Does increasing mileage make for a faster marathon?

Does increasing mileage make for a faster marathon?

Runners training for a marathon are typically advised to increase their weekly mileage in order to perform optimally.  For example, one popular book on competitive running recommends a range of mileage from 30-40 miles per week for a novice marathoner increasing to as many as 80 miles per week for a local champion (1).  The implication of this recommendation is that in order to improve performance, you should increase weekly mileage.  Another popular book on competitive marathoning states “…racing performance improves with increased mileage” and provides training programs starting with 55 miles per week and advances to more than 70 miles per week (2).  Additionally, there is a school of thought that basically states “you can’t really race a marathon until you are averaging at least 60 miles per week.”  The meaning being that you may be able to complete a marathon on less than 60 miles per week, but to perform optimally (i.e. “race it”) requires a minimum weekly mileage of 60 miles.  While there may be debate as to whether the weekly mileage required to “race” a marathon is more or less than 60 miles, the point of the statement is that weekly training volume is a primary influencer of competitive marathon performance.

Why do runners, and marathoners in particular, focus on mileage so intently?  There is a simple answer – performance.  There is a correlation between marathon finishing time and weekly training mileage.  As a general rule, the faster the runner, the more mileage they tend to run per week.  Numerous surveys of both non-elite and elite runners training habits support the observation that faster runners generally train with higher weekly mileage.  This is commonly interpreted to mean that the higher weekly training mileage is responsible for those faster performances, i.e. more mileage leads to better performance.

It is possible that conventional wisdom is correct and that mileage plays a primary role in performance.  There is no doubt that volume of training exerts an influence on performance.  But, it is also possible that the faster performances of the higher mileage runners are not primarily caused by the high mileage training.  Other factors could exert a stronger influence on performance than does mileage, even though high weekly mileage gets all the credit.

How could we confirm the level of influence mileage has on performance?  Some well designed research studies would do the trick.  Unfortunately, at this time there is insufficient research to definitively prove the relative importance of weekly mileage.  That doesn’t mean that there isn’t any research though.  There is quite a bit of research available, giving us some pretty strong clues.  One study in particular has received considerable interest since being published.

Marathon Training Research

Three researchers from the University of Northern Iowa were interested in the importance of weekly mileage in marathon performance (3).  They decided to study the effects of two marathon training programs that included running a competitive marathon.  They wanted to compare the effects of two training programs that had different training volumes and frequencies, but that had equivalent long runs (same time and intensity).  The researchers hypothesized that with all the subjects performing the same long run each week that, even with the differences in weekly volume and frequency, there would be no difference in actual marathon performance.

The researchers studied 51 healthy college students (18 males, 33 females) who had not previously run a marathon.  These subjects definitely fit the category of novice marathoners.  The subjects were matched for peak oxygen consumption and assigned to either a 6-days per week or a 4-days per week training program.  The study was a total of 18 weeks in length and culminated with the subjects running a marathon.

The 6-days per week group trained Monday – Saturday.  The 4-days per week group trained Tuesday, Thursday, and Saturday and one other weekday.  Sundays were non-training days for both groups.  The 4-days per week group averaged 20% less training volume (65 training hours vs. 83 training hours) and 30% less frequency than the 6-days per week group (69 training days vs. 99 training days).  From a weekly perspective, the 6-days per week ran an extra 60 – 90 minutes each week. 

Average training distance for the 6-days per week group increased from approximately 23 miles in week 4 to approximately 48 miles in week 16.  Average training distance for the 4-days per week group increased from approximately 18 miles in week 4 to approximately 39 miles in week 16.  Both groups reached peak weekly mileage two weeks prior to the marathon, and then decreased mileage as part of their taper.

Saturdays were dedicated to the long run.  Starting with a run of 55 minutes in week 4, the long run progressed to 150 minutes of running for weeks 14 – 16.

All runs were prescribed on a basis of time, not distance.  Instead of prescribing a set distance to be run, the subjects were prescribed a set amount of time to run.  This is not an unusual approach to training and, in fact, is an approach recommended by noted exercise physiologist Jack Daniels (4).  Running by time instead of distance allowed all subjects to start and finish together during group runs.  This did result in runs of different lengths for each runner, based on the pace selected by each individual runner for that particular run.  Therefore, while weekly training duration (time) was the same for all members of a group, weekly mileage within a group was not the same for all the members of the group.

