In general, to achieve a target performance and to reduce the risk of detrimental effects of training, effective endurance running plans typically increase the frequency, duration, and intensity of training followed by a taper to maximize performance whilst reducing the possibility of adverse training effects [15]. There are many training plans recommendations that are used by probably millions of Marathon runners, but we know little about how a typical Marathon training plan recommendation for sub-elite athletes looks like and whether typical recommendations are consistent with current evidence from training intervention trials. The objective of this research was therefore to conduct a quantitative analysis of sub-elite marathon training plans to provide a comprehensive overview of current sub-elite Marathon training plan recommendations. While such an analysis has not been conducted before, other studies have sought to review the available literature for evidence-based research, study the training behaviour of recreational runners, or analyse elite training results-proven plans to make recommendations for marathon training [8, 9, 15, 23–27].
4.1 How Do the Recommended Recreational Training Plans Compare to Evidence-Based Research?
In 2007 Midgley et al. concluded that there was little direct scientific evidence to identify the most effective training methods for enhancing long-distance running performance, with even less evidence specifically for the marathon distance [8]. Since then, more work has been published to provide training guidelines for recreational runners and their coaches based on scientific evidence.
4.1.1 Running Training Methods
To improve performance in recreational runners, existing evidence recommends incorporating one to two high-intensity interval training sessions per week along with several sessions of moderate- and low-intensity continuous submaximal running into the training regimen [8, 23]. In the analysed plans, the high volume plans had an average of 8.3% of weekly volume in zone 4 and 5, while the middle volume plans had 5.3%, and the low volume plans had 4.7% at these intensities (Table 3). Despite the lack of clear understanding regarding the ideal volume and intensity of strength training for improving endurance running performance or preventing injury, it is advised to be included in a training regimen as well.
Another component of a training plan for which there is some empirical evidence is the taper before a race, or the intentional reduction in training volume before competition to improve running performance [28]. The varying tapering techniques used in research studies make it difficult to choose the best recommendation. According to a meta-analysis that investigated the impact of tapering on competitive athletes' performance, the most effective approach to maximize general performance gains is to implement a 2-week taper that involves an exponential reduction of training volume by 41–60%, without any changes to the intensity or frequency of training [29]. Intervention research focusing specifically on a 7-day taper found that the run taper group that reduced their training volume by 85% were 3% faster over a 5-km performance than the control group corresponding to an improved measured running economy [28]. With a focus specifically on the marathon distance, one study analysing the training activities of more than 158,000 recreational marathon runners determined that strict 3-week tapers are associated with better marathon performance compared to relaxed and shorter tapers [30]. In the analysed recreational training plans, peak week was found to be between 3 and 4 weeks out from race week, in line with a longer taper before the marathon race. Looking at the reduction in weekly volume following peak week until the marathon race, the tapers in the analysed plans are more gradual with a 22–31% weekly decrease. Focusing specifically on the last week before the race, the training volume decreases further by an average of 50% compared to the previous week in all plans. Among the three examined groups, the low volume training plans exhibit a shorter taper period, characterized by a peak week that occurs in closer proximity to the race week compared to the other groups.
4.1.2 Training Intensity Distribution
When designing a training plan, one crucial element is the distribution of training intensity across various intensity zones. Here a variety of different models are common including polarized, a pyramidal, and threshold models. Using a 3-zone intensity zone structure, a polarized training plan involves spending a significant percentage of time in zone 1 (75–80%) and in zone 3 (15–20%), with little or no time in zone 2, while a pyramid training plan has 70–80% of the volume in zone 1, with the remaining 20–30% in zone 2 and 3. Finally, when training follows the threshold model, the main focus, and therefore a higher proportion of overall volume, is on zone 2 training [23, 31]. Of these, polarized and pyramid training intensity distributions, that share a similar distribution of around 80% in low-intensity training but differ in how the remaining 20% is distributed, are the most recommended models. However, the evidence is inconclusive as to how best to optimize training [23, 31–33]. Based on these definitions and making it comparable, the analysed plans presented in Table 3 consist of a pyramid plan with high, middle, and low volume groups having 82-10-8%, 77-18-5%, and 78-17-5% in zone 1 and 2, zone 3, and zone 4 and 5, respectively. Previous intervention research has indicated that polarized training, with a distribution of 68-6-26% at low-lactate threshold-high intensity respectively, leads to the most significant improvements in various key endurance performance variables for well-trained endurance athletes compared to threshold, high intensity, or high volume training over a 9-week training program [33]. Conversely, a systematic review, which includes both intervention and observational studies, has found that highly trained distance runners tend to follow a pyramidal training intensity distribution approach, which is also related to high levels of performance and significant development of physiological determinants [31]. Another systematic review has analysed pyramidal training, polarized training, and threshold training and concluded that current evidence suggests pyramidal and polarized training to be more effective than threshold training, however among these no single optimal training intensity distribution has been established [32]. Although the inconclusive scientific evidence makes it challenging to recommend only one of these two models, recent research has explored the possibility of periodizing intensity distributions based on the stage of a runner's training cycle. For example, a 16-week pyramidal training plan followed by a 16-week polarized training plan results in the greatest improvement in performance, indicating that this could be a viable method to integrate differences in stimuli from both distributions [34].
