Although our release of the Weyand/Bundle algorithm as a training tool has occurred just recently, the concept of finding some formula—not just for predicting race times based upon training, but for actually assigning workouts times and distances based upon a specific athlete’s ‘speed profile’ is certainly nothing new.  I’m sure coaches have been doing something like this for years.

For me, the pursuit began in earnest back in 1980 when, after four years of coaching, I began to feel as if I did not have any handle on the way to assign repetition times for sprinters on my team.  Previously, I would just give them general guidelines, like five repeats of 120 meters at 75% effort or three repeats of 90 meters at 85% effort, or two reps of 60 meters at 95% effort.

What I didn’t know was the number of repetitions to assign, or what the percentage of effort meant for each runner on my team, and those two things bothered me.  Until I could answer those questions, I was really just guessing.  But the guessing was by ‘intelligent design.’  For example, the longer the runs, the smaller percentage of top speed, which means the more runs I could give them.  The shorter the runs, the faster the times could be, but with fewer repetitions and longer recoveries. It seemed to make sense, because that appears to be what everyone else was doing.  However, the problem was that I had no idea—nor did my athletes—of what 95% max meant.  Ninety-five percent-- of what?  And how was I determining their top speed?  From race performances?

In my pursuit of an answer I was  missing,  in 1979 I ordered  a little book from Track and Field News called:  How They Train, Sprinting and Hurdling, edited by Fred Wilt.  The first athlete presented in that manual was Valeriy Borzov.  That one page section on Borzov’s training began my career quest to quantify speed in order to design sensible, individualized workouts for my athletes.  Here’s what Wilt noted in the section on Borzov’s training:

“Coach Petrovskiy’s studies with the use of an electro-myograph led him to believe that there was a close correlation between times taken over 30 and 60 meters with expected performances over the standard 100 and 200 meter distances respectively.  The correlations were set into tables and served as a guide to what Borzov had to achieve over 30 meter tests from month to month.  In the early stages of Borzov’s training with Petrovskiy it became obvious that, according to the tables, Borzov lacked speed.”

This was exactly what I was looking for.  I never did ‘find’ Borzov’s complete tables, but another track and field manual did contain an article that revealed part of Petrovskiy’s 30-60 guidelines for sprint prediction. This book called Guide to Sprinting, was released by Runner’s World Magazine in 1973. The chart focused on a segment of projected times and shorter repeat speeds from 12.0 meters per second down to 10.0 meters per second, but even the slowest athlete in this plan (10 meters per second) would achieve times far faster than those for my athletes, which meant I had to extrapolate from the available data in order to determine what the 30/60 times would be for athletes running from 11.0 to 14.0 for 100 meters. At the time, I didn’t even know what “meters per second” really meant, but it would quickly become part of my vocabulary. 

 Regarding the tables themselves, they were clearly the product of good thinking, but no real science.  From what I gathered, Petrovskiy put together his initial tables without the benefit of any regression equation. As he notes in the article: “What would be the optimal model for achieving stable performances in the region of 10.0? What level of condition, judged by standard physiological measurements, should a sprinter have to be able to run this fast?  We were able to obtain and analyze statistics on the world’s best sprinters, including precise measurements for those competing in the USSR between 1963 and 1968.  We got the impression that the best sprinters have (a) very high starting speed, (b) very high absolute speed; (c) the ability to sustain their speed or “speed-endurance.”

On the basis of this material, we constructed a model of optimal running factors for the 100 meters. Average values for the best sprinters in the world were taken as the bases for this model. In 1968, Borzov was still trailing these sprinters [from the model] in start acceleration as well as absolute speed.  Comparison of actual and desired running factors enabled us to discover concrete differences and decide on training tasks.

It remained for us to find out how much the components of condition could be raised, and what signs should be weighed to measure the change.  By observing the training state of Soviet and foreign sprinters, we were able to develop and continually refine tables in which we attempted to evaluate the state of preparation of sprinters.  According to the chart, Borzov had to run at least 2.6 for the flying start 30 meters to be capable of his 10.0 performance.  Once he attained this absolute speed, his goal was within reach.

To understand the relationship between performances at 100 and 200 meters, and to have the basis for choosing a main distance, we again had to compare performances run by the best sprinters in the world.  In the course of training, we used test exercises year-round to monitor changes in Borzov’s state of preparation and to judge the effective ness of conditioning techniques.   All methods were used for the purpose of preparing him to do the work necessary to accomplish the main task: development of a specific degree of running speed.

Remember that speed of movement and running speed can only be increased by running fractional distances at maximal and near-maximal speed.  Repeat runs at competitive distance (100 and 200 meters, in Borzov’s case) at racing speeds in training are neither possible nor helpful, because the consequence is rapid exhaustion.  Therefore, the principal method of developing absolute speed consists of s runs over fractions of the racing distances.

I repeat, a person can only reach particular levels in these motor characteristics by repeat run over fractions of the racing distance.

And this concept is at the heart of ASR.  While it is certainly not new, it is important that we recognize the debt we owe to Petrovskiy for 1) coming up with an actual speed model, and not just a series of run distances, percentages, reps, and sets and 2) making sure that the model could be assessed and evaluated in terms of its contribution to the specific speed goal of each athlete.

So there you have it.  In the eighties, I ran a computer spreadsheet which extended Petrovskiy’s basic concept to athletes who were not capable of the goals or meters per second of world class athletes he had used to develop his tables. I still have buried away some of those charts I generated using an old dot-matrix printer.  The problem with those tables is that they really did not fit a universal profile of speed based upon a physiological explanation for the characteristic relationship between performance and event duration.

The Weyand/Bundle research provides us with a speed-model that does not rely on interpretive data from elite runners.

As they note in their 2004 paper:

“We have recently reported an empirical relationship that accurately predicts the high-speed running performances of individuals from two direct measurements and a single exponential time constant “

I’ve waited a long time for research to provide me with science based model that would allow me to structure both time over distance as well as distance over time speed goals while accomplishing what was Petrovskiy’s basic intention over forty years ago: developing a model for determining “how much the components of condition could be raised, and what signs should be weighed to measure the change.”