CrossFit Discussion Board

CrossFit Discussion Board (
-   Fitness (
-   -   CrossFit Programming and Fractals (

Barry Cooper 01-05-2007 07:45 PM

I had this idea a while back, and was just reading the thread on the lack of swimming gains, started to reply, then decided maybe a whole new thread was in order.

Being the touchy-feely Humanities type that I am, I was roughly 20 years late getting up to speed on fractals. Those of you who are the mathematical/physics types please correct me if I'm wrong--and it's overwhelmingly likely I'll be wrong in at least some particulars--but my understanding of what a fractal is is a graphical representation of an orderly chaotic system.

By this I mean that it has been discovered that most apparently chaotic systems can actually be described/created with a relatively small number of equations. What is unpredictable is exactly what will happen in what precise order--these are definitionally non-linear systems--but what can be predicted is that certain patterns will recur, and what a fractal (which I believe is short for "fractional dimension") shows is self similar patterns, on all scales of analysis. They combine the properties of infinite linear distances, in finite space. Creating these things, that look like snowflakes, or spirals, or whatever, was a big deal when PC's started becoming widely commercially available back in the 80's.

Now, CrossFit is constantly varied, if not randomized. According to whatever CrossFit Journal that was, Coach has a relatively simple template for the creation of the WOD's, but is able to create essentially limitless variety with that template. However, there is still clearly a pattern. This, it seems to me, roughly fulfills the requirement for the creation of a fractal. If you could symbolically represent various movements, then you could literally create a visual pattern with the WOD's, I think. If you were much smarter than me.

Now, this raises an interesting question with respect to health and fitness. Our goal is to simultaneously increase and decrease our systemic homeostasis. Increase it, in the sense of physical capacity and capability. Decrease it, in the sense of avoiding accomodation to repeated stimuli, which in this analysis would represent a linear function. Progressive Resistance, of course, is a paradigmatic straight line. Most periodization schedules, though, would also fall under that rubric.

Now, periodization makes sense to people, because they can SEE the underlining patterns. However, the point is made repeatedly in the book I'm stealing these ideas from--Chaos, by James Gleick--that nature, by and large, eschews straight lines. It works better with embedded patterns that are neither fully random nor fully predictable.

Given all of this, it occurred to me that this might be an interesting way to create periodization relative to specific fitness goals, without unesthetic straight lines.

Specifically, the relation between GPP and SPP is still an relatively vague area. We know there is a clear relation between generalized work capacity and the ability to handle varied tasks. This is definitionally the outcome we cultivate, and I suppose in that phrasing is perhaps redundant.

However--and this is my "what if"--we could define programs with specific biasses. For example, decreasing 10K times, or increasing max deadlifts, or pushups, or whatever. Now, clearly, this can be done without getting too fancy just by the application of common sense. But wouldn't it be interesting to represent various elements symbolically, plug them into equations, and pop out a workout schedule that will trend for that individual towards a specific goal. Which would reliably increase SPP without unnecessary sacrifice of GPP.

Now, this would apply more obviously to the physical components of fitness than to the neurological components, but both could theoretically be included. If you want to improve basketball performance, you could allocate basketball practice as one of your variables. However, I think you need neurological cycling as much as physical cycling, so I'm not sure time on the court shooting baskets is always the only good way to improve your ability to shoot baskets. I could be wrong, but that's my hunch.

Anyway, all this may be WAY too complicated, but I'd be curious what the mathematical types think.

This is a hot Friday night for me. Whoo-HOO!!!!

I suppose I could go get a beer somewhere.

Lincoln Brigham 01-05-2007 08:20 PM

Note that even periodized workouts never actually repeat in practice. By the time the entire period is finished and is ready to be repeated, the trainee is at a different fitness level. The system called the human body is so complicated and fluid that true homeostasis never happens. Something has always changed, especially in regards to fitness.

Which means that the human body is too complicated to rely on exact formulas to generate workouts. You'd never get enough relevant data compiled before the body moved on. Heck, most people can't calculate their calorie intake to the nearest 5%. That's a margin of error equal to 10 pounds of bodyfat per year. That's a significant whopper of an error and that's just one variable.

Fitness programming is an art form. Math can only hope to explain WHY it must remain an art and will never be an exact science.

Barry Cooper 01-06-2007 09:39 AM

That may be true, but obviously most periodization programs DO at least start from exact formulas. My point, though, is that they are inherently linear. Very few powerlifting programs include running or handstands.

Perhaps I could phrase this another way. Most of us have heard of Specific Adaptation to Imposed Demands. One of the seeming paradoxes of CrossFit is that specific adaptions often occur without the specific stimulus. Back Squats go up without doing weighted back squats, for example.

Yet, if one were to decide to increase one's max back squat, if one were a non-CrossFitter, they would use one of the Bulgarian or Russian or whatever programs. Week one, day one, 3x8 with 42.7% or whatever. This sort of stuff appears scientific to many people.

CrossFit, the dominant idea is the Maximal Effort Day integrated as the middle day. Given that Days one and three are unknown in advance, this is a chaotic program. It also has the great merit of being simple. And working, from what I hear.

My idea is really twofold. One, looking at generalized fitness as best created and defined by "orderly" chaos. That's the WOD, as it is presented to those who do what's written. And it works. The only new idea there is mathematical. And I'm not sure even that is new.

However, the other idea is approximating periodization for purposes of SPP. In such a program, O-Lifters may go weeks without O-lifting, then lift 5 days straight. If the equations were balanced properly, it might yield better results than current models.

As with all of this, results matter. This just seems like a very interesting approach. Somebody could write a Ph.D thesis on this. As you know, sports physiologists spend an inordinate amount of time on this, and have to this point very little to show for it.

Chris Goodrich 01-06-2007 10:07 AM

DeVaney on fractals and training: [url=][/url]
(work/family safe if you can get past his ego).

Barry Cooper 01-06-2007 03:17 PM


That makes a lot of sense. It occurred to me that the WOD is already fractal, and ME day within that framework DOES act as a targeting modifier. I'm not sure I've added anything new, except possibly a new descriptive term, in this context.

I'll call this my "fun with numbers" hypothesis.

All times are GMT -7. The time now is 05:10 PM.

CrossFit is a registered trademark of CrossFit Inc.