I'm trying to improve my times in the half marathon and the marathon, so I've been reading a lot of training plans for races at those distances. One thing I've noticed is that most of these programs tell you to do your long training runs quite slowly, often recommending 60-90 seconds per mile slower than your target race pace. Why is that? That is, it seems like if I want to become a faster runner at a certain distance, the best way to train would be to run that distance, quickly. Yet most plans instruct you to do your long runs slowly, and to build speed in separate, shorter workouts. I can see the benefit of doing shorter workouts at a faster pace, but I don't see how doing my long runs at 60 seconds/mile slower than my goal pace is going to help me as much as doing them at as close to my goal pace as I can handle. Isn't that how most training works?

So why is that? Is it just to keep my legs fresh for other workouts? (If so, then couldn't I run faster if I know I'm going to have a day or two of rest following my long run due to my schedule?) And is there empirical evidence to support that these long, slow runs train athletes better than faster runs of the same distance? Or is it just unquestioned coaching orthodoxy?

Edit: There are some good answers below, but I haven't accepted any of them because thus far no one has been able to point to any scientific studies. They may not exist; they often don't for this type of thing (it's almost baffling how little hard, scientific evidence we have for so many of our most deeply held beliefs regarding fitness and exercise). That's what I'm hoping to find, at any rate.

  • Interesting question, I think I need to look some more into this.
    – Ivo Flipse
    Commented Jul 11, 2011 at 23:37
  • LSD...long steady distance - typical cardio training plans start with a solid foundation (training block 4-6 weeks)of LSD before moving to speed work - great question!
    – GuyZee
    Commented Jul 12, 2011 at 17:36
  • Great question -- I was about to ask the same one. I was looking at the SmartCoach tool on RunnersWorld.com for training for a half maratho. It recommends running long runs at around 6:15 for the week before, and then running the race at 5:13. I'm comfortable with the distance, but that feels like a big jump in effort for the race itself.
    – Rob Walker
    Commented Aug 17, 2011 at 14:58
  • In the absence of hard scientific evidence, people who get things done rely on A) people who have done the thing before: coaches, athletes, and B) their own reasoning powers. I think an absence of studies is perfectly normal. Commented Oct 25, 2011 at 16:24
  • @DaveLiepmann Yeah, that's obviously true. And I'm not abandoning my own training or faulting others for following the advice of coaches. The question was about whether there was any evidence for the conventional wisdom, though, because I'm skeptical and curious like that.
    – Jeff
    Commented Oct 26, 2011 at 16:57

10 Answers 10


The Furman Institute of Running and Scientific Training (FIRST) put out a good book called Run Less Run Faster. You can find a lot of information contained in the book at the original article in Runner's World.

Basically, the FIRST program outlines 3 types of training runs. The three FIRST workouts are a long run, a tempo run, and a speed workout that are designed to improve your endurance, lactate-threshold running pace, and leg speed.

What this points to is that long runs improve endurance, not speed. Upping the intensity would add extra stress for little gain. FIRST has done multiple studies, and has seen great results. Whether the studies are completely scientific, I cannot say.

  • 1
    The link to Runner's world article isn't working.
    – neves
    Commented Mar 16, 2016 at 21:59
  • None of the link are working...
    – Snostorp
    Commented Mar 28, 2022 at 11:06

I've been to the FIRST running camp (two years when I was in high school). They do have a lot of good information; however, I have anecdotal experience that will help you in particular with the half marathon or marathon.

I would avoid any sorts of interval or short speedwork stuff. Weldon Johnson (co-founder of Letsrun) dropped his 10,000m time to 28:06 off of slower, longer, and more frequent runs than what he had been previously doing.

Here is a great article by his coach John Kellogg: http://www.letsrun.com/jkspeaks4.php

I would say if you want to drastically improve your half-marathon time (as opposed to running for recreation) you will need to:

  1. Run twice a day. Historically, this has proven to be the optimal training frequency.
  2. Run as much mileage as you can handle. Build up slowly -- over a period of months and years.
  3. Add a long tempo run each week, or 2 of them if you can handle the stress. It should be about 8-9 miles and about 85% of the effort of your half-marathon race pace. To make it more interesting, you can start at a slower pace and work down to faster than your HM pace for the last few miles.
  4. Recovery. Increasing your training will inevitably make you feel very tired at first. Don't worry -- this is normal and you will get used to it. However, sleep is VERY important at this point because your body does not build fitness during exercise, but afterwards while you recover. Make sure you're taking in appropriate nutrients as well.

