FormalPara Key Points

This study describes training session models across intensities applied by world-leading coaches in endurance sports.

Training session models vary substantially across sports, mainly due to load tolerance considerations for the locomotion modality, seasonal circumstances, and sport-specific demands.

The interval training session models outlined here are more voluminous, more controlled, and less exhaustive than recommendations from many published intervention studies.

1 Introduction

Numerous studies published over the last ~ 25 years have quantified the training characteristics of elite endurance athletes, in which annual training volumes range from ~ 500 to ~ 1200 h per year, distributed across 300–600 training sessions [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29]. This large variation among equally successful performers is mainly explained by modality-specific constraints (e.g., weight-bearing versus nonweight bearing sports, type of muscle action involved, cycle/muscle-contraction time, and leg-dominant versus whole-body exercise), although individual predispositions also matter [27]. Depending on the specific quantification approach (i.e., categorizing the distribution of whole sessions versus minute-for-minute time in zone), about 80–90% of endurance training is performed at low intensity (below the first lactate or ventilatory turn point), while the remaining 10–20% is performed at higher intensity [2, 7, 14, 18, 22]. While the interaction between training volume and intensity distribution is well described at an annual and monthly level in rowing [1,2,3,4,5], cross-country (XC) skiing [7, 8, 24,25,26], road cycling [9,10,11,12], long-distance running [13,14,15,16, 20,21,22,23], swimming [17,18,19], and triathlon [27,28,29], corresponding information for prescription and execution of individual endurance training sessions are sparse.

Training prescription for continuous exercise sessions include selection of exercise modality, working duration, and intensity, while interval training also encompasses manipulation of number of repetitions, number of series, relief interval intensity and duration, and the between-series recovery duration and intensity [31]. Ironically, interval training sessions are often described in more detail compared with continuous exercise training [2, 20, 31], although the latter training form by far constitutes the largest proportion of training in elite endurance athletes.

Session model comparisons across studies and sports are complicated because of inconsistent methodological frameworks (e.g., intensity zones) and terminology. Moreover, most of the research within this topic has been conducted on well-trained but nonelite volunteers who perform training sessions that may not be consistent with what “elite” endurance athletes perform [31]. Detailed training session descriptions have been presented for world-leading long-distance runners and XC skiers [20,21,22,23,24,25,26]. These studies show large between-sport differences in duration for low-intensity training (LIT), and partly also for moderate-intensity training (MIT) sessions, while the summated duration for high-intensity training (HIT) sessions appear more consistent. Corresponding training session descriptions for sports such as road cycling, swimming, triathlon, rowing, biathlon, and speed skating are clearly underrepresented or missing. Due to the large variations in annual training volume across endurance sports [30], it is reasonable to expect large variations in training session prescriptions as well.

Indeed, more research is needed to improve our understanding of session model features among elite endurance sports. The very best practitioners are often years ahead of sport science in integrating the critical features of training [2, 16, 22], and experienced coaches who have achieved success with multiple athletes over time are likely best capable to describe good session designs and identify factors ensuring high training consistency and quality [32, 33]. Surprisingly, the practices, knowledge and experience of the very best endurance sport coaches have received minimal attention in research literature.

Norway has been one of the world-leading sport nations per capita in the last two to three decades [34], with most Olympic and World Championships medals won in endurance sports such as XC skiing, biathlon, speed skating, rowing, cycling, swimming, long-distance running, and triathlon. One of the advantages of the Norwegian system is that endurance sports use the same framework for defining training content, facilitating valid comparisons across sports. By using successful Norwegian endurance coaches as key informants, the aim of this study is to describe and compare best practice session models across different training intensities in Olympic endurance sports. Within this context, we define endurance sports as disciplines with ≥ 6 min competition duration with an aerobic energy contribution of ≥  85%.