Subjects were assigned an individual training heart rate (HR) range of 60% to 75% of heart rate reserve (HRR) – an easy to moderate level of intensity.  This intensity was used throughout weeks 4 – 18 of the study.  The subjects were taught how to check their pulse and were instructed to periodically monitor their HR during training runs so they could maintain the appropriate intensity.  As the HR decreased over time with training, the pace was increased to keep the HR in the assigned zone.

To sum up, let’s compare the two training approaches – specificity (long run) and intensity (60%-75%) were the same for both groups.  Frequency and weekly volume were 30% and 20% less respectively in the 4-days per week as compared to the 6-days per week training group.


So, how did it all turn out?  Was conventional wisdom supported, with the higher volume group outperforming the lower volume group?  The short answer is no.  At the end of the study there were no significant differences between the two groups.  VO2max, HR, body fat percentage, and lactate levels changed the same amount for each group.  More importantly, there were no differences in marathon finishing time between the two groups.  Average marathon finishing times for both groups were approximately 4:17 hr:min for the males and 4:51 hr:min for the females.  The range of finishing times for the males was 3:36 – 4:52 and for the females was 3:51 – 6:32.

Despite the increased frequency (30 additional training days) and volume (8 additional hours of training) of the 6-days per week group, there was no difference in marathon performance.  The key variables of frequency, volume, specificity, intensity, and performance are summed in Table 1.

Table 1: Summary of training data


Volume:  weekly mileage progression

Specificity:  long run progression

Training Intensity

Performance: Marathon Finishing Time

4 days per week

18 – 39 mpw

55 min. – 150 min.

60% – 75% HRR

4:18:42 males

4:50:54 females

6 days per week

23 – 48 mpw

55 min. – 150 min.

60% – 75% HRR

4:16:48 males

4:51:18 females



There are several very important things to be learned from this study.  First, recall that the subjects in this study, while healthy, were neither runners nor athletes, and none were following a structured exercise program prior to beginning this study.  They were beginner runners.  This is significant for two reasons. 

First, we know that the beginners experience the largest amount of improvement.  When beginners first take up a training program their rate of improvement is usually very rapid and dramatic.  After an initial spike in performance during the first several months of training, additional performance improvements occur at a reduced rate.  In this study there were significant improvements in the indices of fitness (decreased oxygen cost of running, decreased lactate levels, increase in VO2peak, etc.).  We would not expect a group of veteran runners following this same training program to have improved nearly as much as these beginner runners did.  

Second, it is generally promoted by conventional training wisdom that as weekly mileage increases, the rate of improvement decreases.  The gains experienced by increasing from 10 miles of running per week to 20 miles of running per week will be greater than the gains made from increasing from 50 miles of running per week to 60 miles of running per week.  Even though both would be an increase of 10 miles of running per week, increasing from 10 to 20 will result in greater improvements than increasing from 50 to 60.  Or said another way, at lower weekly mileages increases in volume have a greater effect.

Logically then, since this study involved both beginner runners and relatively low weekly mileages, based on conventional training wisdom we should have seen significant differences in marathon performance.  That was not the case though.  If this study had used veteran marathoners running higher mileage then we could have reasonably expected the results to be very similar.  However, this study used beginning runners training at what would generally be accepted as low to moderate mileage.  With the same results for both groups, the idea that volume of training plays a primary role in performance is called into question.

If volume of training exerts a primary influence on performance, then how do we account for the similar results of these two training groups?   On the other hand, if volume of training is not as strong of an influence as conventional wisdom suggests, then something else accounts for the duplicate results for both training groups.

This study suggests that frequency and volume are not as important as is generally believed.  If this is the case, then we must examine the other elements of training – namely intensity and specificity – to see if they can offer a reasonable explanation for the results of this study.

Intensity and specificity were the same for both training groups in this study.  Both groups conducted equivalent long runs (specificity) and trained at the same intensity levels (60% – 75% of HRR).  I suggest that since specificity and intensity were the same for both groups, and volume and frequency were not the same, that specificity and intensity are responsible for the duplicate results of these groups.  I further propose that this study reinforces the theory that intensity and specificity exert the greatest training effect on performance.

