Meb Keflezighi

In a recent post of mine—one that, like a number of my recent posts to this blog, dealt with the subject of pacing in running—I concluded with the following observation: “A masterful pacing performance like Scott Fauble’s 2:08:52 finish at this year’s Boston Marathon, which he achieved with dead-even 1:04:26 first- and second-half splits, are as marvelous to behold as a perfect golf shot, and the science behind such feats is truly mind-blowing.”

As a writer, I am endlessly surprised by the things certain readers get hung up on, and I was more than a little surprised that a few readers got really hung up on the above-quoted sentence. One commenter labelled the statement “controversial,” adding, “No way even split in Boston is optimal.” Another asserted that “even splits may very well be the scientifically ideal way to run a race but it’s just common sense that that isn’t the case with a hilly course.”

I’d like to take this opportunity to address these criticisms, not for the sake of winning an argument but to help runners like you better understand the important skill of pacing. Now, I will concede that I probably should have given an example of masterful pacing that was less vulnerable to being challenged than Scott Fauble’s even-split Boston Marathon, especially given that my purpose in adducing this example was not to provide empirical support for my main argument but simply to convey my appreciation for the beauty of expert pacing. But I chose Scott for a reason, and I stand by my claim that his pacing performance at this year’s Boston Marathon was masterful.

What made it masterful? First, as his 5K split times throughout the race attest, Scott avoided “bad” miles—those throwaway slow miles that have an outsized negative impact on finishing times. On all marathon courses, both flat and hilly, the runners who run the fewest miles at a pace that is slower than their average pace for the race as a whole come closest to winning at the end. While it is not optimal to be rigidly consistent with pace in a topographically interesting marathon like Boston, statistical analyses of pacing data from the Boston Marathon have shown that the least consistent pacers fare the worst, not just generally but also with respect to their own historical standards, and the most consistent pacers fare the best. Scott is now the fifth-fastest American finisher of the Boston Marathon in history, and he owes it partly to his choice and ability to pace the course with remarkable consistency.

The second reason I consider Scott’s pacing execution masterful is that he moved up from 22nd place at the halfway point of the race to 7th place at the finish. All of the fifteen athletes Scott passed in the second half ran positive splits (meaning their second half was slower than their first), in contrast to Scott’s even splits. I’ll delve deeper into the specifics of the Boston Marathon course profile in a minute, but the point to be made here is that, when my Facebook friends say that even splits in Boston are not optimal, what they are inferring is that a positive-split pattern is preferable in this particular event because the second half of the course features more climbing and less descending than the first. But if we accept this definition of “optimal,” then the fifteen runners Scott passed in the second half of the race did it “right” and Scott did it “wrong.” Which is absurd!

The third reason I consider Scott’s pacing execution masterful is the way he finished. I know Scott well, having trained with his team for thirteen weeks in 2017. He’s as tough as they come, and can dig deeper than just about anyone else in any race. Watching his beautifully ugly stretch run down Boylston Street put a lump in my throat, for it was clear he was digging as deep as he ever had, carving himself hollow in the hope of catching one more runner (the fading Albert Korir of Kenya—another positive splitter—who finished just two seconds in front of Scott) before it was too late. Try as he might, though, Scott wasn’t really able to kick per se, or lift his pace much at all, as far as I could discern. But he didn’t lose steam either, as Korir and the fifteen runners he’d overtaken had done.

This is the part of pacing skill that can’t be measured. If thirty-plus years as a runner and twenty-plus years as a coach and student of the sport have taught me anything, it’s that, when a runner is giving absolutely everything he has to give in the last part of a marathon and he neither speeds up (much) nor slows down in relation to the entire rest of the race, that runner his paced himself masterfully.

Grab some popcorn, I’m just getting warmed up.

Now, to the course. Much is made of how much tougher the second half of the Boston Marathon course is than the first half. Too much. Both halves are net downhill, with the first 22 kilometers dropping 72 meters and the last 20.2 kilometers dropping an additional 61 meters. Both halves feature uphill portions that go against the overall altimetric trend, although the lion’s share of the elevation gain does fall in the second half. So, while the second half is indeed tougher than the first, it’s not drastically so. When I emailed Scott Fauble to request his take on Boston, he said that, for an elite male racer like him, an even-split race like the one he ran equates to a 20- to 30-second negative split on a flat course like Berlin’s or Chicago’s. In other words, for a sub-2:10 guy, the second half of Boston’s course is 20 to 30 seconds slower than the first. That’s it.

