Thursday, December 4, 2008

Periodization vs. The Basic Week

“There’s more to being a model than just being really, really good looking”
- Derek Zoolander

Wise, wise words from Derek Zoolander. As many of you know, I have been really into models lately – training models that is. I have received some great feedback from you all, including a number of elite athletes that validates some of the theoretical constructs I presented in my last piece on the off-season. My buddy, Mat wrote a great ‘real world’ piece on his experiences with the off-season on his blog.

The past couple of weeks, I have managed to get my hands on a number of studies by the big players in mathematical modeling of the training process – Banister, Busso, Morton, Fitz-Clarke etc. While these studies contain some pretty sexy series formulae that describe fitness, fatigue and performance at any one time, as Zoolander says, there needs to be more to a training model than just being good looking. As coaches, we want more than theoretical constructs. We want a ‘roll your sleeves up’ computer model that allows us to put in data from a given athlete and accurately forecast race day performance using a variety of training methods, so that, as coaches, we can come up with the optimal training program for a given athlete. Software such as wko+ and RaceDay are a step in the right direction, however, they are much more easily used post-hoc rather than as a forecasting tool. I’m an impatient kind of guy and only have so much time to spend on trial and error. So….

I spent the extra time in the first week of my off-season creating a computer model of the impulse-response formulae of Banister (1975). I am not the first guy whose curiosity got the better of him. Rowbottom (2000) used the Banister formulae to expand on Morton's (1991) study and test different training structures. He found a consistent 3-5% performance difference when using a periodized training structure vs. a flat loading “basic week” structure.

A 24 week output of these two training structures with a training load of 100 TSS/d (with average ‘Middle of the Pack’ fitness and fatigue constants of 45 and 15 resp) is shown below (k1=1,k2=2):

X axis is days of training. Y1 axis is training load, Y2 axis is performance in arbitrary units.

Note: Both of these structures have exactly the same total training load (an average of 100 TSS/d) over the 24 weeks. In the case of the first, this is distributed as a the extreme example of a flat-loading “basic week”, same load day-in, day out over the 24 weeks. In the case of the 2nd, the load is distributed as:

Wk 1: 70 TSS/d (~12hrs of easy-steady training)
Wk2: 120 TSS/d (~21hrs of easy-steady training)
Wk3: 160 TSS/d (~28hrs of easy-steady training)
Wk 4: 50 TSS/d (~9hrs of easy-steady training)

The performance difference between the two structures is predicted at 1485 units for the periodized method vs. 1424 for the flat loading, a difference of ~4%.

Now, 4% is nothing to sneeze at. Most 11hr Ironpeople would welcome a 26 minute PR with no extra training load. Seems like a good deal. So, as Joe Friel asks in his latest blog post, “what’s wrong with periodization? “ Faster times for the same training load seems like a no-brainer. However, there are a couple of caveats that you should be aware of when deciding upon whether to use a traditional periodized training plan for the 2009 season.

The most important caveat is that in almost all cases, the structure of the training load is secondary to the quantity of training load.

In other words, there are many training programs that emphasize an arbitrary periodization structure: Dividing the total load for the mesocycle into 18%/30%/40%/13% etc. As displayed above, this is often a superior way to distribute training load providing the load is equal to what would be accomplished with a flat loading basic week. This is an important proviso. If you want to average 100 TSS/d (~17hrs of easy-steady training per week) this season, but the most training that you can fit into a week (without giving up sleep, increasing stress or compromising recovery) is 20hrs/wk, then obviously the training structure, rather than the total training load needs to be amended.

In other words, to answer Joe’s question, the thing that is ‘wrong’ with periodization is that our lives aren’t periodized. If you have the job flexibility to be able to work 50hr weeks on your easy weeks and 20hr weeks on your hard weeks, without significantly increasing job stress, then you should do it. For most of us, however, this structure is not an option and, particularly as the athlete improves, they will be forced to adopt a structure that more resembles a flat loading ‘basic week’ in order to accommodate the extra load. As a footnote, some of the most successful age group athletes that I have worked with are those with the most job flexibility. In other words, if you are serious about climbing to the very top of the age group ranks, finding a job or a position within your current occupation that offers more schedule flexibility should be a high long-term priority.

Obviously, the two examples I have presented here are extremes and deal with only one ‘level’ of periodization. Even those athletes confined to a flat load weekly structure, will likely have days that they can do significantly more work (e.g. weekends) and so the load won’t be constant on a day to day basis.

Additionally, in terms of the macrocycle, serious athletes can get a significant jump on the competition by taking advantage of appropriately placed jumps in the training load via training camps.

