Do all runners benefit from increasing mileage?

Do All Runners Benefit From Increasing Mileage?

Introduction

I hear these two arguments often – increasing weekly mileage improves performance and high weekly mileage is necessary for optimal performance.  As evidence supporting these beliefs proponents most often point to the training programs of elite long distance runners, who, as a group, all train with high weekly mileage.  The logic is that since these are the fastest runners on earth and they all run high mileage it proves that high mileage is optimal.  In effect the argument is that the high mileage is the reason these runners are elite.  Additionally, many runners report that when they have increased weekly mileage their performance has improved, providing additional support for these two beliefs.

However, there are several challenges to these arguments, chief amongst them being that people are individuals with individual genetic talents.  What works best for those with elite level genetic talents may not be best for those with average genetic talent.  Additionally, when scientists have researched the effect of weekly mileage on performance the results have not shown that weekly mileage exerts a strong influence on performance. 

These contradictory beliefs about the importance of increasing weekly mileage and the necessity of running high weekly mileage raise several questions.  Is weekly mileage as important as is promoted by many runners?  If so, can we determine if increasing mileage is equally beneficial for all runners?  It is possible that some runners benefit from increasing mileage while others do not.  It is also possible that different amounts of weekly mileage are optimal for different people, depending on things such as genetic talent or muscle fiber composition.  The purpose of this paper is to examine some of the research on weekly mileage to see if we can determine if weekly mileage is as important as conventional training wisdom suggests and if increasing weekly mileage is equally beneficial to all runners.

Research

A group of researchers in Switzerland wanted to determine if there was a relationship between training, life-style and running performance.  To investigate this question the researchers decided to collect training and life-style data from all the runners in a popular 10 mile race held in Berne, Switzerland (1).  They collected data on average weekly mileage for the previous year, average weekly mileage in the 3 months prior to the race, bodyweight, alcohol consumption, tobacco use, etc.  The study sample consisted of 4358 male runners over age 16, which was 66% of the total number of runners.  42% of these male runners had been running 6 years or longer and an additional 31% had been running for 3 – 5 years, so nearly ¾ of the competitors were not beginner runners.  As would be expected with any group this size there was a large range of average weekly mileage amongst all the runners, ranging from 0 – 100 km/week.

Performance in the race was electronically collected on all competitors by the race organizers and provided to the researchers.  The performance range for the 5703 male competitors over age 16 was 49:19 min:sec to 1:57:18 hr:min:sec, with an average of 1:13:20 hr:min:sec.  Based on delays due to the mass start and congestion at the finishing line the researchers estimated that the accuracy of the running times for the faster runners was below an error of 1% but may have risen to 2-4% for the slower runners (or about 2-4 minutes for the slower runners).

The researchers then used the survey data and correlated it with finishing time in the 10 mile race.  Combined, the six variables of weekly training distance (-0.46), age (0.37), body mass index (0.23), years of regular running (-0.19), weekly training frequency (-0.11), and cigarette smoking (0.10) predicted 47% of run performance. Note that the correlation between weekly mileage and race performance  was -0.46, which is not a strong predictor of performance – weekly training mileage predicted just 21% of race performance.  The correlation between weekly mileage and performance (age adjusted), is shown in figure 1.

Figure 1:  10 mile running time and weekly mileage (km/wk)

 

The researchers didn’t stop with just this analysis though.  They continued their investigation by dividing the runners into 3 groups based on finishing time, divided as follows:

runners with a finishing time of 55 min. – 1:06 hr:min
runners with a finishing time of 1:06 – 1:14
runners with a finishing time of 1:14 – 1:25

They then correlated finishing time with average weekly run mileage for each individual group.  Finally they compared differences in performance between these runners and the runners who ran just 0.5 – 10 km week and graphed the results.  The results are shown in figure 2.

Figure 2: Relationship between weekly run mileage and changes in 10 mile race performance separately plotted for 3 groups of runners. Percent running time reduction is in comparison to the running performance of those running just 0.5 – 10km per week.

Discussion

Fig. 1

The researchers noted some very interesting things about the relationship between weekly run mileage and race performance amongst the entire study population.

“First, the wide scattering of the 4000 observations makes it clear that weekly mileage is an insufficient predictor of individual race times.  Second, the regression line reflecting the mean relation between habitual mileage and running times is not linear but levels off in the range of 80 to 100 km/week.  Because of the narrow confidence limits of the regression curve (which are in turn due to the large sample size), this departure from linearity is highly significant.”

