Eliud Kipchoge is so dominant in the marathon that I think his brilliance might even be underappreciated. How can you fully understand the context of history when you’re living in the middle of it? Kipchoge isn’t just great. He outran “great” about 5 years ago. He is now something like a miracle.
On Twitter, coach and writer Steve Magness summarized Kipchoge’s accomplishments. Since 2013, he has won 17 of the 19 marathons he has entered. His two slowest times during that timespan WON OLYMPIC GOLD MEDALS. He also broke 2 hours in an unsanctioned event, and almost broke it another time. This is Sandy Koufax in the 1960s mixed with Michael Jordan in the 1990s mixed with a malfunctioning ATM machine that keeps spitting out $20 bills.
Marathons are captivating because they are physiological and psychological nightmares. A 2021 study in the Journal of Applied Physiology examined the physical characteristics of a sub-2 hour marathon and found that it requires holding an extremely high percentage of an astronomically high VO2 max, with metabolic/aerobic efficiency near the limits of what is conceivable. The study indicated that what seemed physiologically impossible just a few years ago actually just required years of training, otherworldly talent, and work ethic to put it all together. It took an athlete at the 99.99th percentile of humanity in talent and toughness and focus.
If you could draw up that perfect marathoner in a simulation, the result would be Eliud Kipchoge.
I love digging into the training of the GOATs while they are at the peak of their powers. When the margins separating athletes at the top level are so narrow, how can someone put a Grand Canyon between themselves and their competitors?
On trails, we have Kilian Jornet publishing his training logs and Strava files, doing podcast interviews, and sharing every secret. On roads, however, it can be harder to know exactly what athletes are doing and how it fits into the broader context of training theory. This article attempts to decipher some of the clues, following breadcrumbs from articles and interviews to lessons from Kipchoge’s unthinkable dominance.
First, a few disclaimers. One day, I plan to go under 2 hours as I recite the list of reasons I might be wrong in my conclusions from uncertain data. (That day is not today.)
The main hesitation is that there is no definite training log for Kipchoge. My favorite source that I will be referring to constantly in this article was written by Cathal Dennehy for Outside in November 2021. Dennehy had inside access to Kipchoge’s training camp and interviewed his coach, Patrick Sang. That access is key for factual accuracy because there are rumors that some of the online sources purporting to describe Kipchoge’s training are approximations at best, and downright fictional at worst. While there is general agreement on some principles applied by Kipchoge and Coach Sang across interviews and secondary sources, the specifics may remain murky for a while. This article will be trying to describe the outline of a yacht from a foot beneath the surface of the water.
In addition, it’s tough to even know whether the reliable sources tell the entire story. In trail running, there are so many variables that go into top performance (many of which downregulate other variables associated with performance), that there is little disincentive to be fully transparent. Anyone can do Kilian Jornet training, but only Kilian will get Kilian results, because there are countless stumbling blocks between process and outcome on rocky, steep trails.
Road marathons are different. That 2021 study laid out the exact physiological metrics an athlete needed. If someone mastered a training approach to reach those metrics, it’s conceivable that copycats with 99.99th percentile genetics would show up at every marathon ready to go GOAT hunting. There are strong disincentives to printing out 1,000 copies of a treasure map while you’re still busy gathering gold.
Finally, even if the information is approximately correct, it’s still N=1. Plus, there is the unfortunate reality that all world-class athletes face questions about performance enhancing drugs, though to my knowledge there is no reliable evidence here other than very, very fast times. And for the love of all sports, we can’t let top performance alone form the basis for doping accusations (though it’s always healthy to ask questions).
Even acknowledging those disclaimers, there is a wealth of online information that can give us a solid feel for Kipchoge’s general approach, which could have takeaways for everyone. Let’s look at 4 lessons from the marathon GOAT, with added context from the scientific literature and training theory. For a more nuanced discussion, my co-coach Megan and I broke down these principles and more on the newest episode of our podcast.
One: Most of Kipchoge’s training is very easy.
The Outside article has so much fantastic insight, and it wastes no time getting to the sexy stuff. “For Kipchoge, recovery runs start at a shuffle, typically an 8:30-to-8:45-minute-mile pace, and slowly build up to finish around 6:30 to 7 minutes per mile.”
Want to see my favorite video of all time? Check out the Kenyan shuffle to start one of their double runs, filmed by Dennehy. Some people put up a poster of Koufax winding up for a pitch or Jordan taking off from the free-throw line, but I want a poster of Kipchoge shuffling his way to greatness.
Okay, time to put on your speculation helmets! For an athlete like Kipchoge, 8:30 pace is nearly double what he does in a marathon, meaning his heart rate may be as low as 100-110 to start. Even if he accelerates to 6:30s, that’s still a huge percentage slower than his marathon pace, probably equivalent to Zone 1 heart rate in a 5-Zone model.
