Identical external load does not mean identical internal responses. External load describes the work the athlete has completed without providing insight into the resultant internal response to that work. [1] External load alone does not determine the training stimulus. The training stimulus is the interaction between the external work performed and the individual’s internal physiological response to that work. [1,2]
The internal psychophysiological response to the work is the stress stimulus to disrupt homeostasis. The physiological and biomechanical responses to the same physical work can vary based on a variety of individual factors: training status, fitness level, recent sleep quality and quantity, nutritional and hydration status, and general life stress, as well as contextual factors such as the psychosocial dynamics and the environmental conditions of the session. Over time, repetition of these stimuli results in the chronic adaptations that underpin changes in training outcomes.
American football combines intermittent high-intensity running and multidirectional movements taking place over a long duration with collisions, static and resistive force production, diverse positional differences, and frequent changes in work-to-rest structure. [3,4,5] Positional tracking data gives you insight on what the athlete did, but it does not tell you how the athlete responded to it.[1,2] That distinction matters when examining an athlete’s stress-response relationship with the goal of mitigating injury risk and maximizing performance.
An accelerometer-derived workload metric or other external load indicators can look similar between athletes while the physiological and biomechanical costs of the session differ.
The goal is not to replace external load monitoring. It is to add an information stream that answers a different question: how is the athlete responding to the external load of their training?
Heart rate (HR) is a complementary lens into a question positional tracking systems were never designed to answer: what cardiovascular response did this external workload evoke in the athlete? And separate from what happens during practice, resting HR and heart rate variability (HRV) are objective response indicators for understanding how an athlete is recovering from and adapting to the demands being placed on them while being influenced by other stressors and factors that are not captured within the training environment. [1,6,7]
Tweet ThisHeart rate is a complementary lens into a question positional tracking systems were never designed to answer: what cardiovascular response did this external workload evoke in the athlete?
@marctlewis

Mind the gap between external and internal load
Internal load describes the relative biological stress imposed on an athlete during training or competition, and the individual physiological responses to the external work. [1]
These factors track reasonably well in more aerobically structured sports. In American football, it’s not that straightforward. The stop-and-go structure of practice, intermittent high-intensity nature of the sport, the positional differences in movement demand, along with the physical contact, static loading, and neuromuscularly demanding actions do not always register meaningfully in positional tracking data. External load metrics can systematically misrepresent the actual physiological cost of training. [3,6,7,8,9,10]
Average HRs are roughly 55–65% of age-predicted maximum during American collegiate football practice. This reflects prolonged recovery periods between plays, the duration of practice, and the time spent in each HR zone. Across positional tracking, HR, and perceptual response data from Division I college football players over 10 in-season weeks, heart rate-derived internal load metrics only showed meaningful correlations with low-intensity movement and total distance. High-speed running showed weak relationships with both heart rate-derived and perceptual (session duration * athlete self-reported rating of perceived exertion) internal load measures.
These results show how these metrics capture different types of training stress. When RPE-based internal load did correlate strongly with external metrics, the relationship was driven almost entirely by session duration, not perceived intensity. [4]
Training stress measures representing both external and internal load serve a distinct purpose. Internal load captured as sessional rating of perceived exertion (sRPE) showed moderate-to-strong positive relationships with most external load metrics. The strongest relationship was with total distance.
Relying on a single metric is not an effective approach in a mixed-model sport like American football, where significant neuromuscular, biomechanical, and cardiovascular demands vary substantially across position groups. [4]
Practitioners need something else to capture the true intermittent cardiovascular demand of the sport. [5,10] Multiple information streams that capture different aspects of athlete training stress and subsequent response take a monitoring program beyond simply relying on positional tracking systems.
The consequences of ignoring internal load in football are not hypothetical. Renaghan and colleagues prospectively followed 30 Division I collegiate football starters across a 13-week in-season period. Both acute and cumulative exposures to high exercise cardiac loads (a validated internal load metric tracking the product of average HR and session duration) were negatively associated with maximum running speed. Athletes exhibiting higher next-day resting HRs showed larger decrements in their maximum running speed, while HRV (rMSSD and SDNN) was positively associated with maximum running speed. [6]
Tracking internal load and next-day cardiac recovery clearly showed that the athletes demonstrating greater autonomic disturbance exhibited greater reductions in maximum running speed.
