about 98% of vVo2max, would have a lower tlimVo2max, than let say a half marathon runner, who runs very close to his anaerobic threshold.

And funnily, this is the link. The variable tlimVo2max is positively correlated to the anaerobic threshold or lactate threshold of a runner (or swimmer, or rider etc…) that is the higher the LT, the higher tlimVo2max. Therefore improving your tlimVo2max will also improve your LT. This is (slightly simplified) why, the sort of sessions designed by Veronique Billat, can simultaneously improve your vVo2max, tlimVo2max, LT and running economy!

The typical 30-30 session in running:

This session has been very popular in Europe for several decades, although there was no real strategy as to what pace, how long, when the session was to be done. It was a traditional session in the early phases of training, to do “some speed work”, but forgotten further in the season as believed to be “too easy”. Though, Billat’s findings contradict this belief. The gains mentioned above among runners who followed the training designed by Billat were done using this 30-30 training session. So what is it?

On a one hour session run at around 70% of your vVo2max, you should include repeats of 30″ of effort at 100% of vVo2max with 30″ of recovery at 50% of vVo2max, the total time recovery included being between 2x tlimVo2max, for beginners to intermediate and advanced in the early season, to 2.5x tlimVo2max for advanced in an intense phase of their training or elite athletes.

So let us consider runner A, with a vVo2max of 18km/h (40″ per 200m) and tlimVo2max of 8′ and runner B with a vVo2max of 20km/h (200m in 36″) and tlimVo2max of 5′. Both are supposed to be advanced runners in the early season. A and B will both run one hour (or slightly more or less) and will do the following:
A: 16′ alternating 150m in 30″ and 75m in 30″ (this can be done on track with cones to visualize the distance to run or on a measured course)
B: 10′ alternating 180m in 30″ and 90m in 30″
In this case, you realized how important it is for the pace and the time of effort to be individual. If A was to do B’s session, it would end up being more of a lactate tolerance or power workout, and B doing A’s session would do something that looks like a LT session where the time of effort is not long enough!

Adapting this session to swimming and cycling:

In the water, the session is easier to do as a set of 50m repeats (or 25m for beginners). This session could be included in a 4k-swim set with some aerobic work. Though, because it is important to maintain a very good technique in swimming (not that this is not the case in running and cycling, but it is perceived as harder to do in swimming due to the “non-friendly” medium!), the set would be better if split in two, with some aerobic active recovery in between the sets.

Consider a swimmer with a vVo2max of 80″ per 100m or 40″ per 50m and who can sustain his vVo2max for a 300m (4′). He could choose between active recovery or complete rest between the efforts.
S1: 6x(50m in 40″ with 25m very easy) leaving on 1’20” – 300m easy aerobic – 6x(50m in 40″ with 25m easy) on 1’20” or
S2: 6x50m in 40″ leaving on 1’20” – 300m easy aerobic – 6x50m in 40″ leaving on 1’20”.

For cyclists, the session is a bit harder to organize. First of all, do not attempt this sort of effort unless you are sure your training area is safe and you are sure your training session will not finish in the door of a car! If you do this session outside, try to do it where you have tested to make sure, the time and speed you have to hold are reliable. A cycle track is ideal.

You can do it on a trainer with power output, or on a regular trainer (on which you can shift the resistance from the handlebars) if you have a bike computer on your back wheel, but a computrainer or Excel would definitely be a plus.

So, consider cyclist C, with a vVo2max of 40km/h. In 30″ he will do 330m (the length of long cycle track, so it would be very handy!), and tlimVo2max of 5′. His session will be 1h30′ (or more if you have time to include it in a longer ride) with:
C: 330m in 30″ and half the track in 30″ recovery during 10′.
An alternative to this, if you know a road with rolling hills, with short enough, it takes about 30” to climb them, then just do 10′ surging up the hill (in this case, you will have to trust your sensations, unless you have SRM cranks) and recover gently downhill before surging again up the next hill. Although this type of session is fun, it will not be as accurate as a cycle track or a flat road where you really have some speed indication (unless you can go as fast uphill!)