I think most of Scott’s fellow elite Boston veterans would agree with this assessment, and if they’re right, then given what we know about optimal pacing in flat marathons, running even splits in Boston is in fact optimal. History shows us that, on a flat marathon course, a 20- to 30-second negative splint tends to yield the best final result among elite racers. The current men’s world record of 2:01:39 was set by Kenya’s Eliud Kipchoge in Berlin in 2018. He covered the first half of that race in 1:01:06 and the second half in 1:00:03, a 33-second negative split. Logic tells us that if a 20- to 30-second negative split is optimal in flat marathons for elite racers, and if a 20- to 30-second negative split in a flat marathon is equivalent to even splits at Boston, then even splits must be close to optimal in Boston, at least for top finishers.

By now my Facebook friends are beginning to sweat a little. “One example doesn’t make a pattern!” Fair point. Even most elites run positive splits in Boston. But most elites run positive splits in every major marathon, including flat ones. That’s because most elites race for position, not for time. They stay with the lead pack as long as they can and then they blow up. Or not. The most successful elite performances in Boston, as in every other major marathon, follow an even-split or a slight negative-split pacing pattern. Forget Scott Fauble. Geoffrey Mutai’s Boston Marathon course record of 2:03:02, which I witnessed from the media center back in 2011, resulted from splits of 1:01:57 and 1:01:05. The women’s winner that year, Caroline Kilel, split 1:11:30 and 1:11:06. Even in the rare year when a runner solos to victory off the front, as Meb Keflezighi did in 2014, we see the same pattern. Meb’s splits that year were 1:04:26 and 1:04:21.

Growing desperate, my Facebook friends move the goalposts, pointing out that what’s true for elite runners isn’t necessarily true for other runners. Scott himself expressed a similar caution in our email exchange, writing, “I don’t think you should take any lessons from a pro race on pacing.” But whereas Scott was talking about the difference between racing for position and racing for time, my Facebook friends, in their desperation, are suggesting that even splitting is unrealistic in Boston for nonelite runners.

I know from personal experience that this suggestion is baloney. I ran my first Boston Marathon in 2009, covering the first half in 1:19:45 and the second half in 1:59:25, finishing in 3:18:10. Oops. When I ran my fourth and final Boston Marathon ten years later, I’d figured out the race and gotten a lot better at pacing marathons generally. My splits on that occasion were 1:27:25 and 1:26:43, and I finished in 2:54:08. Hooray for me. But that’s not my point. My point is that I’m not Caroline Kilel or Meb Keflezighi—there’s nothing special about me. Any runner who is well prepared and who understands the course can pace Boston similarly and reach the finish line quicker for it.

Credit: Kevin Morris Photographer

When I asked Scott Fauble if he could have done anything differently pacing-wise in the 2022 Boston Marathon that might have gotten him to the finish line quicker, his short answer was “probably not.” His longer answer was this: “I could have gone out faster. But the danger of that in Boston is because the second half is harder you’re risking losing a disproportionate amount of time on the hills if you go over the line. And more specifically, at Boston the last 5 miles are some that you absolutely have to take advantage of so I think it’s better to hold back and make sure you still have your legs coming home.”

This, folks, is what it means to understand Boston and how to run it! My Facebook friends’ take on the race is far less nuanced. It’s basically, “First half easy. Second half hard. Even split impossible.” But where these guys see only black and white, Scott and I see shades of gray. Take Heartbreak Hill. Despite the scary name, that thing is little more than a glorified speed bump, rising 91 feet over half a mile. If not for its placement in mile 21, it wouldn’t scare anyone. Sure, it doesn’t feel like a speed bump on tired legs, but only runners who don’t come prepared or who don’t heed Scott Fauble’s advice to “hold back” are slowed much by it. In 2011, Geoffrey Mutai’s fastest 5K split of the entire race (14:12) came between 30 and 35 km, a segment of the course that covers three of the four biggest climbs, including Heartbreak. Not only was Mutai not slowed down by those ascents, he sped up on them! And again, you don’t have to be Geoffrey Mutai to glide over these hummocks. In 2019, I covered mile 21 in 6:42, just 4 seconds slower than my pace for the race as a whole.