No single monitoring metric provides a complete picture of the athlete. Rather, confidence in a decision increases when multiple information streams converge to paint a more complete picture of what is occurring. [11,12]
Tweet ThisTracking internal load and next-day cardiac recovery clearly showed that the athletes demonstrating greater autonomic disturbance exhibited greater reductions in maximum running speed.
@marctlewis
Microcycle shapes monitoring
Our integrated athlete monitoring framework is built around four interrelated constructs: training stress, objective response indicators, subjective response indicators, and intrinsic athlete factors. [5] These address four practical questions: What did we ask the athlete to do? How did the athlete respond acutely? How is he responding residually? What individual or environmental factors might help explain the response?

Figure 2. Athlete monitoring conceptual framework to interpret training stress and subsequent athlete response.
Tweet ThisOur integrated athlete monitoring framework is built around four interrelated constructs: training stress, objective response indicators, subjective response indicators, and intrinsic athlete factors.
@marctlewis
HR contributes to two parts of this framework, but it plays a different role in each.
During training, heart rate-derived measures contribute to training stress by describing the cardiovascular component of internal load. Outside of the training session or practice, resting HR and HRV contribute to objective response indicators by describing the athlete’s recovery and systemic residual response. That distinction is important, as HR indices can provide interrelated yet distinct insight into the athlete’s response to training and subsequent adaptive processes.
Rather than interpreting HR in isolation, we anchor the expected response to the tactical periodization of the football week. In the example below, GD-4 is the primary collision / contact stimulus, GD-3 is the largest running and endurance stimulus, GD-2 emphasizes speed and execution, GD-1 is a low-load recovery / readiness day, and Game Day is the performance endpoint.
The exact structure will differ by program, but the principle is consistent. A metric only has meaning relative to what the day was designed to accomplish.

Table 1. Athlete monitoring framework applied across an in-season microcycle of American collegiate football, aligning the intended stimulus and recovery demands of each day.

Table 2. Practice planning model for an in-season microcycle of American football. Intended internal load targets are group-level guidelines and should be interpreted relative to each athlete’s individual rolling normative ranges.
The organizing principle is game-day-matched, within-athlete comparison. For each metric, compare the athlete with their own rolling value for the same relative point in the week. Between-athlete comparisons can identify athletes responding differently to a similar workload, but longitudinal interpretation should remain individualized. [1,7,12]
Training stress: What did the athlete do and how did they respond within the session?
During practice, positional tracking data describes external work while HR data describes the cardiovascular component of internal load. Average HR is useful as a session-level summary and as a measure to monitor live, but the emphasis should be on measures that retain the duration of the exposure.
Training impulse (TRIMP) is a heart rate-derived metric that combines HR intensity with session duration, with different weighting coefficients based on the relative intensity. There are a variety of TRIMP or HR exertion values with different weighting coefficients and relative intensity zones.[2] This is particularly relevant in football because duration is a major driver of the relationship between internal and external load.
A 30-minute conditioning block and a 90-minute practice can show similar TRIMP values and peak HRs but represent very different physiological demands. Average or peak HR data should only be interpreted in the context of how long the athlete was exposed to that intensity and the external load demands placed on the athlete during the session.
In addition, it is important to interpret heart rate-derived internal load by position and practice type. Different playing positions have different HR and movement demands, as do the types of practice. Heart rate exertion is greater in offensive linemen during full-pad, contact practices, with smaller-field concepts, compared to non-contact larger field concept days, as running demands increase.[5,10] There is no universal HR expectation for every position group on every practice day.
Objective response: How did the athlete respond residually and how are they adapting over time?
Resting HR and HRV are collected separately from the session. These help evaluate the athlete’s systemic response to the accumulated demands of training, competition, and life outside the training environment. rMSSD is practical, has demonstrated reliability in different sporting populations, and is strongly associated with vagal tone, making it an ideal indicator of parasympathetic activity and recovery. [6,7,8]
Standardization matters. Ideally, we are working with data from first thing in the morning, at a consistent time and under consistent conditions, including before physical activity and caffeine.