Other Sessions

The other classical session designed by Billat, in the case of runners is a set of 1000m. Consider runners A and B mentioned above. On a 1hr session, they will perform the following:
A: 5x1000m in 3’20” with 500m in 3’20” recovery while B will perform
B: 3x1000m in 3′ with 500m recovery in 3′

If you have been careful, we will have noticed that, here each runner’s set is longer than 2x tlimVo2max (including recovery). Actually here, each runner’s set is 2xtlimVo2max, without including recovery! Why such a difference? When Billat did these tests, she had access to K2s (portable gas exchange analyzer) that a runner or cyclist can use while training and therefore she could know the exact percentage of Vo2max being used by the runner. She then realized that when doing a set of 1000m, the runners Vo2max would drop very quickly after stopping and when starting their next 1000m, they were at low values of Vo2max (around or under 60%), and what is important is the time actually spent at Vo2max. Then, when doing those supposedly much easier sessions of 30-30, she realized that after the 5th repeats, all runners (elite or not) were at 100% of their Vo2max, even at the end of the 30″ recovery! This gives a new advantage to using these sorts of sessions (compared with the traditional 1000m repeats and so on) as you will spend as much time at your Vo2max but will actually run less at a high speed, and therefore, will risk less injuries.

Anyway, repeating the same sessions week in and week out can be boring, and using other sessions may sometimes be useful.

In cycling, you could perform repeats of 2k-effort at your vVo2max for example with a 1k recovery. In practice, this sort of session will probably be easier to organize than the previous 30-30.

For swimming, we believe that unless you are a very experienced swimmer (maybe able to swim 16’30” or 17′ for 1500m which would be a limit), it is not wise to schedule similar sessions in swimming (like sets of 250m at vVo2max) as it is very likely that you will not be able to hold the technique, and this is crucial. Therefore, sets of 50m would probably be a better strategy to improve vVo2max etc., in swimming.

Scheduling these sessions in the season for short course and long course triathletes

We will go rather quickly on this part, as there may be a lot of different alternatives to organize these sessions in the season. Therefore, we will only list some possibilities.

Short course triathletes

For these athletes, contrary to long course triathletes, being able to tolerate high lactate contents (also something you are working on with Vo2max workouts) is very important, especially for athletes racing draft legal races. This is true for every leg, to close a gap when having a bad start in the swim leg, to break away from stronger runners on the bike, to surge on the run. Therefore these athletes could use these sessions on a weekly basis:

  • One Vo2max session per sport in the week or when trying to work on one discipline specifically. For example, someone working on his running could have,
  • One 30-30 session and one 1k repeat session in running.

In this case, we believe, adding two extra Vo2max sessions (one for cycling, one for swimming) would be too much stress and the athlete depending on his phase of training would have to decide for which he/she would go. Though, Billat has shown that the gain from adding a second Vo2max session were not very significant and therefore should be considered only by elite athletes looking for the little extra that could make the difference.

Long course triathletes

For long course triathletes, the interest of these sessions is essentially improving economy, strength or LT. Many athletes perceive the 30-30 workout as much easier than a traditional LT workout such as 3x3k with 3′ rest which can be pretty stressful, and therefore, Vo2max workout would be a safe and less stressful alternative.

  • One Vo2max session per sport in the week of the last 12 to eight weeks of the building up to a long distance triathlon or an Ironman, depending on how well you tolerate speed workouts,
  • One Vo2max session in running per week for the last 12 weeks leading up to an event, and 1 in cycling and swimming only for the last five weeks. As four weeks was enough to generate significant improvement, and because of the non-impact aspect of swimming and cycling, it would not be a problem to have seven weeks with specific LT workouts, or tempo workouts (LT-5 to 10 bpm), followed by four weeks of Vo2max work (and a last recovery week before the event).

In any case, you will have to find out what works for you! There are many ways to organize your training schedule and to include these sessions, so experiment, carefully monitor your sensations, and make up your mind about what is working for you!


Reproducibility of running time to exhaustion at Vo2max in subelite runners – Medicine and Science in Sports and Exercise, 1994. Billat et al. This paper shows that 1′ or 2′ steps do not influence the values found for vVo2max and tlimVo2max.

Time to exhaustion at Vo2max and lactate steady state velocity in sub elite long distance runners – Archives Internationales de Physiologie, Biochimie et de Biophysique, 1994, N102. Billat et al. Shows that tlimVo2max and lactate threshold are positively correlated and improving tlimVo2max would improve LT.

A comparison of time to exhaustion at Vo2max in elite cyclists, kayak paddlers, swimmers and runners – Ergonomics, 1996, Vol. 39, N2. Billat et al. Shows the difference of average time spent at Vo2max in different endurance sports.

Anaerobic contribution to the time to exhaustion at the minimal exercise intensity at which maximal oxygen uptake occurs in elite cyclists, kayakers and runners – European Journal of Applied Physio, 1997, N76. Billat et al. Similar to the previous paper but includes the %age of anaerobic contribution in each sport.

