Sunday, 20 July 2014

Multi-cycle training: outline of a possibly-new training method

It's about time I pin down the thoughts circling in my head over the past few weeks. I've made analogies about how to perceive running training. A fugue was one, arches another. But these are merely analogies to something that I haven't yet fully described. Here and now I will outline a meta training scheme that may -or may not- be useful. My only claim is that I have not seen it before, hence it could be worth considering.

If you'd like to skip ahead, in a few paragraphs I will describe how overlapping different cycles for different training elements could lead to possible added stimuli in a training plan without a strict need to "up the mileage". Just look for the *****text and asterisks in bold*****.

If you are interested in reading the early stuff, let's outline what these training elements are.

Three training elements

Training is all about cycling different training elements; in particular speed, endurance and strength. Technically these "elements" are not really elemental at all. If you look too closely they disappear. For instance, at what pace does a fast tempo become a slow interval? When does a fartlek become a series of tempo runs? When do strides become sprint intervals? When does a medium-length run become a long run?

But it rarely stirs up controversy in and of itself to define these so-called elements. To refresh, here are the ones I'll include for this post. It is easy to split any of these off into more categories if needed.

1) Distance running: defined as a speed you could theoretically run all day if you had to, and a distance/time that lasts longer than elements #2 and 3.

2) Interval running: a speed at which can be associated with some middle-to-long race distance (anywhere from 800m to 100 km) and not distinguishing between "interval" (1500m-10km pace), "tempo" (10km-30km pace), and marathon pace (30-50km). The speed can be run in any combination of portions.

3) Strength: Any number of exercises (plyometrics, weights, hill/flat sprints) that last only a few seconds and should -in theory, and hopefully in practice- require more power to perform than intervals.

With these three loosely defined -yet clearly distinguishable- elements in play, we are on our way to creating a new training schedule.

Typical training week

First let's look at the most conventional type of plan: the training week

Although completely fabricated, there is nothing particularly unusual about this schedule. All three elements follow a 7-day cycle, there's the traditional long run on Sunday, a few workouts during the week, and some strength work on other days. Each element is rated at four levels of intensity: Hard/Long, Medium, Easy/Short/Light, and "none".

Long/med/short runs are relative, and most non-beginners already know what each feels like. A "hard" interval might be 16x400m at 5k pace, a medium interval might be 30 non-consecutive minutes of tempo, and light could be 8x2 minute bouts of fartlek running. Hard strength might be 8x(7x 80%max) squats, a medium could be some non-trivial plyometrics, and light could be balance exercises and/or light upper body strength.

If we assign a 0-3 value to each element (0=none, 3=hard), a woven patten emerges cycling every 7 days.
Three weeks of cycling three different exercise elements. Additive "difficulty"
ranges from 2-4 and repeats every 7 days.
A 7-day schedule fills up in a hurry. Alberto Salazar, in his guide road racing, suggests trying an 8 or 9-day cycle to balance all the required elements, or, if one wishes to keep a weekly schedule, to use 14-day periodicity. in their 10k training plans suggests using a 21-day cycle. Whatever cycle you choose, the goal has always to cycle various elements to create a short-term balance of training stimuli.

I almost stepped into the realm of controversy in a previous blog post by suggesting adding mileage may not be necessary to stimulate better training. More volume is often necessary, but as a tool to increasing the complexity, hence variety to training. Hence I've been curious if it's possible to complexity training on a 1-2 month timescale without an absolute need to add more volume.

Mesocycle training

An near infinite choice workouts exists week by week. But on a longer timescale, how else can one add "another spice to the spice rack"? I hinted -using a fugue analogy- that re-arranging the order of the elements is crucial to adding complexity. Yet if we consider the 7-21 day cycle system, even if changing the specific workout week-to-week, or extending the long run, or adding a few extra runs in the day, the end result is still a 7-21 day cycle. Certainly volume can be increased with time, or more speed work, or heavier weights. But these long-term goals all count as meta-cycles, also known as annual or season goals. Seasonal goals can last many months, or years in length. Therefore we have a gap of time scales: either we concentrate on a week or so of training (specific workouts), or focus on the entire season (mileage/race goals).  With the above training plans we become trapped in a false dichotomy of thinking day-to-day or year-to-year. Since muscle conditioning lives on a 1-2 month timescale, clearly a mesoscale training should be considered as important. Is it possible then to have a training scheme with all three timescales?