2 Methods

2.1 Study Design

This study is a part of a larger project investigating successful coaches in Olympic endurance sports, where the overall aim is to gain comprehensive insights regarding the holistic training philosophies and practices at the macro-, meso- and micro-level. For the current study, a pragmatic multiple case study design was used to investigate best practice session models successfully used to attain world-class performance in Olympic endurance sports. To investigate the complexity and capture sport-specific dimensions and perspectives, the following cases were selected: XC skiing, biathlon, swimming, long-distance running, long-track speed skating (hereafter referred to as speed skating), rowing, road cycling, and triathlon. To allow for comparison and contrast across sports, all cases were selected within Norway, assuming similar culture and context. Some of the most successful and experienced coaches were chosen as key informants.

2.2 Participants

Twelve male Norwegian coaches participated in this study. They were all currently or previously responsible for the training of world-class endurance athletes who altogether have won more than 370 Olympic, World, and European Championship medals, mainly with Norwegian athletes. All coaches had experience of coaching both males and females. Two coaches were involved in XC skiing, biathlon, swimming, triathlon, and long-distance running, while one coach was involved in speed skating, rowing, and road cycling. One informant coached both swimming and triathlon athletes. Annual training volume measures prescribed by the coaches in these sports are presented in Table 1. All the coaches provided a written informed consent to participate prior to the study and approved the manuscript prior to submission. The Regional Committee for Medical and Health Research Ethics waived the requirement for ethical approval for this study and the ethics of the project was performed according to the institutional requirements at the School of Health Sciences, Kristiania University College. Approval for data security and handling was obtained from the Norwegian Centre for Research Data (reference no. 605672).

Table 1 Annual training volume measures (range) across the analyzed endurance sports
Full size table

2.3 Procedures

Inspired by the key informant technique in ethnographic research, a pragmatic four-step procedure was used to collect and quality-assure comprehensive information on best practice regarding key session models across different endurance sports:

  1. 1.

    Initially, an extensive questionnaire related to planning, conducting, and evaluation of training at the macro, meso, micro, and session level was administered to all coaches.

  2. 2.

    The next step consisted of quality-assurance of data through conversations with the coaches and cross-referencing with historically reported training logs from some of their most successful athletes.

  3. 3.

    Thereafter, a semistructured interview was conducted with each coach by the first and second authors to obtain supplementary information related to the qualitative aspects of session models among elite endurance sports. Within this context, training quality is defined as the degree of excellence related to how the training process or training sessions are executed to optimize adaptations and/or improve overall performance [33]. Each interview lasted approximately 180 min, of which about one-third was directly related to this study. The interviews were audio recorded and transcribed. Formal translation and back translation from Norwegian to English were performed by the first and third author, respectively.

  4. 4.

    During analysis, we involved the coaches in an extensive review process and follow-up interviews to clarify and ensure that the findings reflected their perspectives on best practice training sessions as accurately as possible.

In terms of endurance training intensity quantification, a six-zone scale developed by the Norwegian Top Sport Centre (Table 2) was applied. Modified versions of this scale have been used in several previous studies [35,36,37]. Training zone determination during practice was determined by the coaches based on a holistic interpretation of all the included metrics listed in Table 2. Moreover, a modified session goal approach based on Sylta et al. [36] was employed. That is, training intensity distribution was described in terms of the categorical distribution of prescribed training sessions across intensity zones based on their execution. In comparison with a time-in-zone approach that will overemphasize LIT, this method presents a more representative picture of MIT and LIT prescription within long-term programming. Here, only the main part of the session was considered, while warmup and cool down were excluded. We use the term “accumulated work duration” (AWD) for interval sessions, and this is defined as the summated duration of the work bouts only.

Table 2 Intensity scale for elite endurance athletes
Full size table

For the purpose of this study, cross training was defined as endurance training in a nonspecific mode. Treadmill running (including antigravity treadmill running), roller skiing, roller skating, ergometer rowing, and indoor cycling were considered specific (i.e., not cross training) for runners, cross-country skiers/biathletes, speed skaters, rowers, and cyclists, respectively.