However, on a day-to-day basis the most important consideration in determining the optimal distribution of the training load falls not so much in the realm of the mathematician, but rather the realm of the life coach. Or, put another way, look at your weekly planner first and the training texts second when constructing your 2009 annual plan.

Train smart.

****************Update 12/22/08 ***********************

I received an interesting email from a coach asking about what the model indicates regarding the frequency of recovery weeks within the schedule, i.e. if we do want to insert periodic ‘rest and test’ weeks within the schedule, how frequently should they be planned?

It is not unusual for elite coaches to implemement a ‘recovery on demand’ approach, with the exception of some testing weeks. This approach is used by Aussie Swim Coach Bill Sweetenham to good effect. That said, most coaches will generally want to conduct some tests in a ‘fresh’ state. Additionally, even with a well tolerated basic week, accumulating fatigue is to some extent inevitable and periodic ‘shedding’ of this fatigue is indicated.

Obviously, if an advanced mesocyclic program is used with frequent shock weeks, such as the 70/120/160/50 cycle, more frequent recovery will be warranted. However, in the case of a traditional ‘phase’ periodization approach, the majority of recovery should occur within the week.

I tested a number of different frequencies for the insertion of a 60% recovery week on a 100 TSS/d ‘traditional’ periodized program for an average athlete (T1=45, T2=15, k1=1, k2=2).

I found that recovery weeks more frequent than once every 10 weeks results in a predicted performance decrement if this traditional ‘basic week’ model is used. In general, it is not until the 10th week of a given phase that the addition of a recovery week will actually improve performance. For this reason, I recommend that the coach who uses this model, as opposed to a more advanced ‘stair case’ model (with pretty marked changes in volume from week to week) should only plan to insert ‘rest and test’ weeks at the end of each phase not every 3-4 weeks as needed with the more advanced step loading program.

If recovery is needed more frequently than this then it is a good indication to the coach that the basic weekly load is excessive. For this reason, the coach needs to come up with a basic week that is sufficiently moderate to only require recovery after a prolonged, consistent period of training. The mechanics of this are discussed further in my recent blog on structural consideration in planning the microcycle (basic week)


-Brandon said...

Those %age differences are similar to some that are in one of the Swimming Even Faster books.

To ask another question. Do you have to pick one or the other for the working athlete or the pro/flex-time age grouper?

In other words, how would the model differ if a periodized phase was preceeded by a basic week phase?

What would be the implications of periodizing the workouts and the intensity within the basic week structure? Or is that what the periodized model with various TSS/day values shows?

Chuckie V said...

Any blog that starts out with a quote from Derek Zoolander is alright by me!

AC, I feel as though I should have to pay to read your blog, so useful is its content. Don't get any ideas now though.


Alan Couzens said...

Hey Brandon,

I haven't played around with the first format you suggested, although, I was a little surprised at what the model suggested regarding structure of the macrocycle. Backs up traditional swim/run formats with peak load early-mid cycle.

Regarding your second question, significantly backs up hard-easy training with a 7-10% difference vs. flat loading.

I'll probably write some more blog pieces on optimizing the microcycle and the macrocycle.

Best regards,


Alan Couzens said...

Thanks CV!

Right back at ya.

Your last blog piece was really interesting to put your philosophies together into a "rubber meets the road" program.

Best regards,


Gordo Byrn said...

REST WEEKS -- interesting what you write. A number of coaches that I have worked with use that technique -- rest when you need to.

One risk is that highly motivated athletes in impact sports have the capacity to break something (biomechanical failure) prior to physiological give out.

As well, if we get unlucky then the athlete can totally nuke themselves and write off weeks/months/years. So while recovery weeks might not be strictly required by the model, I view them as a form of insurance.

As well, it is only the genetic freaks that have the capability to survive, say, an olympic level Aussie swim program -- so there is survivor bias in the data.

Good stuff,

Alan Couzens said...

I wasn't that tired :-)

Your point on insurance is a good one and definitely worth adherence when the athlete throws down a week bigger than 'the norm'.

As a coach, I'm finding some advantage in knowing what an athlete's 'norm' (long term load capacity) is. Often if athletes know thay have a rest week in the works, they are tempted to really 'hit' 1 or 2 weeks. If I tell my athletes not to expect a rest week for a couple of months, they seem to be a little more cautious when it comes to exceeding the mean. This doesn't mean that I won't throw a multi-day recovery break in there if needed, but sometimes the threat of no rest is enough :-)

Good stuff.