Jack Daniels was one of the first to identify in-between, mesoscale, training. His studies and intuition led him to break training into 6-week phases. The theory being that 6 weeks strikes a balance between new training benefits and old-training stagnation. That duration need not be set in stone; anything in the range of 4-8 weeks (1-2 months) counts as mesophase.

Six-weeks apparently strikes the balance between "new stress" benefits and stagnation
Jack Daniels goes into a lot of detail about these phases in his Running Formula book.  
Mesoscale controversy

Notice, however, that the above phases are rather artificial. Daniels' Phase 1 comprises mostly easy running, hence a "warmup" to the tough stuff ahead, but Phases 2-4 all mix in a good deal of "quality" work (read: intervals/speed) and a tiny bit of strength [His book section "Strengthening Muscles" comprises just two pages in a 275-page book]. Phase 4 is more of the same, but perhaps faster, while avoiding your training weaknesses (I can't say I agree with this aspect). In practise it is hard to distinguish between his phases. Also there is nothing natural about the transition between Daniels' 6-week mesoscale divisions once you look more closely. On paper it says you should wake up, mark the sixth week of phase 3 on your calendar, and move on to phase 4.

Despite the utility of acknowledging mesophases, in practise you can't really tell when you've crossed the phase X threshold. Endurance improvements, muscle strength and neuro-condition all take place on mesoscale timelines, but of course not all at once. Therefore a top-down imposed phase system does not aid the runner in breaking apart training "elements". Brad Hudson tries to modify this issue (Run Faster, chapter 8) by tweaking these three elements within a mesoscale timeline, but the process is ad hoc.

Perhaps Daniels' phase system is best suited for a single big race, likely a marathon. Ironically he abandons his system precisely when you would expect it to shine. In his words, "The elite [marathon] program groups what would normally be called phases II through IV into one continuous 18-week phase". If the phase system fails at the elite level, its use should not be encouraged at lower levels.

For the record, Gordon Pirie is not a fan of phase training:
Another popular aspect of training which I think is very dangerous is that known as "periodization" - that is, breaking down the training year into various "phases", each of which is divorced from the others. Thus, the beginning of the year may be devoted to a slow distance "build-up", the second portion of the year devoted to hill training, a third part devoted to interval work and then speed training, and finally (though most of these runners never get this far) a racing season undertaken. The difficulty with training in this manner is that you go along quite well with one aspect of training (e.g. long distance running), and then suddenly, on a certain day, "Bang!". You start hill-bounding, or speed training, or something new, and the body simply is not ready for the change, and invariably, year in and year out, you are more often than not injured.
Keeping in mind both Pirie's valid points and Daniels' recognition that training ought to be measured over a timescale longer than a few weeks but shorter than half a year, here now I would like to suggest a modified approach to training cycles that introduces mesocycles more naturally.

****If you were skipping ahead, this is where you can start reading again******

A novel(?) mesoscale method

Up until now I have outlined that (micro)cycles and certain training elements exist; the former consist of 7 to 21 day periods while the latter comprises three or more types of training: Distance, Intervals, and Strength. Jack Daniels and others recognized that having only meta (seasonal) and micro (weekly) cycles is insufficient; a mesocycle (1-2 months) should also be included t monitor progress. Daniels then introduced a popular system for breaking training into phases (and which many disciples copied), but it seems there is something rather forced about these divisions.

This is where I'd like to add my two cents. I agree with Gordon Pirie that one should maintain all three training elements (Distance, Intervals, Strength) throughout a season. I also agree with Daniels that training should somehow comprise mesocycles.

So…What if each of the three elements were on its own cycle?

Using the concept of least common multiples, I realized that not all micro cycles are the same length then it will take several weeks until the three elements re-synch. Without much difficult one may create a natural multi-week mesocycle built upon several simple units.

In the training plan shown above, distance runs micro cycles last 7 days while intervals and strength last six. The lowest common multiple between 6 and 7 is 42, hence the mesoscale length of this training scheme is 42 days, or six weeks. Plotting the mix of cycles as a function of daily stress we see that with this one small tweak we create considerable variety within each six-week span (after which the cycle repeats).