What they are saying in the first sentence is that weekly mileage alone is not a good predictor of performance.  As noted above, it predicted just 21% of race performance.  An examination of the provided data shows that lower mileage runners often outperformed higher mileage runners.  The data also shows that although multiple runners ran the same weekly mileage performance in the 10 mile race was not equal, with a wide variation in finishing time amongst those with the same weekly training mileage.  As an example the second fastest subject ran abut 23 km per week and finished in about 50 minutes while another runner also ran 23 km per week and finished at the back of the pack in about 1 hr and 30 minutes. 

The second thing the researchers are saying is that while race performance generally improved with increases in weekly mileage across the entire group (those who ran higher weekly mileage generally performed better than those who ran lower weekly mileage) that the race performance leveled off somewhere between 50 and 62 miles/week.  This means that further increases in weekly mileage beyond 50-62 miles/week were not associated with better race performance.  Those who ran 62 miles/wk performed the same as those who ran 50 miles/wk.  Contrast this finding with the oft stated belief that the high mileage run by elites is necessary for optimal performance.  The data presented in fig 1 suggests that, on average, mileages beyond about 50 mpw did not result in faster race times.

Based on the data found in figure 1, we reach the conclusion that weekly mileage does have an influence on performance in that increases in weekly mileage can result in improved performance.  However, the influence of weekly mileage is not strong, predicting just 21% of finishing time.  We can also conclude, based on the leveling of the relationship curve, that continually increasing weekly mileage does not cause continual improvements in race finishing time.  There appears to be an upper limit to improvements due to increases in weekly mileage.  Finally, this data shows that, on average, the relationship between increasing weekly mileage and race performance levels off at about 50 mpw.  This leveling occurs at a significantly lower weekly mileage than that recommended by those who use the training mileage of elite athletes as their guide.

Fig. 2

The data in figure 2 provides a very different perspective on the relationship between weekly mileage and race performance.  By segregating runners based on finishing time it provides a more detailed picture of the relationship between weekly mileage and performance for different groups of runners.  Recall that the % change in 16 km run time (the “y” axis in fig. 2) is derived by comparing performance to those runners who ran just 6.2 miles or less per week.

Fastest 1/3:  The data shows that increases in weekly mileage were linearly associated with better performance for only the fastest 1/3 of runners.  In this group those who ran higher weekly mileages outperformed those who ran lower weekly mileages (just as conventional training wisdom suggests).  On this topic the researchers noted, “The runners from the fastest third of the sample showed significantly greater differences in 16km times with higher levels of weekly mileage than runners from the medium and slowest third…”

Middle 1/3:  For the middle 1/3 of runners, race performance was better for those who ran higher weekly mileage up to a peak of about 40 km/wk.  Performance for those runners in this group who ran more than about 25 miles/wk was not better than those who ran 25 miles/wk.  Those who ran a weekly mileage more than 25 miles/wk did not run faster race times but they didn’t run slower race times than those who ran 25 miles/wk either. 

Slowest 1/3:  For the slowest 1/3 of runners, race performance was also better for those who ran higher weekly performance up to a peak of about 25 miles/wk.  However, race performance for those in this group who ran more than 25 miles/wk was worse than those who ran 25 miles/wk.  The relationship between weekly mileage and performance of those in this group stands in stark contrast to the relationship for those runners in the fasted 1/3 group.

There are several important conclusions we can draw from the data in fig. 2.  First, we can conclude that faster runners benefit from increasing weekly mileage, as the graph clearly shows a linear increase in performance with higher weekly mileages of the competitors.  Those who ran higher weekly mileage performed better than those who ran lower weekly mileage.  The implication of this linear relationship between weekly mileage and race performance is that increasing weekly run mileage results in improved performance for those in this group.  Many runners, as noted in the introduction, believe that increasing weekly mileage produces better performance.  This finding supports that belief.  Many runners also point to the high weekly mileage of elite runners as evidence of the effectiveness and necessity of running high weekly mileage.  Again, the graph of the relationship between weekly mileage and race performance for the fastest 1/3 of runners supports this belief.

Compare the graph of faster runner performance in fig. 2 to the mileage/performance graph in fig. 1.  No leveling in race performance was seen with increasing weekly mileage for the faster runners as shown in fig. 2.  A clear leveling of performance occurred at around 50 mpw for the entire subject population as seen in fig. 1.  This indicates that the leveling of race performance at 50mpw seen is fig. 1 is not accurate for the group of faster runners.