4.2 How Does the Training Behaviour of Recreational Runners Differ from the Recommended Training Plans?
To compare how established training recommendations align with the actual training behaviour of marathon runners, additional studies that describe these behaviours were considered. Gordon et al. examined the training characteristics of 97 recreational marathon runners including both males and females sub-grouped by different finishing times (2.5-3 hr, 3-3.5 hr, 3.5-4 hr, 4-4.5 hr, and > 4.5 hr). This study found race speed for a marathon to be correlated with distance covered per training session, and weekly training distance [24]. Comparing these running behaviours, such as distance per week, distance per session, and the longest run of the plan, to the recommendations in the analysed plans, the training patterns of the 4-4.5 hr group (56.2 km/week) was similar to the middle volume (58.1 km/week), and the > 4.5 hr group (43.8 km/week) to the low volume plans (43.5 km/week). Only the weekly distance in the high volume plans of 104.9 km/week differed from the fastest finishing group of 2.5-3 hr, which on average ran 91.7 km/week. When training for a marathon, it appears the actual training behaviours of recreational runners correspond well with the recommended most popular training plans for marathon performance [24].
For further analysis, Doherty et al. performed a systematic review, meta-regression, and meta-analysis on 127 cohorts of runners to determine the relationship between training behaviours and marathon race performance [25]. This analysis examined the average weekly running distance, number of weekly runs, maximum weekly running distance, number of runs ≥ 32 km in the pre-marathon training block, average running pace in training, longest run completed, and hours of running per week and found that increases in any one of these training parameters coincided with significant faster marathon finish times [25]. Based on the formulas they created, the marathon finish time calculated from the training recommendations for high volume training plans is 3:04, followed by 3:36 for the middle volume, and 3:50 for the low volume group. These predicted finishing times are faster than those suggested with the plans themselves and those predicted based on training behaviour [25].
4.3 How do Training Plans for Recreational Runners and Elite Runners Differ?
To relate the examined recreational training plans analysed in this report to elite populations we compared our findings to the training habits of elite marathon runners. Billat and colleagues examined the training characteristics of top-class and high-level elite marathoners and while the absolute distances of these runners are very different from the plans investigated here, the average intensity distribution revealed 78% of the total weekly distance was run at velocities less than marathon pace, 5% at marathon pace, and 17% greater than their marathon pace, matching a typical polarized training model [26]. While the exact comparison cannot be made due to discrepancies between intensity distribution methods, the high volume group comes the closest to such a polarized model with an average of 83% of training at less than marathon pace (zone 1 and 2), and 8% greater than marathon pace (zone 4 and 5) while the middle and low volume groups follow a typical pyramid training model.
Additionally, giving further insights into the training behaviour of elite long-distance runners, Haugen and his colleagues published a review integrating scientific literature and results-proven practice to understand the training and development of elite long-distance runners [9]. For marathon runners, this review found the weekly running distance in the mid-preparation period to be between 160–220 km per week, again significantly higher than the examined training plans. The intensity distribution of this distance, in line with our examined plans, was made up of ≥ 80% of the total running volume being performed at low intensity (zone 1 and 2), 5–15% at middle intensity (zone 3), and 5–15% at high intensity (zone 4 and 5) inversely related to the middle intensity training [9]. The tapering for these athletes started 7–10 days out from the main competition, whereas for our analysed plans the peak week was around 4 weeks out from the competition, with an additional pronounced decline the last week before the race (Table 4 and Fig. 1) [9].