Now, buckle up your conjecture belt, because it’s MATH TIME. Dennehy broke down an entire training week that gives us clues into the start of one of Kipchoge’s builds (disclaimer: perhaps his training changes a ton later in cycles). It seems that 4 of Kipchoge’s training days are purely easy, consisting of 3 x 18 mile doubles (as 12 miles in AM and 6 miles in PM) plus 1 x 2 hours at easy effort (maybe 16-18 miles). Add onto that 2 x 6 mile easy doubles on his workout days. That puts him around 84 miles that are almost purely easy, and I am guessing firmly in Zone 1. There are probably some miles of warm-up and cool-down for workouts on top of that, also very easy. Let’s estimate 100 easy miles in a week around 130 miles.
That would indicate that 77% is in Zone 1 or just a bit higher! Perhaps that approximation is wrong and some of those miles are in lower Zone 2, or maybe there is more steady running on some of those recovery runs than is reflected in the article. But still, that’s a lot of easy running. People gasped when Jornet’s training summary included 58% in Zone 1, but perhaps that was just a normal range for the GOATs.
The low-end aerobic training supports the high-end training in athletes doing very high volumes. On our podcast, we discussed where an athlete might want to start thinking about specific Zone 1 work as a training focus, and we approximated around 8-10 hours a week. So it doesn’t mean you need to go this easy, but you should definitely feel permission to go very easy relative to your physiological limits. Again, no pace is too slow for aerobic development.
If the GOAT can start at a shuffle, we can all start at a shuffle.
Some people put up a poster of Koufax winding up for a pitch or Jordan taking off from the free-throw line, but I want a poster of Kipchoge shuffling his way to greatness.
Two: Kipchoge does high volume training, but not higher than other world-class athletes.
How does someone become the best of the best? The tempting answer is to say: “By doing more than the wannabe best.” The right answer is more complicated.
Dennehy indicates that Kipchoge does 124-136 miles per week, largely in repeating weekly structure with workout and long run variation. That likely equates to around 16-17 hours of running per week, similar to many of Jornet’s 20-hour weeks. A 2022 review article in Sports Medicine–Open found that world-class marathoners train around 99-136 miles per week and 500-700 hours per year, so Kipchoge falls within the normal range.
In terms of intensity distribution, Kipchoge likely does most of his training purely easy in Zone 1, with some work in Zone 2 on long runs, a lot of Zone 3 tempo in workouts and long runs, a small amount of Zone 4 threshold on workouts, and very little Zone 5 of pure intensity. It’s classically Pyramidal training, with most easy, some moderate, and little hard, a lot like we saw with Kilian Jornet (or with cyclists and skiers).
I think it’s significant that a couple months before his 38th birthday, Kipchoge is still getting faster and faster. He has done that by prioritizing sustainable aerobic development every single training cycle. Discipline with easy training allows athletes to build up volume that supports more economical workouts, creating a positive feedback cycle over the many years required to explore the limits.
Three: Kipchoge runs controlled workouts, with disciplined intensity control.
The hardest place to get athlete buy-in is right here. Faster is often not better in workouts, particularly for advanced athletes.
I can already feel some readers recoiling from that idea. If you can do mile repeats at a 6-minute pace, why would you ever run 6:30s?!
Kipchoge’s training is another pillar of support for controlled workouts. Dennehy quotes Kipchoge: “I try not to run 100 percent,” he says. “I perform 80 percent on Tuesday, Thursday, and Saturday and then at 50 percent Monday, Wednesday, Friday, and Sunday.”
Those 80% days are workout days. The 50% days are the easy run days. Then he shows up on race day with 100% of his capabilities and proceeds to annihilate our wildest imaginations.
Now is a time to reemphasize that we’ve ventured into murky waters. Workout descriptions abound online, but how many are perfectly instructive? Honestly, I don’t know. I have a habit of cleaning out my ear wax with my car keys. I use Chocolate Chex as an iron supplement. Please take everything I say with a grain of salt (I also put salt in my tea before AM runs, with creamer and maple syrup).
But we can approximate a zoomed-out description of his sessions, with the workouts seeming to involve very relaxed intervals between marathon and half-marathon pace, sometimes combined with faster intervals closer to 10k speed. The example given by Dennehy is this combo session, all on a rough dirt track at altitude:
- 8 x 1600m in 4:40 with 2 minutes easy recovery
- 8 x 400m in 63-64 seconds with 30-50 seconds recovery
Those mile intervals look fast, but for Kipchoge, that’s a casual pace around his marathon effort (or a bit harder considering the track and altitude). The quarters are around 10k pace or slightly faster, depending on altitude adjustment. That’s a big session in terms of volume, but probably right around the 80% in terms of his effort (which is mind-blowing on its own).
The principles are great to apply to all athletes–a combination workout of more aerobic intervals followed by speed/power is a strong setup. However, the volume is likely excessive for almost anyone occupying space outside the tail-end of the human talent bell curve. A comparable session for an advanced athlete might be 4 x mile at 1-hour effort progressing to 10k effort, followed by 6 x 200 meters faster.
You probably noticed that there is an offset between 4 x 1 mile at 1-hour effort (what I recommend for most athletes) versus 8 x 1 mile around marathon pace (what Kipchoge did). Why the offset? And now, my fellow intrepid training explorers, it’s time for us to run head first into a training theory brick wall of uncertainty.