Biomechanical events in the time to exhaustion at maximum aerobic speed – Archives of Physiology and Biochemistry, 1997, vol 105, N6. Gazeau et al. Shows that runners who are able to hold stable running styles were able to hold longer their vVo2max, tlimVo2max and economy are positively correlated.

The Vo2 slow component for severe exercise depends on type of exercise and is not correlated with time to fatigue – The American Physiology Society, 1998. Billat et al. Explains why the tlimVo2max differs between running and cycling (from a biomechanical perspective).

Interval training at Vo2max: effects on aerobic performance and overtraining markers – Medicine and Science in Sports and Exercise, Dec 1997. Billat et al. Shows the improvement one can expect with vo2max workouts in vVo2max, tlimVo2max, LT, economy, and the less stressful aspects of these sessions compared to LT workouts, as well as the need of individual design for these sessions.

The role of cadence on the Vo2 slow component in cycling and running in triathletes – International Journal of Sport Medicine, Vol. 20, 1999. Billat et al. Shows that most runners will choose the most economical stride length but in cycling the most economical cadence is not the best to choose as it does increase muscular fatigue. From a Vo2 point of view, the most economical cadence (the one leading to the lowest oxygen uptake) is around 60rpm, but this cadence is increasing muscular fatigue. These cadences coincide in running.

Influence of light additional arm cranking exercise on the kinetics of Vo2 in severe exercise cycling – International Journal of Sport Medicine, 2000. Billat et al.


Duathlon Nationals

A couple of weeks ago, my good friend Brian Stover at Accelerate 3 told me come race with him in Tucson. At the time, I didn’t know that when he said race with me, it would really be that. I wasn’t too fond of the idea of racing a short distance either. Short means painful. Duathlon means really painful. Since I am kind of a jerk at times, I told nearly local John P. to sign up for the race too, so we can share the pain a bit. After all, I’m his coach, so he’d better listen!

Anyhow, here I am at the start line of a 5km-35km-5km duathlon in the Oro Valley, with about 400m of flat for the whole course. The moderate hills were likely going to be an issue for my right achilles tendon, which decided to flare up a few days before the race. Therefore, I make the decision to run the hills easy (which meant the whole course pretty much) to avoid aggravating it. And off we go. Aside from one guy who breaks away right at the gun, things stay relatively compact for the bigger part of the first run. I can still see Brian in his Trisports gear, about 30secs up, when I enter T1 in a very pedestrian 19.10 5km. The first transition is a total disaster, since I lose one of my bike shoes. Stop. Pick it up. Wow, can I be more rusty than that at short distance races?

The bike goes well. I pace the two laps very evenly, in the aero position 99% of the time with my big dorky helmet. I pass a whole bunch of people, Brian included, avoid a few slower riders with, let’s say, not exactly great bike skills in a couple of corners, and get back in transition with a relatively decent 50.10 bike split, which turns out to be a 51.55 since I lose a shoe again while entering T2. To add insult to injury, I run by my shoes in T2! Turn back, run to my rack and see Brian who seems amused by my so awesome transition skills. We head out for run 2 together. I get a small gap, but start getting really annoyed by the ‘Go Brian! Hang in there Brian! Looking good Brian!’ Hey come on now! I’m racing too here. I know I’m short, but I’m not invisible [Matt Q, insert joke here]. Brian ends up getting a little gap on the final climb where my achilles is not happy one bit.

Crossing the finish line, just under 1h33, with a decent ride, 2 slow runs, but good enough for a 6th spot, in a very competitive AG, where the winner comes from (obliterating the rest of the field). All in all, not bad at all for some short stuff, which happens to be quite fun actually! It was fun to see John race also (congrats on a good race mate). And I’m now 0-1 against Brian.

Note to self: avoid walking in transition saying ‘wow, I hope they let me race, I forgot my compression socks.’ That should do the trick to not get nasty looks.

Slaying the dragon

Matt E. went, saw, and conquered in St George UT. On one of the most feared courses of the Ironman circuit, Matt perfectly executed the race strategy he had trained for over the past few months, to finish his first Ironman in 12.09, in hot and windy conditions. I will paraphrase him a bit: the swim was easy and went by fast in 1.07, the bike was tough with stronger winds on the second lap (6.06), and the run was so tough that “it should be illegal”. Matt went into survival mode after the end of the first lap to finish in 4.49, and take 78th in the 35-39 AG.

I definitely like this as a first post. As a coach, it is always extremely gratifying to see one of your athletes doing exactly what you knew he was capable of, and delivering the great performance he has worked hard for. Although, let’s face it: Matt is tough. There was never a doubt in my mind that he wouldn’t be at the finish! Congratulations Matt, you are an … can’t say it; I think it’s trademarked.