About a year ago (May 2011) I received a Garmin 610 GPS/Training watch that incorporated a firmware version of the Firstbeat Training Effect (TE) software (here). The following is a summary of my experience with using this tool as part of training and periodization in endurance sport.
Is training effect just another datum that provides little real value to your training program planning and execution? I argue that it is a useful tool and one that can help ensure adherence to a training program and/or provide real-time feedback on training progress provided that you use an appropriate training effect analysis protocol.
Training Effect (TE) and Excess Post-exercise Oxygen Consumption (EPOC) are two parameters that Firstbeat uses to assess the activity level of a given workout session. A review of the physiologic basis of the calculation of TE and EPOC are contained in these white papers under the heading “The science behind EPOC”. Briefly, what Firstbeat have done is to correlate detailed training session heart rate (HR) data to detailed post-exercise recovery data and developed an algorithm that coverts one to the other.
Detailed real-time and recorded heart rate data is readily available with current technology (e.g. the plethora of heart rate capable training watches on the market). Many athletes are challenged with the prospect of the utilization of this data to guide their individual training programs. Analysis and interpretation of the HR data is not straightforward and can lead a user to faulty conclusions and therefore incorrect feedback to a training program. A coach can be very helpful in this regard; I wholly support the concept that anyone who is serious about their training and race performance will be well served by a qualified coach, whether the interaction is daily, weekly, bi-weekly, or monthly. However, there exists a reasonably large proportion of athletes in trail ultrarunning that cannot afford a coach, are not comfortable with coaching, or just prefer to guide their own program by educating themselves. It is this later group that can take most advantage of software/firmware offerings such as the Firstbeat EPOC and Training Effect algorithms. Coaches would also have access to training session specific information of their athletes that some are now lacking.
The importance of perceived effort
My experience in both national-level and Olympic-level endurance training programs and the nitty-gritty of the execution of these plans in real life, is that the primary variable of perceived effort and the associated training effect of that effort is often very difficult to reliably determine. This is important because any structured training program (with or without periodization) requires good measures of expended effort in a given workout for both completion of the program and for feedback to the program to allow for inevitable modifications. Additionally, one of the most common negative outcomes of a structured, periodized training program is overtraining and the resultant potentially long recovery time. Although other factors play into such overtraining, an athlete’s ability to correctly assess perceived effort during a given workout is critical to success and the avoidance of overtraining.
“To go fast you must go slow”
Numerous endurance Olympians have related to me that a fundamental basis to their successful training has been the mantra: “to go fast you must go slow”. What is meant by this is that although many athletes have no issue with pushing themselves to accomplish high intensity workouts (say L4-L5 intervals) in a structured program, the low intensity, L1-L2, efforts can be very difficult to adhere to. Yet these lower intensity sessions are critical to the success of the overall training plan, the ability to properly execute upon the high intensity training sessions, and to produce peak performance for focus races. Hence “to go fast you must go slow”.
We have all had the experience of going out for an easy L1-L2 run or ski only to find ourselves well above this level due to inattentiveness, a sense of well-being (i.e. “boy I feel great today, I am going to push it a little harder”), or just plain defiance of the structured program (which as many coaches know is not uncommon). Such effort-positive excursions from scheduled workouts can significantly impede training progression and lead to overtraining or lackluster performance both in high intensity workouts and on race day. Having a measure of your effort level via detailed time series HR data can help with this issue. The “time series HR data-to-EPOC-based” TE measurement provided by Firstbeat and/or the Garmin 610 (and 910XT) is one useful approach. In addition, the real-time measurement of training effect given by the 610 allows the athlete to respond within the workout to attain the goal for the day. As a good friend of mine (and endurance Olympian) says, this “puts an electronic leash” on you as if your coach was right beside you ensuring that you are keeping to the plan.
How well does it work?
Having used this technology for a year now, I have made a reasonable assessment of the effectiveness and accuracy of the provided training effect measure. The most critical part of the use of this technology is to establish an accurate evaluation of your effort levels as a function of HR. Your maximum heart rate and the associated L1-L5 effort levels are the basic inputs to the firmware which will calculate the training effect. Spending time on determining these parameters is critical to obtaining accurate information from the calculation provided by the watch. I conducted a maximum HR determination session by doing 5 X 3 minute maximum effort hill intervals. My maximum HR was in accordance with the Karvonen formula and I found that the Karvonen zone calculations were very close to my perceived effort levels and I have made only minor modifications. So, I will suggest that the Karvonen formula may be a good place to start but do not feel constrained to modification of the zones as you see fit as this will affect the training effect calculation. The other input is your current fitness level according to Firstbeat’s scale of 1-10 where 1 is low fitness and 10 is highest fitness. I have used the highest level as I train everyday and throughout the year. It is important to realize that the training effect will only be informative if you thoroughly and accurately establish your max HR and zone levels. It is well worth the effort and time to do this well.
Although no reference is made by Firstbeat about the details of their algorithm, it seems apparent that they are at least utilizing an integrated total heartbeat measure, time spent in each zone, and the differential of your training session heartbeat function (i.e. recovery rates) as inputs. As such, the algorithm takes about 1-2 weeks of daily training to “learn you” and populate a database that parameterizes your HR activity and recovery rates to calculate an accurate training effect. Once established, the training effect calculation is stable and, at least in my experience, accurate.