Other Possibilities

The evidence from this study provides compelling support for the belief that intensity and specificity are the most influential training elements.  However, despite the strength of this argument, two other possible explanations for the results of this study can be made and should be examined.  The first of these two possibilities is that the differences in frequency (33%) and volume (20%) were not large enough to cause a difference in performance or that the length of time (18 weeks) was insufficient for the higher mileage to exert a positive effect.  Perhaps if the difference in weekly training volume had been greater than 20% or if the training program had lasted longer than 18 weeks there would have been differences in performance.  This is a possibility, though a remote one because of the above mentioned facts about beginner runners and increases in relatively low weekly mileages. 

If it were true, it would mean that large increases in volume or long training periods of higher mileages are required in order to positively influence performance.  That fact alone would validate the idea that volume of training plays a less significant role in performance than conventional wisdom teaches.  If more than a 20% increase in volume is required in order to measure a difference in performance it would seem that training mileage is not as influential as other variables.  For example, relatively small increases in intensity have been shown to measurably improve performance.  Whether higher mileage would have caused a difference in performance or not, either way results in the the importance of volume of training on performance as being over-emphasized.

For supporters of conventional wisdom or those who wish to maintain the belief that higher mileage is the most influential variable for improving performance a second possibility exists to explain the results of this study.  This possibility is that higher volume training causes an unidentified, positive adaptation in the body such that, while it doesn’t make the runner faster, it does allow him to improve more if and when higher intensity work is added.  This idea would fall under the “base building” theory of training.  The idea is that high mileage training causes greater adaptations in the body (or “builds a bigger base”) than does lower mileage training, so that when faster paced training is added on top of the higher base a better performance results.  If this were true it would explain this study’s results.  In this particular study the higher mileage group would have built a larger base but failed to outperform their smaller base cousins because neither group conducted a period of higher intensity training. 

The main challenge to this possibility is that the currently accepted version of the “base building” theory preaches that an improved base will result in improvements in performance.  The theory holds that sub-maximal paces, or “aerobic” paces, will improve while the faster, “anaerobic”, paces will not improve.  Since the marathon is run at an “aerobic” pace, building a bigger base would seemingly result in at least some improvement for this distance.  Since this did not happen with this study, it does not lend support to the “base building” theory.  Additionally, a second challenge is that from a physiological perspective it isn’t logical that the body’s capacity would increase without some corresponding improvement in performance.  I know of one study that addresses the base building theory and will write a review of this study in the near future.


An analysis of this study offers some interesting ideas for consideration.  First, both training programs prepared the subjects equally well to run in the marathon, despite the significant differences in both training volume and frequency between the two groups.  This result calls into question the belief that volume plays the most significant role in performance.  In personal communication with Dr. Dolgener, one of the researchers on this study, he revealed that since this study was published they have trained 500+ novice marathoners and have permanently adopted the 4-day per week training plan for all.  He also stated that not a single person who completed the training program has failed to successfully complete the marathon. 

Since intensity and specificity were the same between the two groups, as was performance, this supports the belief that these two factors play a more significant role in performance than does volume of training.  The relative importance of intensity and specificity, i.e. is one more influential that the other, is not answered here.  While this study suggests that training volume is not the primary influencer of performance, it also does not give an indication of what level of influence, if any, volume may have on performance.  Other possible explanations for the results do exist, but are either unlikely or pure speculation at this time.

The one thing that is clear though is if you want to run a marathon high mileage training is not a necessity.


  1. Glover, B., Glover, S: The Competitive Runner’ Handbook:  pg 43: Penguin Books: 1999
  2. Pfitzinger, P., Douglas, S.: Advanced Marathoning: pg. 27, Ch 8-10: Human Kinetics: 2001
  3. Dolgener, F., Kolkhorst, F., Whitsett, D.: Long Slow Distance Training in Novice Marathoners: Res Q Exer Sci; 1994; 65(4); 339-346
  4. Daniels, J.: Daniels’ Running Formula:  pg. 88-89:  Human Kinetics:  1998


Does increasing mileage make for a faster marathon? — 12 Comments

  1. Pingback: Does increasing mileage make for a faster marathon? | Training Science

    • Marc,

      This particular study didn’t report on the rate of injuries. But studies that have examined injuries report that increasingly weekly mileage leads to a higher rate of injury.