2.4 Analyses

Numerical information on training session organization across intensity zones was systematized in Microsoft Excel (Microsoft Corporation, Redmond, WA) for descriptive presentations. Thereafter, this information was compared with information from training diaries of successful athletes in the respective sports and calibrated among the authors and coaches.

To identify similarities and differences within and across endurance sports, summaries of common session-model features across endurance sports and sport-specific features related to planning and execution of training sessions were outlined.

3 Results

Table 2 presents commonly applied training session models across intensity zones among Norwegian world-leading athletes in Olympic endurance sports, while Table 3 provides an overview of how the loading factors are typically organized across intensity zones within the same sports. Overall, LIT sessions account for approximately 75–80% of all sessions. These are dominated by continuous exercise, although swimming, rowing, speed skating, and road cycling also apply low-intensive intervals. Athletes from all disciplines surveyed performed the vast majority of LIT sessions in Z1 and only to a limited extent in Z2. The duration of typical continuous LIT sessions spans from 30 to 420 min across sports, with long-distance running on the lower end and road cycling on the upper end of the scale. Long-distance running, road cycling, swimming and triathlon perform most LIT sessions in the specific modalities, while nonspecific cross-training accounts for a considerable proportion of total LIT sessions in speed skating, rowing, biathlon, and XC skiing.

Table 3 Commonly applied training session models across intensity zones in Norwegian world-leading endurance athletes
Full size table

MIT sessions (i.e., Z3) account for approximately 10–15% of all sessions across the annual cycle. These are mainly performed as intervals, although with large sport-specific variations. AWD is in the range of 20–90 min, and interval times are mainly in the range 5–20 min, while work-to-rest ratio is mainly in the range 6–4:1 (Tables 3 and 4). Biathlon, XC skiing, road cycling, and swimming also apply continuous work in Z3, with accumulated work duration in the range 40–60 min. In several sports, particularly road cycling, competitions account for a considerable part of the overall Z3 volume.

Table 4 Loading factor organization in typical training sessions across intensity zones
Full size table

HIT sessions comprise about 5–10% of all sessions and are mainly conducted as intervals and competitions in all sports. AWD is in the range 15–50 min for Z4, 10–30 min for Z5, and 3–15 min for Z6/7 (warmup and cool down not included), while work bout durations are in the range 1–10 min, from 30 s to 7 min, and from 20 s to 3 min, respectively (Table 3 and 4). Work-to-rest ratio is in the range 3–2:1 for Z4, 2–1:1 for Z5, and 1–0.1:1 for Z6/7, but sport-specific differences are clearly present. Competitions account for a considerable part of Z4/5 work among elite athletes, as most sports have at least 15–20 competition days per year (Table 1).

Most coaches reported applying a limited set of session models within each zone for week-to-week calibration/control of performance development. Within these sessions, several common features were identified across endurance sports that are described in detail in Table 5. These include the application of hard–easy rhythmicity, few but well-known session models, lactate measurements for intensity control, limited use of all-out endurance training sessions, mixing intensity zones within sessions, slight progressive intensity increases throughout the hard session(s), adjustments of session models during altitude training, and a preference towards passive instead of active recoveries during interval sessions. Moreover, some common characteristics related to the coaches’ focus before, during, and after the sessions are also present (Table 5).

Table 5 Common session-model features across endurance sports
Full size table

Although several consistent approaches were observed in terms of training session organization and implementation, some sport-specific features were identified related to planning and execution of training sessions (Table 6). Differences in session models across sports are mainly explained by competition-specific demands, seasonal considerations, logistic factors, movement constraints for the modality, and associated load-tolerance considerations.