An individual rolling baseline allows deviations to surface, which convey more information than any single value. A modest elevation in resting HR after a high-load day can be an expected training response (see Table 2). Meanwhile, a larger deviation, especially when accompanied by suppressed HRV, increased soreness, or other wellness changes, warrants investigation.

Time spent above 150 bpm during American football training is associated with elevated next-day resting HR and impaired HR recovery with stronger associations at higher intensity thresholds.
That does not mean >150 bpm is a universal red flag. Repeated exposure above that intensity should influence what we expect to see the following morning. [6]
Heat is an important modifier. Greater exposure to elevated core temperatures is associated with reduced autonomic recovery the following day. Environmental conditions are best considered as part of the training stimulus rather than as background information. This is especially important during preseason and early in-season time periods, as well as in geographical areas with extremely hot and humid conditions. [8]
Tweet ThisrMSSD is practical, has demonstrated reliability in different sporting populations, and is strongly associated with vagal tone, making it an ideal indicator of parasympathetic activity and recovery.
@marctlewis
Subjective response: What is the athlete telling us and how does it align with objective data?
Objective data provides distinct insight into the athlete’s response. But it does not measure all constructs relevant to understanding the athlete’s response and the subsequent adaptive processes.
One low-cost, practical tool to build the subjective internal picture is a wellness questionnaire. Practitioners can adapt existing, validated questionnaires to their specific contexts or purposes. Key factors to look for in a wellness questionnaire are sleep quality, sleep quantity, general muscle soreness, any specific muscle soreness, general stress, mood state, and general fatigue. [13] These measures provide insight into actual physiological response.
There are also questions capturing insight into non-training factors that influence the stress-response relationship.
This information is especially useful to gain insight on different training stresses not measured well within the sport, or when there is disagreement between the external and internal load.
For example, an athlete may have a normal external workload and cardiovascular response, but elevated soreness after a collision-heavy practice or a small-field practice with increased change-of-direction demands, indicating increased tissue damage. Meanwhile, another may have elevated resting heart rate after a practice with a normal external workload because of poor sleep, illness, academic stress, or travel.
The subjective data help determine whether the response is likely driven by training, or whether other non-training related factors could be negatively influencing athlete response and recovery.[1,2,13,14]
Intrinsic factors: What individual or environmental factors might explain the response?
The fourth construct is the context around the athlete. This includes certain non-modifiable characteristics such as age, injury history and recency, injury location and severity relative to their job demand. There are also longer-term modifiable factors such as current fitness level or body composition, and short-term modifiable factors such as loading status.
These are not just variables to add to a dashboard. They should anchor the interpretation of other monitoring data. Practitioners can use them as potential mediators and moderators of the observed stress-response relationship. [1,2,5,15]
Understanding the picture that emerges from internal and external load
The value of the framework is organizing the purposeful collection of interrelated, but distinct, information streams that allow for more informed decisions.
Same external load, different cardiovascular response
Two linebackers complete a similar GD-4 practice. Their positional tracking reports show similar workloads, high-speed running distances, acceleration counts, and collision exposures.
Athlete A records a cardiovascular response close to his individual GD-4 normative range. Meanwhile, Athlete B records a substantially higher average heart rate and TRIMP value compared to his own GD-4 normative ranges, including considerably more time in higher heart rate zones.
These athletes experienced a similar external stimulus but different internal responses.
Instead of automatically modifying Athlete B’s training, carry the elevated internal load forward into GD-3 decision-making. The next morning, look at his resting heart rate and HRV while conducting a targeted athlete check-in for subjective data. If those signals remain normal, the athlete is tolerating the increased physiological demand well. However, if several response signals converge, illustrating the external load resulted in a more significant stress response than normal, the case for intervening and modifying GD-3 becomes stronger, dependent upon other contextual factors.

High external load, normal internal response
An offensive skill-position player completes a high-load GD-3 practice. Their positional tracking report shows a meaningful increase in workload, including meaningful increases in high-speed running distance and acceleration counts. However, the athlete’s heart rate response is close to his individual GD-3 range. The next morning’s resting heart rate and HRV remain within his normal GD-2 ranges, and he does not report any increases in muscle soreness or fatigue on the GD-2 wellness questionnaire.
