Feature

Training to Win

Keen athletes could benefit from research which shows how to maximise fitness and minimise fatigue.

by Dr Hugh Morton and Professor Eric Banister

Take a bout of strenuous exercise, and what is the immediate reaction of your body? It is one of tiredness and fatigue. Another, slower reaction, sees your body adapt, or become fitter. No athlete can perform well if they are fatigued, however fit; nor can an unfit athlete perform well, however rested.

It is the interplay between these reactions which determine a good or bad performance on race day. We have developed a theory which allows an athlete to design a training schedule to maximise fitness, yet minimise fatigue on the day of competition.

The amount of training undertaken affects fitness and fatigue, raising both responses to higher levels. These responses can be combined -- fitness positively and fatigue negatively -- into a single performance output for analysis.

Training Input

The theory first requires a good measure of the training input. It isn't good enough to say "I run 10 km per day, 5 days per week". Instead, we have devised a measure which can be used for running, swimming, cycling and rowing.

It is based on the proportion of heart rate reserve elicited by the exercise, and on its duration. Heart rate reserve is the difference between resting and maximal heart rate.

Thus, one individual with resting and maximal heart rates of 72 and 202, exercising at 137, and another with 50 and 188, exercising at 119, are both at 50% of their reserve. Multiplying this proportion by the duration of the exercise in minutes, and by a weighting factor which emphasises shorter duration, high-intensity training over longer duration, low-intensity training, provides the desired measure.

The units of measurement are arbitrary, and are called "trimps", a contraction of "training impulses". As an illustration of the scale, running 14 kilometres in one hour at a heart rate of 150 produces about 125 trimps. Triathletes going for a long cycle ride or marathon run can generate several hundred in one arduous session.

Quantifying the
Responses

Secondly, we need to theorise precisely how fitness and fatigue respond to the training stimulus, and recognise that athletes train repetitively. We suppose that fitness and fatigue both respond by an immediate increment, according to the measured amount of training. If no further training is undertaken, both responses will die off, but at differing rates, fitness more slowly than fatigue.

Fitness may be enhanced and its decay prevented by successive bouts of training. However, successive increments for the same training impulse become progressively smaller, and fitness plateaus after a while unless the number of trimps is increased.

On the other hand, successive training bouts will maintain and increase the fatigue level also, particularly if the bouts are severe or insufficiently spaced. This, too, will plateau, but at an earlier time than fitness.

Integrating Fitness
and Fatigue

Fitness contributes positively, and fatigue negatively, to performance. Thus if fatigue is high and fitness low, a poor performance can be expected; if fatigue is low and fitness high, then a good performance is likely to result.

More precisely, predicted performance at any time is given simply by the difference between fitness and fatigue levels, allowing for an interconversion factor, or fatigue multiplier, between the two arbitrary measurement units.

Thus, given a measure of trimps and the dates on which they occurred, together with the values of the two time constants and the fatigue multiplier, we can calculate predicted performance scores. These can be extrapolated based on a number of training scenarios, such as continued regular trimp inputs, or a rest period, or a period of lower (or higher) than normal training.

Fatigue and fitness both respond to regular training as shown in the diagram. The initial worsening of performance is simply a response to the imposition of an unaccustomed training regime. It soon gets better as the athlete's body adapts. After this adaptation has established itself to the now accustomed regime, it too levels off.

A rest period has an obviously beneficial effect on performance, as performance initially improves. This phenomenon of improvement after rest, known as peaking after a taper, is now becoming more widely accepted amongst athletes and their coaches.

For the theory to be any use to an individual, we must know his or her parameters in order to identify the exact nature of his or her performance characteristics. To do this, we need to correlate actual performances, recorded as time trials and converted to a points score, with predicted performance.

A computer program can be used to search for the parameter values which jointly produce the best correlation between actual and predicted performance scores. These then describe that individual at that time.

Optimisation

Given a competition date, can we develop an optimal training schedule in terms of numbers of trimps, frequency of training and taper characteristics, which will produce optimal performance on that date? Or more generally, how can performance be maximised at any future time, given the previous training history?

Simulations suggest that in order to achieve maximal performance while avoiding overtraining stress, intensive training on alternate days over a period of approximately five months is best. In particular, the taper period of around two weeks is longer than currently thought.

We interpret this to indicate the importance of recuperative periods. On the shorter time-scale, 48 hours of rest should be allowed after heavy training bouts. On the longer time-scale, incessant periods of training for months on end are not beneficial, and rest periods of four weeks or so are best after about five months. This suggests two such cycles per annum for the serious competitive athlete.

We do have some experimental data which supports this theory of performance, but the studies are small and not wide ranging. Several barriers exist to the theory gaining enough acceptance to be put into widespread practice.

One is the difficulty in obtaining capable and motivated athletes as subjects who will, with their coaches, participate in training patterns which may perhaps appear significantly different from those to which they are accustomed.

Another is the notion that theoreticians like ourselves are in some sense "distant" from the real world.

A third is the as-yet not fully established correlation between fitness and/or fatigue and some real biochemical or physiological variables, which would make the arbitrary nature of fitness and fatigue become real.

Education may resolve the first two difficulties, and we are pleased to report that some progress is being made on the third.

Eric Banister is Professor of Kinesiology at Simon Fraser University, Vancouver, Canada.
Hugh Morton is Senior Lecturer in Statistics at Massey University.