Table 6 Sport-specific features related to planning and execution of training sessions
Full size table

4 Discussion

This is the first study to describe and compare training session models across intensities and endurance sports. The duration of LIT sessions varies substantially across sports, ranging from 30 min to 7 h, mainly due to modality-specific constraints and load tolerance considerations, while MIT and HIT sessions differ less across sports and are mainly conducted as intervals (or competitions) in specific modalities. Overall, both MIT and HIT interval sessions are characterized by a high AWD, a progressive increase in intensity throughout the session, and a controlled rather than exhaustive execution approach. In the following paragraphs, we will discuss the quantitative and qualitative aspects of these session models and potential underlying mechanisms in more detail.

This study clearly demonstrates that LIT is the most prescribed type of training session in elite endurance sport, in line with previous studies based on quantification of training performed [2, 7, 13, 14, 18, 22, 38]. Most LIT sessions are prescribed and executed in Z1, interspersed with Z2 once-to-twice per week, or as part of progressive Z1-sessions. This distinct feature would not have been detected by the commonly applied three-zone scale (LIT, MIT, and HIT), emphasizing the advantage of a more categorized scale (e.g., a six-zone scale as in this study). In this context, it is important to emphasize that elite endurance athletes have a broad intensity range below the first lactate turn point compared with recreational and moderately trained performers. This makes the potential range of intensity and duration combinations within LIT larger for elite athletes and a more important programming consideration for their coaches. One might speculate that Z2 training costs too much for elite endurance athletes, making them less recovered and poorer prepared for the subsequent intensive sessions. According to several of the present coaches, Z2 training is mainly implemented for technical reasons, as an effective force signature in the movement cycle sometimes requires a minimum speed or power output. Some of the analyzed sports perform LIT as long intervals to provide short intermissions for nutrition/fueling, technical feedback, and lactate measurements. However, since the latter intervals are quite long and the intermittent recoveries are relatively short, such sessions practically act as continuous LIT sessions from a perceptual and physiological perspective. Similarly, terrain variations in sports such as biathlon and XC-skiing make LIT sessions more stochastic [39, 40]. The prevailing notion is that most LIT sessions must be sufficiently easy to ensure that the subsequent hard sessions can be conducted with sufficient quality. LIT sessions have misguidedly been termed “recovery workouts” by several practitioners over the years [22], suggesting that these sessions do not elicit adaptations themselves but rather “accelerate” recovery prior to the next hard session. We argue that this interpretation is erroneous for two important reasons. First, the concept of any form of recovery acceleration from an intervening workout lacks support in the scientific literature, although the “low” load of such sessions likely causes limited interference with the ongoing recovery process. Second, frequent and voluminous LIT is considered an important stimulus for inducing peripheral aerobic adaptations [41] and improving work economy [42, 43]. At least three adaptive signaling pathways (through which exercise of different intensities and durations can impact protein composition of working muscle over time) have well-demonstrated signaling roles, mediated through specific kinases, and aggregated by the PGC1a gene [44]. The pathways triggered by high energy phosphate depletion (i.e., large reductions in ATP/AMP ratio) and by elevated production of reactive oxygen and nitrogen species both show rapidly evolving feedback inhibition of signal amplitude as signal mediated adaptations occur [45,46,47]. That is, adaptive feedback inhibition reduces the adaptive return from these pathways with repeated HIT bouts over time. In contrast, it appears that elevated intracellular calcium concentration associated with the excitation-contraction coupling process remain responsive across longer training time frames due to greater potential for modulation via exercise duration × intensity interaction, with essentially no feedback inhibition of the primary signal within and across motor units. Accordingly, it may seem that years of accumulated wisdom among elite coaches is consistent with how different signaling pathways coalesce to determine the overall adaptive enrichment of the endurance phenotype.