That type of strict intensity control looks a lot like Norwegian training, where athletes do many relaxed intervals around threshold effort. But I think there’s a problem. Does strict intensity-controlled training like that done by Kipchoge or the Norwegians work for normal people?
The answer to that question is complex. Yes, it will improve aerobic development at the cellular level. But when an athlete has mechanical limitations that prevent them from running efficiently at these controlled efforts, the workouts may result in an aerobic beast without the mechanical power and efficiency to actually run fast. It’s notable that Kipchoge, Jornet, and the Ingebrigstens were all fast as hell as teenagers. Take someone that is already fast as hell, focus mostly on their aerobic development, and guess what? You have someone that can run fast as hell for long as hell.
But what about everyone else?
In coaching, I have made some compromises around this problem. For intermediate athletes, the controlled intervals are often best around 10k effort–not overly intense to allow for aerobic development, but not so relaxed that they just run slow. For advanced athletes, it’s somewhere between that 10k effort (approximating critical power) and threshold effort (1-hour effort). For pros, it’s usually half marathon effort to threshold/CP. It’s obviously more complex than that on a weekly basis, I just think it’s key to be careful applying the principles that work for a Ferrari to all of us Chevy Volts.
Four: Kipchoge does quality long runs most weeks.
One of the reasons why I got so giddy when I saw Kilian Jornet’s training was the consistent repetition of long tempo runs. Throughout his training, he would do sessions on trails in Zone 3, sometimes over 20-30km. And most sources say that Kipchoge is famous for his long, progressive runs that start easy and end moderately hard. Two GOATS make a trend! That also overlaps with the training approach for Coach Renato Canova, one of the best marathon coaches in history. I bet we’re seeing something very cool about how physiology works in these long events.
Dennehy’s article describes an alternating weekly schedule. One week will be 19 miles, the next 25 miles, over hilly terrain. Other sources describe a run that starts easy, before progressing to steady, and inching down toward 5-minute miles by the end. That’s an extended time of moderate running over hilly terrain, with most likely around Zone 3 tempo.
You see these long, quality runs in the logs of tons of world-class athletes. The likely rationale is that long runs progressing to sustainably fast (but still aerobic) speeds approach the maximal strain on the aerobic system and the underlying lipid oxidation that powers it. That improves the body’s ability to preserve and recover glycogen stores, a key determinant of how hard an athlete can push in endurance events.
I think all athletes can benefit from a bit more steady tempo in long runs. However, it’s key to make sure it doesn’t turn into a hard, race-like effort. Exceeding lactate threshold excessively could counteract some of the aerobic development from the sessions.
And it’s even more complicated in trail running, when heart rate gets higher on climbs. That’s why I’ll often have athletes do their long runs “easy/moderate,” not referring to a specific effort level, but giving permission to move up toward threshold on uphills. After seeing this overlap between Jornet and Kipchoge, I am planning on adding more consistent 1-2 hour controlled tempos around 50k effort for advanced athletes who are not doing many training races (which can serve a similar purpose).
After running the new world record, Kipchoge had a tweet ready to fire off. He does everything fast.
“Limits are there to be broken. By you and me together. I can say that I am beyond happy today that the official world record is once again faster. Thank you to all the runners in the world that inspire me every day to push myself.”
Even in his moment of triumph, he shows gratitude to every other runner. He seems like such a great human. What is it about running that makes our heroes so uniquely kind and caring? Eliud, Kilian, Courtney, Clare, countless others–all wonderful people, using their platforms to spread kindness and love.
I don’t know the answer to that question, but I have a theory. When Kipchoge pushes the limits of what humanity can achieve, he is feeling all of the same chemical and emotional sensations that you or I feel when we push our own limits. Yeah, we might be doing it a bit slower. But I think there is some universal truth to be found at the personal edge.
Peer over the edge, and what do you see? I have done hill workouts where I have lost faith and then found faith over and over and over. Running sucks, and it’s also the most beautiful thing in the world, sometimes simultaneously. It can be a brutal reminder of death and a transcendent affirmation of life.
As runners, we are constantly asking questions: Can I run my first 5k? Can I finish this hill workout? Can I overcome this stress fracture? We all feel the same physical sensations in that journey, a chemical shitstorm of pain and joy. Over time, with patience and grit, we learn that we can do all of those things and more! So we ask bigger and bigger questions, layering leaps of faith on top of one another as we explore our personal edges.
And for a few athletes on the planet, all of that questioning leads to a really big one: Can I run a world record? I think that in saying that we all inspire him, Kipchoge is saying that his world record attempt shares an evolutionary through line with our first mile, our hill workouts, our injuries. His achievement is the pinnacle of a pyramid of questions that all get at a similar idea.
What is actually possible with my time here on Earth?
Therein lies the fun part of an often not-fun sport. You have to layer one thousand leaps of faith on top of each other to start getting a hint of the answer. And the answer is that the answer doesn’t really matter.
But the leaps? Nothing in the world matters more than taking those leaps.
David Roche partners with runners of all abilities through his coaching service, Some Work, All Play. With Megan Roche, M.D., he hosts the Some Work, All Play podcast on running (and other things), and they answer training questions in a bonus podcast and newsletter on their Patreon page starting at $5 a month.