Cyclists, for instance, used to track mileage as a way to quantify training. While it is very easy to gather mileage data—in its most simplest form, all that is needed is a reasonably accurate map—tracking mileage alone does not give a true profile of the stresses associated with a training session, let alone the stresses that accumulate over the course of weeks and months. Mileage alone does not factor in such variables as terrain (was the loop hilly or flat), temperature, or wind; riding ten flat miles on a warm spring day is very different than riding ten hilly miles in the middle of a January freeze. And while one could draw a very rough equation between total miles ridden and overall physiological impact, such correlations will be extremely rough and non-predictive of future performance.
The next step for cyclists, then, was to track both mileage and time per session. By adding in the variable of time, a rough approximation of intensity could be determined. As heart rate monitors became common, a third variable was added to the calculus, and today power data promises to provide a fuller, more detailed overview of athletic performance.
The key concept to take away from the example of cyclists is that a variety of data need to be used in order to develop, assess, and predict individual athletic performance.
Avg HR * Minutes Training
Example: 120 average HR for a training session of 60 minutes would yield a TRIMP score of 7,200
The problem with this simple form of calculation is that it is possible to register a 120 average heart rate for 60 minutes in a number of ways, such as performing at a HR of 120 for the entire session (a steady effort), or by doing a series of high HR efforts of 190, followed by recovery at lower HRs (an interval session, for instance). While the TRIMP scores for both the steady effort session and the interval session are the same in this implementation, we know from experience that an interval session is more physiologically stressful than a steady ride at a conversational pace.
To better quantify training stress, it is possible to factor in time spent in specific heart rate zones, which provides a much more granular record of a workout session. In this slightly more advanced model, a TRIMP score could be calculated as the sum of a series of time in specific HR zones:
Total Training Stress = 7,785
Total Training Stress = 7,440
The advantage of this more detailed calculation is that it factors in different levels of intensity that might comprise a workout, but it still is problematic in that heart rate is not necessarily an accurate predictor of work actually performed. Heart rate can drift upward during the course of a session due to environmental factors (temperature, etc.), or heart rate can be depressed by fatigue, over-training, or cold temperatures. With the advent of cycling power meters, watts were added to training stress calculations, which provide an even more accurate accommodation of actual work performed. Power meters work great for quantifying training stresses for cyclists, but their data does not accommodate training stresses associated with swimming, running, and other athletic activities.
Enter Perceived Exertion
By carefully monitoring performance values such as heart rate and power, an athlete can quickly normalize their sense of perceived effort with actual field metrics. Incorporating perceived exertion can involve little more than adding an overlay to TRIMP scores like those calculated above:
TRIMP * PE = Total Training Score
Example of a Hard Interval Session: TRIMP of 7,785 * PE of 8 = 62,280/100 = 623
Acute and Chronic Training Stress
There are two different types of training stresses that we need to consider: acute and chronic. Think of acute training stress as the short-term impact of a training session. For instance, we know from practical experience that we will be tired the day after an interval session, so we try to build in easier workouts between intense workouts to facilitate recovery. Similarly, chronic training stress should be regarded as the cumulative effect that training has on use over an extended period of time. For instance, most good training programs will be based on a pattern of relatively intense of training cycles that might extend over a period of a month or two, followed by one to two weeks of light activity, again to facilitate recovery.
Putting It All Together
· The duration of training sessions
· The physiological stresses of training sessions
· The short-term (seven day rolling average) acute physiological stresses of a training program
· The long-term (say, forty day rolling average) chronic physiological stresses of a training program
For multisport athletes, it is critical that metrics are used to do like-kind assessment within specific sport sessions, as well as metrics that can assess the overall impact of multiple sport training session. What is important to keep in mind is that whatever method being used to calculate session stress should be considered over short and long-term horizons and that individual session scores have value only when taken in aggregate. The easiest way to see emerging patterns is to graph acute and chronic training stress over time:
The chart above tracks both my acute (blue) and chronic (red) training stress for the months of November and December, 2010. Note the significant variations in intensity as signified by the peaks and valleys described by the blue line; this line is marking a pattern of intense workouts, followed by recovery sessions, on a weekly basis.
What is more informative is the pattern described by the red—the chronic—line in that accumulated training stress has been gradually increasing since the beginning of November. Based on my annual plan, I would anticipate that the red/chronic line will continue to rise for the months of January and February, after which I have scheduled two weeks of easy to moderate training to prepare for my next macrocycle. As I accumulate more data as the season progresses, the peaks and valleys of these initial curves will flatten, making it even easier to track long-term trends.