Notably, AWD for LIT sessions varies markedly, both within and across sports. Within-sport differences are mainly explained by session purpose (i.e., extra-long versus short, long-slow distance), while between-sport differences were explained by competition-specific demands, movement constraints for the modality and associated load-tolerance considerations. The latter aspect is in line with Sandbakk et al. [30], who recently developed a theoretical framework for the impact of physiological and biomechanical mechanisms associated with different locomotion modalities on training load management in endurance exercise. According to their theory, the combination of weight-bearing exercise and rapid plyometric power production in long-distance running puts high loads on muscles and tendons during each step, likely explaining why the duration of LIT sessions in long-distance running is relatively low compared to most other endurance sports. However, elite runners seem to compensate for this “low” volume by training twice a day and performing some of the LIT sessions in the upper range of Z1, sometimes approaching Z2 [22].

Speed skating is also muscularly demanding but for other reasons. The small angles in the hip and knee, in addition to the static upper body position and long duty cycle of an effective skating stroke, together induce intermittent blood-flow restrictions in the working muscles [6, 48]. Hence, speed skaters typically prefer cycling instead of the skating-specific modality during LIT and MIT sessions, as well as for warmups and cool downs. Nils van der Poel, double gold medalist in the 2022 Beijing Winter Olympics, followed this approach to the extreme with 6–7 h rides on the bike five times per week during preparation training and considerable amounts of MIT cycling in the subsequent phase [49].

Road cyclists perform longer but fewer training sessions compared with the other sports. The preference for and tolerance of voluminous road cycling sessions can mainly be explained by the concentric only and nonweight bearing loading, the long-duration competition format, and the fact that cyclists draft behind teammates/competitors and coast downhill in substantial parts of the sessions. Careful examination of elite cyclists reveals that 10–20% of all cycling sessions are spent at a power output < 0.75 W/kg [10]. While a runner absorbs a huge mechanical load when running downhill, a cyclist coasting downhill is normally resting the active musculature.

Swimming also involves nonweight-bearing exercise and low contraction velocity movement [30]. Swimmers perform shorter LIT sessions than road cyclists. To obtain a relatively high training volume, these athletes seem to compensate by consistently swimming twice a day, with the first session performed in the early morning. This approach can at least partly be explained by restricted access to swimming halls in the middle of the day, as school swimming is generally prioritized by local authorities. Interestingly, in contrast to their high-volume training, most swimming events are dramatically shorter in duration compared with road cycling and most other traditional endurance sport events.

Rowing also involves nonweight-bearing and low contraction velocity movement, but the injury risk of overloaded back and ribs, particularly with modern “cleaver” rowing blades, has led rowers to implement a larger proportion of LIT as cross-training [30]. In XC skiing and biathlon, the athletes distribute training time across varying sub-techniques while skiing on snow or using roller skis [7, 8, 25, 26]. However, the best athletes do not perform longer LIT sessions than cyclists, rowers, or swimmers. This can be explained by the moderately high muscular loads of skiing uphill, in addition to the strong focus on and accompanying strain associated with maintaining effective technique (and appropriate switching between multiple subtechniques) in complex movements [30].

This study shows that many of the best practitioners within endurance sports supplement their LIT sessions in the specific modalities with cross-training, in line with previous studies [1, 7, 8, 25]. The application of cross-training differs substantially across sports, not only for movement constraints and associated load management but also for seasonal reasons. Because of the limited access to snow during the summer, XC skiers and biathletes perform many running and cycling sessions, respectively. Likewise, rowers execute numerous land-based sessions as running, cycling, or XC skiing (perhaps a distinctly Norwegian cross-training modality) during the winter. Other supporting arguments for cross-training in research literature include injury prevention, general central capacity effects and prevention of training monotony [50, 51]. A plausible question within this context is whether long-distance runners should compensate for their “low” volume (compared with the other analyzed sports) by adding more cross-training sessions to maximize the training stimulus with lower muscular-mechanical load. However, a common notion among the interviewed coaches was that cross-training modality must bear sufficient physiological and mechanical resemblances to the specific demands to maximize the odds for positive adaptations (Table 5), in line with the principle of specificity [52]. Alternative locomotion modalities for runners (e.g., cycling and XC skiing) are less used (in most cases limited to injury rehabilitation processes) and may be too removed from the specific demands, increasing the odds for maladaptations. Running is also unique among endurance sports in that cycle frequency/cadence does not and cannot be manipulated very much across a broad range of intensities/speeds. More specifically, the cadence may only increase 10% from LIT to HIT for a distance runner. For a rower or kayak paddler, cadence can vary at least twofold from Z1 to Z5, with the force signature maintained relatively stable. These issues may partly explain why cross-training in long-distance running mainly is restricted to injury rehabilitation processes. Overall, the underlying mechanism of cross-training remains poorly understood, and future longitudinal studies should aim to explore the training transfer efficiency of varying types of cross-training.