High external load is not automatically excessive load. Both objective and subjective data support the conclusion that the athlete may be tolerating the stimulus well. The increase in external load suggests the athlete is experiencing increases in general workload and in specific neuromuscularly demanding work.
The convergence of data that provides insight into the response of both metabolic and biomechanical stress allows for this interpretation.
Tweet ThisThe convergence of data that provides insight into the response to both metabolic and biomechanical stress allows for this interpretation.
@marctlewis
Expected HR response vs. unexpected HR response
The same heart rate response can mean different things on different days. A high HR response on GD-3 may be expected because the day is designed to create a cardiovascular stimulus. The same response on GD-2 is not desirable because that day emphasizes brief, high-quality speed and power outputs with long recovery periods and limited cardiovascular load to minimize metabolic stress.
The target is not always to minimize the internal response. If a player experiences an increased physical workload on a practice day that is designed to create a more significant cardiovascular stimulus, it can be effective in enhancing cardiovascular fitness.
The goal is to align the stimulus to maintain or enhance physical capacities, while maximizing game-day physical readiness.
On GD-3, a meaningful cardiovascular response can be part of the plan. Meanwhile, on GD-2, excessive cardiovascular responses may indicate that the practice has become too dense, too long, or too physiologically demanding for the intended objective.
Morning after a hard practice
An athlete experiences a meaningful increase in GD-4 workload and high-intensity actions and subsequently demonstrates altered resting HR the next morning. In terms of HRV, rMSSD is modestly suppressed, but subjective indicators such as soreness and fatigue are within normal ranges and the athlete reports feeling great. This is different from an athlete with an altered resting heart rate who also shows a meaningful rMSSD suppression, increased soreness, reports poor sleep, and poor perceived recovery.
The first athlete is showing a response that may be consistent with the intended training effect. The second has multiple information streams indicating that he has not recovered as expected. Any decisions the performance staff subsequently makes are therefore not triggered by one metric, but rather convergence and context.
Heat as a load modifier
Two practices can have the same planned external load but very different physiological consequences due to different environmental conditions. [8]
Tracking environmental factors lets us interpret the stress-response relationship within the appropriate context. For example, if a moderate GD-4 session occurs in extreme heat and the next morning shows disproportionate increases in resting HR and reductions in HRV across several athletes, the response may reflect an environmental load modifier rather than poor practice design alone.
The intervention may not be to reduce the planned football stimulus but to modify the environment by shortening outdoor exposure, lengthening recovery periods, moving selected periods indoors, adjusting equipment, increasing recovery support, or a combination of these. The monitoring data help identify that the problem is not necessarily the external football workload itself.
Guidelines for combining internal and external monitoring
Practitioners should decide what they want to know before choosing the specific tool and output measures or variables. Establish the intended stimulus first so that the practice or training plan creates the expectation. Monitoring tells you whether the athlete experienced what you intended.
Within-athlete comparisons, individual rolling standard deviations, individual game-day matched means, and individual coefficients of variation for a given metric, are more useful for identifying change and for longitudinal monitoring than population thresholds applied equally to every athlete. Look at the distribution of the data within each metric for each athlete, while identifying trends to better understand singular time point changes within the overall distribution of an individual’s data.
Heart rate indices do not always behave as expected: interpret them with caution. For example, a trend towards a lower resting heart rate can be an indicator of increased fitness in one athlete, while being an indicator of overreaching in another.
Examine each heart rate measure in the context of all the other information streams and leverage longer-term trends to gain insight into athlete response.
A single abnormal metric is worth investigating but doesn’t automatically result in an intervention, nor does it need to be factored into decision-making. Multiple independent signals pointing in the same direction increase confidence in a decision. Look for convergence before intervening.
Similarly, treat divergence as information. When external load, internal load, and other response indicators do not agree, do not force them into agreement. Often, the disagreement allows for more and better questions and conversations, potentially resulting in valuable insight.
Tweet ThisWithin-athlete comparisons, individual rolling standard deviations, individual game-day matched means, and individual coefficients of variation for a given metric, are more useful for identifying change and for longitudinal monitoring.
@marctlewis