The second observation we make is that the graph of the relationship between weekly mileage and race performance for the middle 1/3 of runners is very different than that for the fastest 1/3 of runners.  The graph for the middle 1/3 of runners shows a clear leveling of performance at about 25 mpw.  Those runners in this group who ran more than about 25 mpw did not perform better than those who ran only 25 mpw.  The higher mileage runners didn’t run any faster but they didn’t run any slower either.  Performance for those running 25 – 62 miles/wk was equal.  Recall that in the fastest 1/3 of runners, those who ran 62 miles/wk outperformed those who ran 25 miles/wk.  The conclusion we can draw from this is that middle of the pack runners also benefit from increasing mileage.  However, the leveling of race performance of the middle third of runners indicates that beyond a certain weekly mileage these runners can expect race performance to level off and that additional increases in mileage will not result in faster race performance.  Note also that the leveling of race performance occurs at around 25 mpw, a significantly lower weekly mileage than is recommended by many training programs.

Our third observation is that the graph of the relationship between weekly mileage and race performance for the slowest 1/3 of runners is very different from both the fastest 1/3 of runners and the middle 1/3 of runners.  In this graph we see that race performance was highest in those who ran about 25 mpw.  Runners in this category who ran more or less than 25 mpw performed worse than those who ran about 25 mpw.  The greater the difference between a runners weekly mileage and 25 mpw the worse the performance in comparison to those runners who ran 40 km/wk.  The conclusion we can draw from this data is that slower runners, like middle of the pack runners, also improve race performance with increases in weekly mileage up to about 25 mpw.  However, at this point their performance diverges significantly from the other two groups in our study.  In the case of slower runners additional increases in weekly mileage beyond 25 mpw were correlated with a slower race performance.  Those running 12 mpw performed very similarly to those running 40 mpw, with both groups running slower than those who averaged about 25 mpw.  This indicates that, for this group of runners, increasing mileage beyond about 25 mpw not only didn’t produce improved race performance, but actually resulted in a slower race performance. 

This data quite compellingly shows that increasing mileage does not benefit all runners equally.  By correlating weekly mileage and race performance for 3 distinct groups of runners rather than simply taking the average of all 4000+ runners it becomes clear that the relationship between weekly mileage and performance is not the same for all runners.  The correlations lead us to conclude that some runners benefit more from increasing weekly mileage than do others.  We also see that the benefit of increasing mileage levels off much sooner for some runners than others and at a much lower weekly mileage than is suggested by many as the optimal weekly run mileage.  For some runners increasing mileage beyond a certain point not only doesn’t produce additional improvements but actually causes a decline in performance.

This is not the only study showing that increases in weekly mileage do not equally benefit all athletes.  A study of cross-country skiers found that 4 years of increasing mileage failed to cause performance improvements in half of the subjects (2).  In this study subjects were divided into two groups based on changes in performance from a full year of training.  One group, the higher performing group, responded to increases in training volume.  The other group, which was the lower performing group, did not respond to increases in training volume.  A review of the training data of both groups revealed that the lower performing group had not improved performance during the previous four years of training despite significant annual increases in training volume.  Though this study divided subjects into two instead of three groups, the results are consistent in showing that some athletes respond well to increases in weekly mileage while others respond to a point and then stop improving despite additional increases in weekly mileage.

Genetics

I speculate that those genetic factors that most influence race performance also play a strong role in ability to adapt and improve from increases in weekly mileage.  The genetic factors that make faster runners faster runners also allow them to improve with increases in weekly mileage, at least to some amount of weekly mileage beyond that measured in this study.  Those genetic traits of middle of the packers that result in this group not running particularly fast also prevent them from continually improving from increases in weekly mileage beyond a certain point.  Their genetics do seem to allow them to run additional weekly mileages without a decline in performance, at least up to some weekly mileage greater than that measured in this study.  And the genetic talents of the slowest runners unfortunately prevent them from either running fast or improving with increases in weekly mileage beyond very modest amounts. 

I further suggest that this explains the contradiction between the observation that elite runners all run high weekly mileages and research studies, such as this one, that show a modest correlation between weekly mileage and race performance.  The data from this study clearly supports the belief that elite runners do benefit from increasing weekly mileage to high levels, while others do not.  I suggest then the same genetic factors that allow elites to run so very fast also allow them to adapt and improve from very high weekly mileages.  Those with less than elite level speed are less likely to benefit from increasingly high weekly mileages.  I believe the correlation between race performance and weekly mileage for the 3 groups of runners in this study supports this genetic explanation.  There are surely exceptions, but I suggest this genetic explanation to generally be the case.