Based on the large variations in LIT session duration across the analyzed sports in this study, it is reasonable to question well-established training load assessment tools such as training impulse (TRIMP) and session rating of perceived exertion (session RPE). While these concepts only take training volume and intensity into account [53,54,55], it seems clear that the choice of exercise modality influences effort beyond commonly applied external and internal load measurements. We argue that these methods are not valid for comparisons of training load across exercise modalities, for example, by comparing sports or when comparing the load across different modalities. Foster et al. [56] have also indicated that session RPE is mode dependent, but more studies are warranted to verify this feature.

Intensive sessions in the form of MIT and HIT are considered fundamental for performance progression by all the participating coaches in this study, and the planning and implementation of training are mainly centered around such key sessions. Most MIT sessions are performed as intervals, although several sports also apply continuous work. Competitions account for parts of MIT or as elements of LIT, and the stochasticity of many competition formats such as cycling and XC skiing results in some intensity undulation. A specific feature for triathlon is the application of combined modalities (so-called brick workouts), where swimming/cycling or cycling/running are frequently applied during MIT sessions to manage modality transitions. Overall, the analyzed sports implement considerably more MIT than HIT sessions across the annual cycle. This strategy has been a part of the training philosophy in several Norwegian endurance sports over the last two decades [57]. Here, a fundamental feature is the application of double threshold sessions (i.e., both morning and afternoon) twice a week, with blood lactate concentrations in the range 2–4.5 mmol/L. Marius Bakken, a former Norwegian 5000 m record holder (13:06 min) is considered the originator of this concept, and he has argued that Z3 intervals (particularly microintervals lasting only 45–60 s) allow for accumulation of work at faster and more race relevant running speeds than continuous training in the same lactate-based zone, without the negative consequences of HIT in the form of fatigue and subsequent recovery [57]. Half of the coaches in this study and numerous elite coaches worldwide have adopted double threshold sessions in their weekly preparation training, representing a novelty in the current training of elite endurance runners.

The HIT sessions presented in this study are mainly conducted as intervals, although competitions constitute a substantial part of most sports. In road cycling, elements of HIT are conducted during LIT sessions. In triathlon, Z4 sessions are often conducted as brick workouts. Overall, a common and logical trend across all sports is that interval times and accumulated working duration for interval sessions decrease with increasing intensity. Recovery time between intervals depends on interval time and intensity, but we observed a clear trend towards lower work-to-rest ratio with increasing intensity. The variations in MIT and HIT session design across sports can mainly be explained by corresponding movement constraints and load management considerations as explained previously for LIT sessions, although the differences between sports diminish with increasing training intensity.