Two micro cycles of 6 days and one of 7 create a 42-day mesocycle. Net difficulty ranges from 0 to 6, broader than the 2-4 range created from the all-weekly cycles. 
New combinations of days are formed by constantly re-juxtaposing different training elements. For instance, on Day 8 "hard strength" is paired with a long-distance run. This could be interpreted as a long morning run followed by an afternoon of challenging upper body strength. Day 9 is just intervals and not distance, which could mean a large volume of strides and a brief warmup/cool-down. Day 36 involves a long run and hard interval, which may be interpreted as a medium-distance run (8 miles?), followed by some hard intervals (8x1km?), and finished with another medium distance run. Combinations can be interpreted as the athlete/coach see fit, and particularly stressful days can safely be ignored if premature or unnecessary. The exact workouts and distance are unimportant. What is important is that a complex-looking 6-week cycle is actually the sum of three simple 6 & 7 day cycles.   

Some questions that emerge from this staggered scheme:

1) Can other mircrocycle lengths be used? Of course. If one desires a longer mesocycle, one may choose to create a (6,7,8) day triplet of microcycles, which will not repeat for 168 days; a (3,5,7) triplet will cycle every 105 days; or maybe try (5,8,10), which cycles every 40 days. And so on.

2) Will this system interfere with mileage building? No, and mileage can be increased whether or not a mesocycle has completed. One of this method's benefits is to separate total distance from workout intensity without being restrictive about their absolute values. That said, it's probably best to use the length of a mesocycle as a yardstick for changing training intensity.  

3) Won't this mean having to do workouts any day of the week? It is true this system is a more demanding schedule. It is not recommended for beginners but serious athletes that can adjust their training to suit a larger-pruspose plan. I would also recommend having at least one element cycle on a 7-day period to prevent days of week becoming meaningless.

4) Has this method been tried before?  I'm really not sure. I haven't come across it, and I've read enough training books that I suspect it's not widely used. At the very least there's no harm in suggesting this workout method, but I do feel this has been attempted somewhere (Russia maybe?).

5) Will mixing training days increase risk of injury? Since all three elements are constantly present it is unlikely an athlete will over shock their system when exposed to new combinations. Ideally one should practise each cycle separately for a couple weeks before combining them together, or holding difficulty constant for two while varying the third. In the long term this mixture should result in a more robust athlete, capable of many combinations of exercises.

6) If a workout day is missed, does that ruin the entire mesocycle? Like a pianist missing a note, one does not typically start from the beginning but rather keeps moving forward. A missed day can be skipped, or switched for nearby an easy day. The overall cycle then continues as normal.

7) Should one repeat all the same elements each new (42-day) mesocycle? It depends on the training goals and progress. Some may benefit from new or added difficulty inside each microcycle every six or so weeks, others less so. Some experimentation is necessary for all interested persons. One should

8) Can one add more than three elements, or reduce to two? Of course. I only list three as it adds sufficient complexity to keep this method interesting. Two elements can work well (say distance and speed only), and four is possible, though could make for a very long mesocycle.


Now that I've outlined this plan, I hope that it might stir some new long-term training ideas in the curious reader. It's not a fantastically difficult scheme to implement, and it allows mesoscale (4-8 weeks) training cycles to arise naturally. Personally I haven't tried it on myself...yet. Partly this is because I only began considering it a few weeks ago and partly because I don't know if there is more information already on this topic that may come to light. I am interested to hear any thoughts on why this may or many not work. I believe there is little harm implementing this method to any non-beginner runner. Following a challenging training schedule is worth the effort only if it create a better athlete. Therefore I expect new methods to be challenged before serious time be devoted to a new one.

Cheers all.


  1. Hey Graydon. I will start with this recent link:

    Alex Hutchinson posted a very interesting article about top skiers and biathletes’ training. The pattern is lots of easy work, with some of the rest sprinkled in at around 10-20% which is what has been found to be the ideal ratio. Their work is expressed in yearly hours, which breaks down to about 90min/session, and just over 2hr/day. Top runners are mentioned and we end up at around 100min/day.