Summary

The argument is frequently advanced that increasing weekly mileage improves performance and that high weekly mileage is necessary for optimal performance.  The speed and training habits of elite runners are frequently used as evidence supporting these beliefs.  In contrast to these beliefs, research shows a modest correlation between weekly mileage and race performance.  The research reviewed here leads us to the conclusion that different athletes respond differently to increases in weekly mileage.  The data suggests that the performance of faster athletes improves with each increase in weekly mileage, up to some high level.  Performance for runners with average genetic talents levels off at a much lower weekly mileage with no further improvements despite additional increases in weekly mileage.  Finally, performance peaks at relatively low weekly mileages for those athletes with seemingly low genetic talents and then declines as weekly mileage continues to increase.  I suggest that the genetic factors that greatly influence an athlete’s natural, inborn speed also determine that athlete’s response to increasing weekly mileages.

References:

  1. Marti B, Abelin T, Minder C.  Relationship of training and life-style to 16-km running time of 4000 joggers – The ’84 Berne Grand-Prix Study  Int J Sports Med 1988, 9, 85-91
  2. Gaskill, S., Serfass, R., Bacharach, D., Kelly, J., Responses to training in cross-country skiers, Med Sci Sports Exerc, 1999, 31(8), 1211-1217

Comments

Do all runners benefit from increasing mileage? — 10 Comments

  1. Pingback: Do all runners benefit from increased mileage? | Training Science

  2. I don’t understand what the idea was behind this figure 2. Average times for runners in a range of finishing times, like 66 minutes to 74 minutes (middle) would just be a near normal distribution of times. Statistically it would be insignificant. It’s already reported that the accuracy for times falls off to up to 4% for slower runner, this would pretty well bury any possible signficance in that graph. Who created this graph?

  3. Richard, I’m just following up on my previous comment. Can you provide a link please or at least a reference to the study contain this graph in figure 2?

  4. This was something you created? In your blog above you’re telling people that this was performed by the same researchers that wrote the published Berne study:

    “The researchers didn’t stop with just this analysis though. They continued their investigation by dividing the runners into 3 groups based on finishing time, divided as follows:”

    That doesn’t seem very forthright to me.

    • Vinnie,

      Everything in the article, with the exception of the quotes, is something I created. Not just the graphs – the entire article is my own creation. I’m reporting the information and data found in the cited studies and I’m doing it in my own words. I don’t have publication rights to the cited research studies so I don’t include the original graphs, charts, ect from the studies in my article.

  5. I spent a little time looking at the cited Bern study to see if the graph was indeed there and why the researchers created it. I did have my doubts about it, given the way you were presenting it and your follow-up comments, but it is indeed there. My apologies for that. I would say that you’re reading far more into it than they intended. This is the item marked as ‘figure 4’ in the cited study, right?

    They didn’t look at the specific details of the three regression curves themselves and draw conclusions on them or even focus discussion on the subtle differences in them as you did. You quoted part of what they said, but not in it’s entirety or give the lead-in sentence. Here’s the whole quote:

    “Figure 4” (your figure 2) “also focuses on sample heterogeneity with regard to the association of weekly mileage with 16-km time. The runners from the fastest third of the sample showed significantly greater differences in 16-km times with higher levels of weekly mileage than runners from the medium and slowest third indicating that the regression curve of mileage and 16-km time would be steeper in the faster runners.”

    The last part indicates what they were looking for. It wasn’t what the optimal training mileage was shown to be for each group, as you have done. It appears that all they wanted to find out was whether the curve was sharply right-ascending, or sharply right-descending, for one or more of the subgroups to ascertain how the overall trend revealed in their mileage/time analysis (figure 1 in your blog) might be altered for different subgroups. All they said was that the response relationship between mileage and finish time for faster runners would be steeper than for slower runners. No specifics, just the general trend ‘steeper’, and they didn’t specifically compare the two slower groups in any way or look at any details of these curves.

    They spend more time looking at related factors for the faster group, like weight, smoking habits, years running, than they do on the table itself, so clearly they didn’t any weight at all into the slight variations in the two slower groups as you have done.

  6. It’s been a while since I last visited here. I thought I would offer some followup to clarify a few items from above.

    Most readers won’t have access to the original published studies, at least not without paying for the downloads. The Bern study authors reach very different conclusions than the blog author here and I’ll give them here for comparisons sake.