Ever since the first studies on interval training were published in the 1960s [58, 59], a plethora of research has been devoted to this topic. Interestingly, the best practice interval sessions described in the present paper differ considerably from most of the models tested in previous intervention studies that are the building blocks of current established scientific recommendations [20, 31, 60]. First, our analyzed sports perform interval sessions across a considerably wider intensity range compared with research literature. Comprehensive and highly cited review papers recommend athletes to reach at least 90% of their maximal oxygen uptake during interval sessions (or ≥ 95% of the minimal velocity/power that elicits maximal oxygen uptake) to elicit both maximal cardiovascular and peripheral adaptations [20, 31, 60]. Secondly, AWD for interval sessions is also considerably lower in most scientific studies [31, 60] than those presented here. Fundamentally, elite coaches use AWD to adjust both stimulus and progression “between” intensity adjustments (stair-step model) far more than most recreational athletes and researchers, who emphasize mainly intensity as a “lever” for managing the HIT prescription. Third, the observed trend towards lower work-to-rest ratio with increasing intensity has not previously been established in scientific studies. However, the predominant application of passive recoveries is in line with recommendations from research literature, as active recoveries can lower muscle oxygenation, impair phosphocreatine resynthesis and, thereby, trigger anaerobic system engagement during the following effort [31].

Another notable finding from this study is that very few interval sessions are performed to the point of power or pace “failure.” Instead, these sessions are characterized by an even pacing across bouts or even a small but progressive increase in intensity (crossing through, for example, upper Z3 to upper Z4), a semiexhausting effort and high AWD. Importantly, maintaining good technique (i.e., avoiding technical collapse and “floundering” near the end of work bouts) is emphasized. Some intervention studies have applied intervals with maximal sustainable work intensity, aiming to achieve the highest possible average speed or power (so called “maximum session effort” or isoeffort approach) [61,62,63]. The interviewed coaches argue that such an all-out approach is not sustainable over time for several reasons. In a short-term perspective, an all-out session execution approach can lead to an undesired and poorly timed peaking response (provided that the recoveries between such hard sessions are sufficient). In a long-time perspective, an all-out approach limits the accumulated load of MIT and HIT due to shorter work time in single sessions and longer recovery time after sessions. Concurrently, this increases the odds for overtraining and burnout due to the physical and mental strain associated with such sessions. The best practitioners are, therefore, especially cautious not to overuse all-out intensive sessions or introduce them too early in the annual cycle [7, 16, 22, 25, 35], a notion in line with traditional periodization thinking [64]. Alternatively, controlled and semiexhausting interval sessions may effectively stimulate adaptation through the interaction between high intensity and larger accumulated work that can be achieved before the onset of fatigue, compared with an all-out approach [61, 65,66,67].

All the key informant coaches in this study consider training quality highly important for performance development. Here, “quality” is not synonymous with “intensity” as often seen in popular science literature. Instead, training quality is defined as the degree of excellence related to how the training sessions are executed to optimize adaptations and/or improve overall performance [33, 69]. This includes the ability to optimize processes that affect the execution of training sessions in relation to the intended purpose. Intensity discipline in relation to the training prescription is an example of this quality emphasis observed at the elite level. Training quality can be developed and fine-tuned over time through optimal application of monitoring tools and good communication among the athlete, coach, and supporting staff. Obtained information related to readiness, exercise load, and recovery state form a basis for subsequent decision making [69], and this is continuously subject to improvement through a circular learning process where planning, execution and debriefing/evaluation are the fundamental stages [33, 68]. The present coaches describe a culture of continuous learning and development through constructive interactions with the athletes.

Although this study has described a variety of session models across sports, the best practitioners tend to apply a limited set of session models within each zone. In this way, each key session acts as a test where heart rate, blood lactate concentration, speed/power output, and perceived fatigue/exertion can be compared from week to week. The principle of control is a fundamental feature of elite sport to determine whether athletes adapt to the training, identify individual responses, monitor fatigue and accompanying need for recovery, and minimize the probability of nonfunctional overreaching, illness and injury [55, 70]. It is also reasonable to assume that implementation of well-known sessions increases the likelihood for increased training quality. Interestingly, all coaches have primarily adjusted their training session models to the individual athlete and sport-specific demands, rather than based solely on sex, as previously described more generally [71].