    I think we need to reconsider the idea of mileage as a weekly total, and think of it more on a daily or yearly scale. Daily because the day-to-day is what will accumulate over time, and like it or not, we live in daily cycles. I think this is much more natural than the weekly cycles. I’m in agreement with AS there. Yearly because it is a much more natural cycle. We go through the seasons, there are usually one of each championship/goal race each year (I mean there is only one Tely Ten, but there may be multiple 10k or 10 mile races).

    Weeks are convenient only in that they are a more artificial construct of our lives. I don’t think we should ignore them completely, but I do think we should be flexible in our workout days: the long run doesn’t HAVE to be Sunday. In your example, you end up with a 42-day cycle which is interestingly the same as Daniels’ 6-week cycle. Does that mean there is something to it? My observations tell me there is, but I agree with you that it could be 4-8 weeks as well. It does seem to take about this long to absorb a certain amount of training and see a change in performance though.

    The staggered cycles are also interesting. Ryan and I have discussed this, as he noticed that if all three of strength, volume and intensity are cycled together, sometimes the load is too much, and the recovery weeks too little. I might argue in favour of this more extreme method, if only because a larger stimulus and a bigger recovery should yield a greater adaptation. But the danger is too big a load.

    In terms of how the cycle should look, there is an interesting chart somewhere (I’ll dig it up and email you) that gives an idea of what should go where. For example, if you are doing intervals at a very fast pace, like 1500m or faster (Daniels’ reps let’s say) then you should probably be pretty rested for that. So you would want to do those first in the cycle. If you are doing a tempo run, the pace doesn’t matter as much as effort does, so if you do it tired, you still get the same training effect, even if you lag. So you could put that up against another workout, and be ok.

    Apparently the ideal time to do strength would be on the morning of an interval day. It’s easier to build strength in the morning, and also, when you get used to it, you don’t really feel the weights in the afternoon. If you do weights on Monday night, and try to do a workout on Tuesday though, you’ll be sore and maybe not get the most out of the running workout.

  2. Part 2: (my response was too long for the comments)

    Building all this together gets one quickly to realize that 7 days is not enough to fit everything in. One way to do it, if you must, (or even on a 8-9 day cycle this is fine) is to mix paces in workouts. Start with some tempo, then intervals. There’s a tendency to want to finish with speed, and I think that’s a good habit, but it might also be good to start with speed (once warmed up) to train fresh.

    I think if you did all this and also, on the side, without letting it dictate anything, tracked weekly mileage, you’d find that in order to get in the necessary work and in order to achieve improvements, that mileage would go up.

    A couple things: You can’t really have a day or session that is ONLY intervals. There will always be some “distance” in warm up or cool down.

    Re: Hudson and ad hoc adjustments. The whole thing is pretty much ad hoc. lol.

    Periodization: this is a word that refers to both ways of doing things. There’s “Block” periodization, where you focus on only building a couple elements, while the others are merely “maintained.” This can be done long-term or short term. For example, a certain Canadian who had a big breakthrough a couple years back would do two weeks of very high mileage (distance) and then a week of two a day intervals. This is high-risk, high-reward, but it is still periodization, in the same way as training all the elements pretty much evenly throughout.

    There are no “off” days in your 42-day cycle. Surely recovery has to be an element of training that is considered.

    In response to your question: has this been done and does it work? I think the answer is yes to both.

  3. Keeping in mind this plan is completely artificial, I see off-days on day 23 and 37, and "easy" days on days 7,11,13,16,17,25, 31, 35, and 41. Or do you mean these don't count ? Not sure I understand.

    Interval-only days would be a case of having to do a warmup and cool down, naturally, like you might do on a race day. Again, there are adjustments to be made in reality. Perhaps that day would mean in reality not running in the morning of an afternoon session.

    Also if 6,7 day cycles are too tight, one can easily choose longer cycles. Say the endurance cycle is 14 days, while workouts run a 8-day cycle. That would make for a 8-week (56 day) mesocycle.

    I'm not advocating a more rigorous training scheme, just one that gives rise to more diverse set of combinations using the same basic ingredients we always have.

    A more concrete example: at McGill we used to ALWAYS run workouts on the same days (MWF), until CIS approached, and we'd switch to one satruday workout. By moving the workout one day forward what had been easy was now a bit harder. Small changes like that can have an impact.

    Glad to hear it's being tried!