Finally, research from Karp found that among analysed qualifiers for the United States of America Olympic marathon trials, the large majority of the training was performed at low intensity, with men running 74.8% and women running 68.4% of their weekly distance, at a pace slower than marathon race pace [27]. In more detail, the distribution of training intensity for men and women was 75-10-10-5-3% and 68-13-12-7-5% for intensities below marathon race pace and at marathon race pace, lactate-threshold pace, ≥ 10k race pace, and ≥ 5k race pace, respectively [27]. In comparison, the distribution of the data presented here is skewed towards the lower intensities for all volume classifications with 82-10-5-3% for high, 77-18-2-3% for middle, and 78-17-3-2% for low for intensities of zone 1 and 2, zone 3, zone 4, and zone 5, respectively.
4.4 Limitations
Although our research has revealed new and potentially valuable insights that could assist coaches, athletes, and recreational runners in improving their training routines, there are several limitations to classifying the training plans in such a way that must be acknowledged. Firstly, the classification process involves subjective interpretation, as different training plans were written in various ways, making it necessary to analyse based on subjective decisions to ensure comparability. Moreover, the analyses here are limited to the last eight weeks before the race, as the research team believed this period to be the most critical for marathon training, which is another subjective decision. Additionally, both the subjective classification of the specific sessions into the five intensity zones and the classification of the training plan itself into low, middle, and high volume are subjective interpretations based on the range of training plans collected and the descriptions of the training sessions themselves.
Most training plans are not developed with a five-zone model in mind, and the intention of specific sessions may not always be apparent. Furthermore, we noticed discrepancies across the analysed training plans with different sources having varying definitions for commonly used phrases. We classified such sessions based on their descriptions in the plan rather than our understanding of the terms. For instance, several plans defined ‘steady’ runs differently, leading to varying categorizations. When steady was defined as a “purposeful pace … similar to marathon pace that helps to familiarize yourself to speeds you should set off on marathon day” [35], we classified this into zone 3, however for different plans steady runs were defined as the “runs to build the base for the rest of your training where conversations are still possible but only in shorter sentences” [36] or as a “continuous easy-medium pace” [37] which classified the sessions into zone 2. Some plans were also more detailed than others, and this may have affected the classification process. For example, one plan describes in detail a fartlek session starting with 20 minutes of easy running, then transitioning into 10 repetitions of 1 minute hard where “you should be running fast enough that you cannot sustain the pace for more than a few minutes”, followed by 1 minute at a very easy jog before completing the rest of the run at an easy running pace [38]; whereas another plan just includes 45 minutes of fartlek running with the explanation that “rather than running a set distance in a set time, you play with different running paces and distances until you feel you’ve completed the workout” [39]. Additionally, one plan might include 20 different types of sessions included in a plan, while another plan consisting entirely of easy and long runs [40].
Finally, as previously mentioned, another limitation results from converting time-based training sessions into distance-based measures, considering the variability of paces of runners that might intend to follow the plan which will in turn affect the distance covered in a given session. For example, as part of a tempo run, one source includes 30 minutes in zone 3 [41]. For an advanced goal marathon time of 3:00 hours, based on the included pace descriptions, this would mean running this session at a recommended pace of 6:12 minutes per mile and therefore covering around 4.8 miles. However, for the same exercise, if the goal time is around 4:00 hours, the pace for this tempo run would be around 8:10 minutes per mile meaning this session would cover 3.7 miles. While here for the analysis, we used the information available in the descriptions of the training plans to make the best calculation for how much distance would be covered in sessions written with only a time variable, there may still be considerable variability.
4.5 Lessons Learned and Recommendations for Future Training Plans
The limitations identified in this analysis have highlighted significant differences in how training plans are developed and presented for recreational runners, which could potentially cause problems for those attempting to follow such plans. This lack of standardization in training plans makes it difficult to compare different plans, which limits the overall evidence base in this field. It is recommended that future training plans should be developed using consistent language and descriptions to ensure clarity and ease of understanding for those following the plans. By standardizing the language used to describe training sessions, runners can better understand what is expected of them during each session, and researchers can more effectively compare the effectiveness of different training plans. A clear and comprehensive training plan may incorporate the following elements: setting a target marathon time as the desired goal, utilizing a standardized 5-zone model for intensity recommendations, specifying the intended volume for training sessions, indicating the running speed in minutes per kilometre as the intensity measure, and providing information on the training plan structure, whether it is polarized, pyramidal, or follows a different framework. On top of that, this analysis has revealed limits in the existing evidence regarding the best tapering techniques and the optimal training intensity distribution for marathon performance with current research being inconclusive. Additionally, future training recommendations should consider how to optimize marathon preparation for different genders and age groups as well.