    For starters, the study authors do not downplay the role of training volume, they instead cite it as the single most significant factor that influences performance. Consider that some factors cannot be controlled at all, factors such as age or height, volume would then have an even greater relative level of influence than other controllable factors, such as lifestyle (think smoking, etc.), diet and weight.

    Second, they conclude that given the nature of the study, which concentrates on recreational runners rather than elites, optimal volume is in the range of 100-120KM (65-75 miles) per week, not the low levels (25 mpw?) the blog author here states. Part of this is likely due to confusion over the purpose of the graph shown here as Figure 2, which was neither intended nor designed to find optimal mileage for individual groups of runners on it’s own. I explained this part in my previous comment.

    Third, they note that the volume requirement to improve performance increases as endurance capacity increases. This makes sense since the body primarily responds to time and effort. As we improve and get faster, we cover more distance in given levels of time and effort and would translate to a corresponding increase in volume. Volumes around 65-75 mpw for a slower rec runner would be more like 80, 90mpw or more for more developed athletes. In relation to the blog topic, runners don’t benefit quite equally, but pretty close.

    Some other notes from the study authors’ conclusions:

    – The across-the-board difference in performance levels between low mileage levels (~10mpw) and their highest graphed levels (~60mpw) represent roughly 27% increase. They speculate that faster runners would likely experience performance increases higher than this.

    – If training volume is very low (10-15mpw), how often you run seems irrelevant. As volume increases, more frequent running is associated with better performance, although they do state that it’s hard to ascertain if the relationship is simply due to habit, convenience, preference, etc. It would seem to make sense that the more one runs, the more frequently one would likely run.

    Finally, this is quoted directly from the study to summarize their findings:

    “We conclude that this cross-sectional study of joggers who were mostly not highly trained confirmed that (in order of importance) weekly training distance, age, relative body weight, years of regular running, training frequency, and cigarette smoking were independently related to endurance capacity. Endurance capacity of joggers was, on average, much superior to that of the general population. A few training and life-style characteristics were able to explain the whole difference in mean endurance levels between joggers and the general population and half of the individual variance of endurance capacity among joggers. Thus, on average, the joggers investigated appeared not to be selected concerning biological predisposition and genetic endowment, but selected concerning behavior.”

  7. Hi, Build LSD to 30 minutes to start 6 weeks nice and slow with good form, rest 30 minutes, Now build to another 30 minutes same as we did before. Now you are able to jog with good form for 45 minutes daily. Why, because now you have a release of growth hormone at around 40 minutes.

    For me its all about the quality and quantity of the base lSD work. Thats when to double up slowly. Tripples are even possible, then go to 45 minutes.

    Jog aerobically 45 minutes 5 days, slow but with good form, in the afternoon every second day run for only 1 minute 30 seconds anaerobically at your goal pace or race pace 90 to 95% effort (every second day) for 1 min 30 seconds, after 3 to 4 weeks add, increase by 1 extra run after another 3 to 4 weeks increase another so that after 4 weeks you can run 4 intervals every second day for 1 minute 30 seconds 3 minutes rest between intervals. Run long runs once a week run 22 kilometers easy to start but build up speed of the weekly 22k slowly,you must be able to talk in sentences, not paragraphs. After 3 to 4 months trial over race distance at 90 to 95 percent effort.

    If you can run 4 to 6 intervals for 1 minute 30 at or slightly above YOUR race pace every second day your total anaerobic system just needs to be fine tuned to work with your aerobic system. This fine tuning is done in the long runs and you have been doing this along the way. The longer you train the easier it all becomes and the faster you become over the 22k. When you can run 22k at half pace easy, you are ready for a trial. The first trial is run to see how far you can run, so expect to have to pull out of the trial at some stage, may be not? My point is this, pulling out of the trial tells me how unfit you were at the begginning of the programme. By the way, if you cant run the 4 to 6 intervals, like you did last week, then you overtrained in the 20 to 25k long run on the weekend, simple. Easiest way to fix that is 3 days straight jogging again, pick up where you left off.

    Now its only a matter of time, just slightly increase the speed of the 22k every 3 to 4 weeks. every 4 weeks comes another trial.

    When you can run at 95% effort for the distance in a trial we will reduce the work load for a period of time so you are fresh for your new PR.

    My own Half marathon training. 1 hour 4 to 5 minutes and i am 57 years old.

Leave a Reply

Your email address will not be published. Required fields are marked *