It’s beyond the scope of a single blog post to say everything there is to say about pacing the Boston Marathon, or pacing marathons more broadly, or pacing in general. My modest goal here is to impress upon you how much more subtle and nuanced is the art of pacing than folks recognize. Squinty eyes see only dyads: fast/slow, hard/easy, up/down, first part/last part. The reality of pacing is far more textured, in multiple dimensions—kinesthetic, perceptual, affective, cognitive. My message to you is this: If you want to find yourself walking up Heartbreak Hill, listen to my Facebook friends. But if you want to finish your next race, and the one after that, and the one after knowing you couldn’t possibly have done any better, as Scott Fauble did at the 2022 Boston Marathon, then do yourself a favor and read my full take on this important topic, ON PACE: Discover How to Run Every Race At Your Real Limit. Or read this free sample chapter and then decide if I know what I’m talking about.

Recently I had the opportunity to read a prepublication copy of a book by Bryan Green, who ran with Meb Keflezighi at UCLA and now cohosts (along with retired elite middle-distance runner Jon Rankin) the Go Be More podcast. Titled Make the Leap, the book is based on the premise that, as Green puts it, “the better we think about our training, the better we will train.” He writes, “The workouts are not the problem. Having a better mental framework to understand training is what’s missing. It doesn’t matter how good the training plan is if you’re holding yourself back mentally.”

I think he’s right. And I also think the same truth holds for diet. In my experience, what distinguishes athletes who are happy with the results they’re getting from their diet from those who aren’t is not so much what they eat as how they think about food and eating. Science backs up this notion. Consider the following studies:

Canadian researchers reported in the journal Personality and Individual Differences in 2007 that, within a population of 364 college students, emotional intelligence was a strong positive predictor of healthy diet strategy.

Seven years later, in the same journal, scientists at Laurier University and Gettysburg College reported that trait-level mindfulness was strongly associated with healthier eating.

In a 2015 study published in Psychology & Health, Researchers in Florida reported that, within a population of 5,150 adult subjects, those who scored high for neuroticism in a standard test of personality traits had a significantly higher body mass index.

Another 2015 study, this one done by an international research team and published on PLoS One, reported that the well-established relationship between eating consistency and successful weight management was mediated by the personality trait of self-control in a group of 164 women.

Finally, New Zealand researchers reported in a study published in the journal Appetite in 2014 that “[p]articipants with a weight-loss goal who associated chocolate cake with guilt were less successful at losing weight over a 3-month period compared to those associating chocolate cake with celebration.” This effect was mediated by lower levels of perceived behavioral control over eating.

I’m not suggesting that the food a person eats doesn’t matter. As you well know, there’s tons of research out there showing that the healthiest people tend to eat lots of certain kinds of foods and small amounts of certain other kinds of foods. But who is it that actually consistently eats a lot of healthy foods and not a lot of unhealthy foods? People who think a certain way about food, that’s who!

I’m not an expert in food psychology, so I probably won’t get this right, but in my experience as an endurance coach and nutritionist (which is extensive), I see two types of individuals as far as thinking about food is concerned: those who problematize eating and those who don’t. To problematize a thing is to make a problem of it even though it doesn’t need to be. Healthy eating comes easily for healthy eaters. That’s because, on a purely practical level, healthy eating really is easy.

This doesn’t mean it’s your fault if you’ve struggled to find and stick with eating habits you’re fully satisfied with. Things like neuroticism and non-mindfulness that make healthy eating difficult for many people aren’t flaws to be ashamed of but limitations to work on. The message of this post is that if you really want to reach a point where healthy eating is easy, stop searching for the right diet (which has been right in front of you for your entire life) and take steps to improve how you think about food.

One way to do this is by practicing mindful eating. There are several smart phone apps, including the popular Am I Hungry? app, that can help you in this process, as well as books like Savor: Mindful Eating, Mindful Life by Thich Nhat Hanh and Lilian Cheung.

As for improving how you think about training—I’ll revisit this topic when Bryan Green’s book is released. Promise!

If you ask the average running coach what good running form is, he or she will probably answer with phrases like “midfoot strike,” “high stride rate,” and “low vertical oscillation.” It is true that these and other form characteristics are common in top runners and less common in slower runners. But there are exceptions. Meb Keflezighi won the 2009 New York City Marathon as a heel striker, Ryan Hall set an American record of 59:43 in the half marathon with an unusually bouncy stride, and Mo Farah won 10 Olympic and World Championships gold medals with an exceptionally low stride rate.