A distinct feature across all the analyzed sports in this study is the alternating rhythmicity of hard and easy workouts. The legendary track and field coach Bill Bowerman popularized this concept in the 1960s [22]. This was also a fundamental feature of Matveyev’s traditional periodization model founded at the same time [64], with strong links to the principle of stimulus and response (also known as the overcompensation or training adaptation principle). That is, training stress leads to acute fatigue and damage to physiological structures, and during the subsequent restitution phase, the organism does not only return to the original condition but overcompensates to be better prepared for the next stress. The larger the training stress, the longer the restitution time required [72]. Importantly, many of the coaches regard the training day (not only each session) as the unit of stress being managed. Therefore, amplifying the intensive load during a planned “high stress” training day is more sustainable than adding an additional high stress training day to the microcycle. MIT and HIT both induce high stress, particularly given the high absolute intensities and associated metabolic flux of elite performers, combined with the AWD that is prescribed and executed. The present study, together with other recently published studies, shows that consecutive hard training days rarely occur and that the hard–easy rhythmicity also holds true for today’s elite endurance practitioners [7, 16, 22, 25, 35]. Hard and easy sessions seem to stimulate a complex set of overlapping and complementary adaptations [73, 74], justifying the systematic training intensity variation for performance development in endurance sports. Overall, we would argue that elite coaches use this day-to-day rhythmicity to carefully manage and, to a substantial extent, “polarize” training stress, not work intensity, to ensure that recovery is achieved. Elite coaching is about managing the systemic cost of maintaining a high training frequency and volume, and thereby a high adaptive signal. Across sports, the success of these elite coaches is quantified in terms of long-term thinking and “staying healthy and being able to do the work required for success.”

Some study limitations should be acknowledged. First, it is likely that the present results are influenced by a Norwegian “group culture” bias, and other roads may also lead to Rome. Although the key informants in this study have coached numerous world-leading athletes, they have also applied the same training system to several other less successful athletes. Moreover, the intensity scale outlined here (Table 2) has been used by Norwegian elite endurance athletes over the last two decades. Previous studies have presented several arguments to explain why standardized intensity zone systems are imperfect tools [2, 16, 22, 37]. Slight inconsistencies in AWD within the same zone can be observed when comparing present findings with previous studies [16, 22]. Inconsistencies across studies are expected because (1) the intensive zones are “narrow” (i.e., small differences in heart rate, blood lactate and RPE), and (2) MIT/HIT sessions tend to overlap intensity zones. Although intensity scales can be criticized for several reasons, we argue that the potential error sources are outweighed by the improved communication among practitioners that a common scale facilitates.

Finally, we do not believe Norwegian athletes are “physiologically biased” towards a genotype that is uniquely responsive to the training characteristics described here. More likely, the Norwegian endurance sport success is grounded on a culture which includes an appreciation for “endurance.” Norwegian champions often describe an active childhood that included lots of hiking, skiing, cycling, etc., just as a function of living. So, relative to the size of the country, we could argue that a large fraction of Norwegian children has good local conditions for (1) sampling a variety of endurance sports and (2) meeting local coaches with a good understanding of the endurance training process.

5 Conclusions

The unique training session templates presented here are derived from world-leading coaches, whose athletes have won more than 350 medals in international championships. Overall, large variations in session loading factors were observed across sports, although the differences diminish with increasing intensity. AWD for LIT sessions ranges from approximately 30 min to 7 h, with differences being mainly explained by modality-specific constraints and accompanying consequences for load tolerance. For the same reasons, in addition to seasonal considerations, several sports perform large amounts of LIT using cross training. Intensive sessions (MIT and HIT) are considered paramount for performance progression by all coaches, and all sports perform considerably more MIT than HIT sessions across the annual cycle. Although most intensive sessions are conducted as intervals, competitions also account for a large proportion. Best practice interval sessions are characterized by a controlled, nonall-out approach, high AWD, and a slight progressive increase in intensity throughout. We also observed a trend towards lower work-to-rest ratio with increasing intensity.