In my opinion, the "long run" each week is highly overrated. I could run a marathon or even a 50k faster than most of the population based on my college training for the 8k (one person on my team actually decided to run a well-known marathon just for the heck of it and won), which proves you don't need to throw in a single high mileage run to become fast.

One day with disproportionate mileage compared to the rest of your week is only going to get you injured or tire you out. Yes, there are physiological benefits that come with running longer, but you will reach this automatically as your day-to-day mileage steadily increases.

Also be careful with the long, slow distance (LSD) trap. Don't purposefully trudge through your runs at a pace that feels uncomfortably slow or unnatural. There's a famous quote -- I think by Sebastian Coe -- that "long, slow distance produces long, slow runners" (also quoted in this book). The point of the slow runs is mainly to recover from your workouts. Many people make the mistake of running their hard runs too easily and their easy runs too hard. As another example, a different person on my team would run 14 miles at 5:20 pace followed by two days at 8:30 to 9:00 minute pace. This pace is slower than most neighborhood joggers. He PR'd every single race that season and moved to the front of the team.

So I think that's about it. I would recommend finding someone to train with as well (and maybe you already do), because as you increase your mileage you will find it much more enjoyable with a training partner.

  • 2
    Twice a day seems excessive - do you have a link for this recommendation? Most runners I know either run daily or every other day.
    – Eyal
    Commented Oct 30, 2012 at 14:39
  • 2
    Well, it really depends on what your goals are. If you are training to run as fast as you can, you'll definitely want twice a day (almost all elite runners do this). If you're training to keep in shape, then daily or every other day is perfectly fine and probably advisable.
    – Nick
    Commented Oct 30, 2012 at 16:27
  • Great answer. I like the anecdotal evidence. Would like to hear even more examples from your experience or your that of your teammates if you've got them.
    – gary
    Commented Feb 7, 2016 at 16:07

The simple reason for a LSD is to get you comfortable with being on your feet for a long period of time. The reason for doing LSD runs at a much slower pace is because running at pace is very demanding on your body, especially over several hours. It's all about preventing injury.


I'm going to start with addressing your edit first re: some scientific studies. You're right research on this type of stuff often leaves something to be desired. However, long distance running is actually one sport we have a lot more research than most.

Notice the theme of these studies, bolding mine:

"Like Slovic, other investigators have demonstrated a relationship between quantity of training mileage and final time."


"The better runners were seen, on the whole, to have been running longer, and to have more strenuous regimes, both in terms of intensity of training and distance run per week."

  • Anthropometric and training characteristics of female marathon runners as determinants of distance running performance

"Marathon Performance Time was also negatively related to previous marathons completed (r= -0.47), workout days (r= -0.47), two a day workouts (r= -0.52), total workouts (r= -0.56), km per workout (r= -0.58), total training min (r= -0.56), training pace (r= -0.66), max km per week (r= -0.70), km per week (r= -0.74), km previous 12 weeks (r=-0.74), and km per day (r=-0.77).

Marathon Performance Time for our population of runners may be predicted (r= 0.82, R2= 0.68) by the following equation:

MPT, (min) = 449.88 - 7.61 (km per day) - 0.63 (training pace)"


"Marathon performance correlated to 5-km, 10-km, and half-marathon performance and to years training, average and peak weekly distance, number of weekly runs, and number of runs ≥32 km for women."

  • Training Characteristics of Qualifiers for the U.S. Olympic Marathon Trials

"the average weekly training volume in kilometers (r2 = 0.224, p < 0.01) and the personal best time in a marathon (r2 = 0.334, p < 0.01) were significantly associated with total race time, whereas no anthropometric variable was related to race performance (p > 0.05)"

  • Training Volume and Personal Best Time in Marathon, Not Anthropometric Parameters, are Associated with Performance in Male 100-KM Ultrarunners

"body mass index (r = 0.24), the sum of eight skinfolds (r = 0.55), percent body fat (r = 0.57), weekly running hours (r = −0.29), weekly running kilometres (r = −0.49), running speed during training (r = −0.50), and personal best time in a marathon (r = 0.72) were associated with race time. Results of the multiple regression analysis revealed an independent and negative association of weekly running kilometres and average speed in training with race time."