Training effect analysis approaches
There are numerous ways to use the training effect data for analysis. Training effect data can be used on a daily basis to ensure that your scheduled workout was at the right intensity level and to provide feedback for any daily training program adjustments needed. Training effect data can also be used to monitor the longer term intensity periodization of your training program, particularly leading up to races. The longer term perspective will give insight as to how your training has progressed and what specific periodization protocols work for you. All of this is important information to have to assist in reaching your ultimate potential.
Often athletes rely heavily on detailed volume accumulation, be it time (such as for Nordic skiers) or distance (such as for ultramarathon runners), with a less detailed understanding of the specific intensity attained for the workouts. You will likely be exposed to athletes referring to consistent 100 mile or 15 hour weeks as a summary of their training regimen. Without the corresponding intensity data, the mileage or time can be quite meaningless as it relates to training progression. The training effect data can very effectively supplement other training data to give a more complete picture of the efficacy of your program.
Runners will often report weekly mileage or, more appropriately, a rolling 7 day (or other relevant period) average value for mileage (although many adhere to a calendar week due primarily to constraints associated with their work schedule). Similarly, a rolling average value for training effect can also be used as a metric and one that is perhaps better suited to providing solid guidance for intensity management and the inevitable needed modifications. It can also provide warning as to potential over-training since consistent elevated training effect for known lower intensity workouts is a direct indication of over-training.
Plotted below are my recorded daily (BLUE) and 7 day rolling average (RED) TE data for about 1 year. Firstbeat defines various ranges of training effect according to the impact on training: 1-1.9 is Easy Recovery, 2.0-2.9 is Maintaining Fitness, 3.0-3.9 is Improving Fitness, 4.0-4.9 is Highly Improving, and 5.0 is Overreaching.
Clearly the greatest majority of my training time has been in the “Improving Fitness” level (L3). According to coaches that I have consulted with, these data indicate that I have spent way too much time in L3 and have not achieved an optimal fitness level for the time spent training (about 750 hrs./yr.). The same data is plotted below as a distribution across training zones where over 40% of the training volume (in terms of time) is in zone 3.
It is suggested that although I have importantly built a considerable base, a more structured program with more extended periods of intensity followed by more extended periods of recovery will lead to higher performance and optimal tapping of my ultimate ability. I am now designing a more structured program to accomplish this with a couple of focus races as goals along the way. Without this intensity data (training effect) neither I or a coach would have been able to make a thorough evaluation of my training nor would I be able to reliably execute upon a structured, periodized program (due to factors outlined in “The importance of perceived effort” above). This makes the training effect data quite valuable to those interested in achieving their potential. This is not to say that utilization of perceived effort is flawed, just that the TE gives a more analytic, repeatable, and likely reliable, assessment of the intensity for a given workout. It is a tool not a requirement.
Further analysis of these data however, reveals a more robust metric: intensity minutes. The training effect-only data is useful from a “zone analysis” perspective for daily monitoring of training effort level, but the training effect data combined with the time spent at this level is actually the most valuable metric as it truly reflects the effort expended. Plotted below is the same training effect dataset combined with the training time data to produce the intensity minute metric (intensity minutes = training effect X time).
A certain level of calibration is necessary to interpret these data and this can be developed over time for each individual. For me, a periodized rolling average of somewhere between 300-500 intensity minutes is typical for my training time allotment (about 2 hrs/day) with a low of about 180 and a high of about 600. Large positive excursions (up to about 1500 intensity minutes) correlate to hard and/or long training sessions or races, the latest one being a 50km tempo run of 5hrs at a training effect of 4.9.
Another way to analyze the data is as a 7 day rolling total intensity minutes as plotted below. This gives an assessment of training load going into a session and can help parameterize “volume” periods from “intensity” periods from “rest” periods in a structured program.
Once again a calibration of these data to personal capability and effectiveness can be developed with time. I appear to be able to “support” extended periods of 7 day rolling totals of 3000+ intensity minutes. This is good input for an individual or their coach in developing training programs as “intensity minute” volumes can be periodized with confidence that the program is, in fact, doable for the athlete. This is particularly important in designing intensity periods, as it is quite frustrating to the athlete when he or she cannot complete or properly execute upon a program due to too much intensity and/or volume.
An interesting correlation that I have found is that my intensity minute values are highly correlated with accumulated vertical ascension. Plotted below is the corresponding 7 day rolling total vertical ascension. This correlation may be due to a particularly inefficient climbing ability and may be identifying an area where I should be focusing in fitness and technique.
Certainly there are numerous other extractable observations from all of this data, some that may provide insight and others that may just confuse. My use of this data has been to develop a detailed structured training plan going forward now that I have re-established a proper training base in the past 1-1.5 years. I will give progress updates on the usefulness of the training effect and intensity minute metrics in the structured program in subsequent posts.
I hope that I have demonstrated the utilization of Training Effect and the derivative metric of intensity minutes as a useful tool for developing, analyzing, and monitoring structured training programs. These data combined with dedicated execution upon a rational training program can help anyone achieve their endurance athletic goals.
This post summarizes a work in process and I am by no means expert in this area. I welcome any critically constructive comments and/or observations.
I have posted and update on this subject as of 21 October 2012.