  2. It seems that the exercise, like most sports” science”, does not actually tell us anything and the largest part of the article is spent in guess work at what caused the program to tell us nothing.

  3. The subject article, Long Slow Distance Training in Novice Marathoners, gets cited a lot as support for lower mileage vs higher, but this is off-the-mark from the goal of the study itself. The researchers were actually looking at the effect of the long run on marathon performance moreso than volume. This from the article body:

    ‘Specifically, the purpose of this study was to compare the effects of two marathon training programs that had unequal training volumes and frequencies —but used equivalent long runs—. It was hypothesized that as long as all subjects performed the same long run each week, the two training programs would result in equivalent benefit.”

    Peak performance online gives a better analysis of the study and more in line for what I consider.

    “The Northern lowa researchers tried to suggest that their study validated the importance of a weekly long run for marathon preparation, but it really didn’t do that. To test the true effects of the single long run, it would have been necessary to include a third group of subjects who ran 39 miles per week but without any extra-long (120 to 150-minute) efforts. All the lowa research is saying is that if you train at moderate intensities about four times per week (-)for your first marathon, doing 20-per cent more of this moderate work isn’t going to help you.”.

    Basically pponline points out, and validly so, that the researches blew it in their goal since all subjects had the same long runs, and therefore made it impossible to ascertain it’s effect compared to the two other factors (volume/frequency) involved. They may very well be right, especially with such modest training, but it can’t be known for sure due to the study parameters. Perhaps this is why the abstract reads differently than the body, maybe trying to salvage something from the work put into it.

    The study itself doesn’t give us any real hints on lower vs. higher mileage at all since it was bound to a pair of rather inadequate marathon training programs for beginners, made obvious by the rather pedestrian results of the subjects. All it does is tell us that if you are training for a marathon, you should choose an appropriate plan for your ability level at the start. It reveals nothing regarding how lower mileage vs higher mileage affects long term training.

    Think of it another way, if you are woefully undertrained for a marathon, a 20% difference in undertraining won’t mean anything.

  4. I look at this study as neither group ran low mileage or high mileage both groups ran middle of the road distance that is why no difference was seen.. low would be three days a week 20-30miles and high being 70 plus miles 6 days a week

    • Totally agree and can testify that higher miles = faster marathon times. the difference between 39 and 48 MPW is not worth mentioning. I ran my best at 60-100 mpw

  5. Getting faster must surely be a function of training intensity which is a combination of both distance and pace. Tanda (2011 & 2013) provides a good analysis of this. If a high mileage program is at too low an intensity no improvements would be expected – it will produce a similar set of adaptations to a lower mileage program at higher pace. Understanding the average relationship between the two is critical. If a runner can maintain the average weekly pace but covering a greater distance then performance improvements should be seen. I have tested the Tanda formula on a series of runners from VLM 2013 using data from Strava and its predictions match the performance profiles fairly well. The Tanda prediction is usually faster where runners have adopted a positive split strategy. Of course shoe weight, clothing, weather, hills etc all play a role. Someone who trains in heavier shoes wearing warmer clothes in warm weather and on hills will be subject to much higher stress than someone who does the same distance and pace in race wear on the flat. Predictors of stress can only predict with variable that can be quantified. Will more mileage help? In my opinion – yes – as long as it results in more training stress. It is certainly an easier and safer way of adding training stress than increasing peak pace during speed work.

  6. Regarding other possibilities with respect to the strength of the argument:

    In this section it was mentioned:” The first of these two possibilities is that the differences in frequency (33%) and volume (20%) not large enough to cause a difference in performance or that the length of time (18 weeks) was insufficient for the higher mileage to exert a positive effect”

    I would say that the ladder in terms of the length of time could very well have been insufficient for the higher mileage to exert a positive effect. I sense this based on two observations:
    1. My own personal experience where I significantly increased my volume for a significantly increased my mileage (around 42% over an additional 8 weeks roughly compared to the study) and found significant ANAEROBIC performance.
    2. It is also my understanding that the above result may be do to the fact that the additional time above the study may have allowed my body to create both more capillaries and mitochondria which directly exerted a positive performance effect.

    So, it seems like of there were a similar study where two groups were compared based on their capillary as well as mitochondrial development over a longer period of time, then we may get closer to isolating the most relevant causal factors.

Leave a Reply

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