  • What is associated with race performance in male 100-km ultra-marathoners – anthropometry, training or marathon best time?

"In men, the mean weekly running distance, minimum distance run per week, maximum distance run per week, mean weekly hours of running, number of running training sessions per week, and mean speed of the training sessions were significantly and negatively related to total race time"

  • Sex Differences in Association of Race Performance, Skin-Fold Thicknesses, and Training Variables for Recreational Half-Marathon Runners

So one reason to engage in long, slow, distance training, is because it helps you run more miles, and the more miles people run, the faster their times tend to be. However, that doesn't mean it's optimal or the cause of faster times.

One reason to not immediately buy into this philosophy is it's really quite different than other human performance events. There are far and few events where merely doing the event more and more will keep increasing performance.

I'll give two areas for food for thought.

  1. For non-professional athletes, time is a limiting factor.

It's not easy fitting e.g. four+ hour workouts into a week.

You almost always find everyday people basically do the best they can during the work week, then on the weekend have one long workout, trying to simulate the event the best they can.

This is why first timers to these types of activities often end up destroyed after the event. By far the biggest mistake I see first timers do is,

"Eh, I mean, I've done a 10 mile run in training. I think I'm ready for a marathon."

This is no different than a lifter going,

"Eh, I mean, I've squatted 200 pounds in training. I'm sure I'm ready for 400 pounds. Or more."

You're talking more than a 100% increase in volume, overnight. Cue the stress fractures.

In other words, the average rec person has a hard time getting the required volume in, then ends up delusional thinking what they've been doing is enough. Until they get hit by a wrecking ball on event day. The more running they can do before the event, the less likely this happens, the better they do.

  1. Body weight

Again, more a concern for rec people. The more they run, the more likely they are to weigh less, the better performance in longer distance events tends to be.

That is, a person could likely do just as well running less with more attention to their eating.

In fact, there is really only so many calories a person can burn from exercise. (See: New insights as to why you can't avoid exercising.) The idea someone burns e.g. 10,000 calories is false. Where I wouldn't be surprised if those engaging in more training simply do two things 1) blunt their appetite more 2) have less time to eat.

One way to back this up is in many of the above studies you can see anthropometry is not correlated with performance. Well, at the elite level, we know that's not true. Kenyans dominate, and they have a very specific body type.

For the majority of people engaging in long distance competitions, the ability to merely do the event is an enormous hurdle. The ability to then run the whole thing is another hurdle. The ability to do that with any level of regularity and not got hurt yet another. Finally, we have speed.

You will see average running pace is also correlated with performance in the studies above. Speed of course matters. But for the average person looking to run these distances, it's not a priority. Finishing without having to limp for a month afterwards is.

Say you're an everyday person following a common marathon training plan, where there's one long run on the weekend. Your options are:

  1. Run 10+ miles at a relaxed pace
  2. Run 10+ miles at race pace or faster

The second involves more injury risk, and a greater chance you won't even hit the distance. Marathon training has an absurd injury rate. You have to get people to the distance before you worry about speed.

But once someone's body can regularly handle the distance, then you're right. Speed needs to be addressed, and not just in shorter workouts. (That said, even at the elite level, they will often run longer distances slowly, just to maintain the ability of their body to run that far, while lessening the stress of how intensely they run it. You can't run hard every workout.)

  • One item that I have observed anectdotally and in myself is that you gain 3-5 seconds in pace per mile for the same effort with each pound less that you weigh. So if you weigh 180 and run at a 6:30 pace, and lose 10 lbs you generally will run 30ish seconds faster per mile for the same effort. Of course you also have the correlation that in order to lose the weight you are probably training more, so it's not just a "lose weight run faster" kind of equation. Otherwise great answer with some solid research to back it up.
    – JohnP
    Commented Sep 6, 2017 at 15:49
  • Interesting observation! The cycling world is notorious for hyper-focusing on their power to weight ratio. An interesting longterm study would be to follow endurance athletes and see at what weight is this ratio optimized. At some point, there is going to be diminishing returns. After all, part of generating power is having some muscle.
    – b-reddy
    Commented Sep 13, 2017 at 12:36

Continuing on from the answer from Nick, I think that it is a strong case for learning the principals of training.