The Basis of a Good Running Form

On the basis of the fact that no two champion runners run the same way, I would argue that no specific movement pattern defines good running form. What, then, is the true essence of good running form? One possibility is efficiency. This idea has a lot going for it. After all, what is it that you really want your running form to do as distance runner? Not maximize power–that’s for sprinters. Rather, you want your running form to minimize the energy cost of running so you can go faster and farther before you become exhausted. It doesn’t matter how you minimize the energy cost of running, biomechanically, but only that you minimize the energy cost of running. Obviously, some movement patterns are generally more efficient than others, which is why most of the best runners are midfoot strikers. Nevertheless, each body is unique, so we shouldn’t expect every runner to minimize energy cost in the same way.

As sensible as it seems to define good running form as that which maximizes movement economy in an individual runner, however, I think this definition misses the mark as well. The true definition of good running form, I believe, is the particular way of running that feels easiest to an individual runner. I can’t claim to have come up with this theory. I got it from Samuele Marcora, father of the psychobiological model of endurance performance. Here’s an excerpt from an email message Marcora wrote to me about his perception-based theory of biomechanics:

“Many specialists in this area. . . still think that the [central nervous system] controls locomotion to optimise energy. They envisage an optimal solution in which the CNS senses energy expenditure and adjusts locomotion parameters (e.g., step frequency) to minimise energy expenditure. There are some basic neurophysiological problems with this proposal. However, the biggest problem is that there are several instances in which the CNS chooses locomotion parameters that optimise perception of effort at the expenses of energy expenditure (e.g., walking to running transition). When energy seems ‘optimised’ this coincides with the optimal perception of effort as well. So perception is always optimised but energy expenditure is often not. The. . . conclusion is that the CNS optimises perception of effort rather than energy expenditure.”

Energy expenditure and perceived effort are closely linked. For the most part, movement patterns that reduce energy expenditure also make running feel easier. It is very likely that the evolutionary reason individual runners adopt the particular way of running they do is to minimize energy expenditure. But it is perceived effort that actually controls this process. In other words, running in the way that minimizes perceived effort is the means by which runners achieve the end of minimizing energy expenditure.

Consider what happens when you’re running along and you turn into a strong headwind. You hunch forward, don’t you? This natural adjustment reduces the surface area of your body and minimizes the energy cost of running into the wind. But the choice to hunch is made consciously, and what you are really doing as a living organism is trying to make running feel as easy as possible despite the headwind. What Marcora and I are suggesting is that this is what runners are doing all the time, although mostly in subtler ways not involving headwinds.

Perception of effort is tied to brain activity. In crude terms, the harder you have to think about your movements, the harder those movements feel. So your goal as a runner is to run as thoughtlessly and unconsciously as possible. This is why consciously changing the way you run is such a bad idea. No matter what you change, you will have to think more about what you’re doing because what you’re doing is unnatural, and this will make running feel harder.

This does not mean you are stuck with your current stride forever. Every runner’s stride evolves gradually over time as the CNS learns through repetitive practice how to move in ways that feel easier (and that usually also reduce energy cost). Consider something like running with a higher stride rate. If you force yourself to run with a higher stride rate right now, running will feel harder (if only very slightly) and you will probably also be less efficient because you have to think about what you’re doing. But guess what? If you just leave your stride alone and keep training, it is probable that your stride rate will increase somewhat over time, and this natural evolution will be associated with reduced perception of effort and likely also with increased efficiency. You have to earn your stride changes.

Evidence that thinking about your running is bad for your running comes from studies in which researchers look at the effects of attentional focus on running economy. A series of experiments conducted by European researchers has shown that runners become less efficient when they focus their attention internally versus externally, even if their internal attentional focus does not involve any effort to alter their running form. Let that sink in for a minute.

I only wish these researchers had measured the effect of attentional focus on perceived effort as well, because it is perception of effort, not physiology, that directly limits endurance performance. Becoming more economical doesn’t do a runner any good if this benefit comes at the cost of increased perception of effort, as no runner can run faster than he feels he can run. And based on his understanding of what good running form really is, Marcora believes it would be a mistake for runners to try to increase their economy through interventions requiring conscious form changes (i.e., internal attentional focus, i.e., thinking), even if these interventions actually worked. “What happens to perception of effort if we make people run in the way that gets the lowest energy expenditure?” he asked in our email exchange. “I bet it would go up considerably.” And when perceived effort goes up, performance goes down, because the definition of good running form is the particular way of running that feels easiest to an individual runner.

$ubscribe and $ave!

  • Access to over 600 plans
  • Library of 5,000+ workouts
  • TrainingPeaks Premium
  • An 80/20 Endurance Book


30 day money back guarentee

For as little as $2.32 USD per week, 80/20 Endurance Subscribers receive:

  • 30-day Money Back Guarantee