I'll expand; If you train at 100% all of the time you will burn out within 6 months, this is because you are pushing your body to the limit all of the time. That is not good for you as our bodies were not designed to do this.

I think you should research periodisation as well. At certain times of the year, you want to focus more on different things. Eg, winter is for putting in the miles, summer is sharpening up for the races etc. This will vary from sport to sport and disciplines within sport.


I am a 100/200m sprinter and we would following something similar to this and bare in mind that it is specific to sprinting:

Sept -> End/Jan

Putting in the miles, so lots of over distance training like 3/400's (we need the base endurance to increase the speed later on). This is typically about the 80% effort mark.

Feb -> End/April

In these few months, we would up the pace to 90% for about half of our sessions keeping the rest at 80% effort. One key term we use here is lactate stacking, so short recoveries in our interval training.


2 (sets) x 6 (reps per set) x 150m (distance) with each run in 20/21 seconds off of 45 second recoveries and 5-8 min set recovery. We would then do something like 10 x 40m at near maximal effort to push our bodies even harder.

Mar -> July

This is usually race season, so the intensity gets turned up to 90% effort + every session. Loads of 100% speed work.

Remember it's all specific to your distance. Research periodisation and different training types (interval,fartlek etc).

  • +1 good one about year-planning, short-term planning is very short-term thing :)
    – user2598
    Commented Dec 28, 2011 at 5:03

Short answer: Safety and culture. There are no studies proving it.


Most people using a training plan are amateur runners, and many are beginners. They are self-coached. Elites have coaches.

As others have said, for beginner runners, LSD is essential for building endurance, since it is either not possible or not safe for them to run long and fast.

For trained runners and particularly elites, long and slow runs have not been demonstrated in controlled studies to be better than specific training for marathon performance.


People do what is familiar, but not all reputable coaches are staying within the mold. A vocal example is Italian coach Renato Canova. From his 2017 Valencia talk:

Specific quality is the key for running at the best possible level. Running long and slow, combined with short and fast, doesn’t produce the possibility to run long and fast

Source PDF

Video of the talk

Slow running is still essential. Canova does prescribe long slow runs. His athletes do them earlier in the season, before the "Specific Period". Canova also prescribes many short slow "regenerative" runs. These easy runs enable the tougher quality workouts:

In order to produce “Quality”, we need to use big modulation between High Intensity and Recovery

Never athletes can run very fast, if are not able to run, sometimes, very slow and easy


I've always found that if I run faster, I can't run as far. It's as simple as that for me. If I want to run a long distance, I can't run as fast. Eventually, as I either run that distance more, or increase my distance even more past that, I can run it faster.

Also, I'm always surprised by my times on race day. There's a lot more motivation to keep you going a little harder then.

If you find that your long runs are too easy and you think you should be pushing harder, you may consider upping your target time. Maybe you underestimated your ability!


I've just reviewed Tim Noakes, Lore of Running, 4th ed (2001) - specifically the section on Base Training (pp303-317). This book is very well regarded as the authority on the science of running. I can find nothing in that section which refers to a peer-reviewed study. All the advice is the collated advice from some very well regarded coaches - e.g. Arthur Lydiard.

I will note that Lydiard had his athletes running long but fairly fast and on a hilly course.

If you really want Dr Noakes does respond to email.


Building upon what others have said, the major part of the long run is to improve your endurance, and not your speed. Yes, you can argue that running your long run at closer to target pace will help to bring progress, however, you are not building your endurance.

I point you in the direction of Hadd's Approach to Distance Running - it may not be as strictly scientific as you are after, but I do find it a very interesting read (with some references to lactate studies).


The part that stands out most to me, is when the author relates training to a tube of toothpaste - if you consistently train the upper thresholds (pace/heart rate), then you are only squeezing toothpaste out of the top of the tube. If you want to get all of the toothpaste out, you must squeeze all of the tube.

I certainly do recommend reading the document - I hope it answers a portion of your question!


Based on my readings & knowledge, running slow (relative to your performance) triggers adaptations that can't be achieved any other way namely eccentric heart development, mitochondrial development for processing oxygen. Also, it regulates your VLa max i.e. how fast does the glycolytic system works. LSD runs tend to temper down the glycolytic system. This can be desirable or not (a 800m runner does not have the same requirements as a marathoner)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.