What I’m Doing to Run My Fastest Marathon Ever

As a corollary goal in my Rubber Decade, I hope to eclipse my personal record in the marathon. Set when I was fully ten years younger and about ten pounds lighter, my PR of 3:13:23 has haunted me. With each passing year, that finishing time appears more and more like a flash in the pan.

At the time, I had no structured training plan—no interval workouts on the track, no consideration for carbohydrate intake, no means to monitor my heart rate while training. I simply ran as often and as far as my college schedule allowed.

Given the present demands on my time, a body entering its fourth decade, and a near-pathologic desire for order and structure, I can no longer approach marathon preparation with caprice.

For those uninterested in the full post, here is a summary of my strategy:

1. Training in specific heart rate ranges rather than target pace ranges

2. Performing the majority of my training at a low aerobic intensity

3. Leveraging indoor sessions during winter months

4. Augmenting carbohydrate and electrolyte intake

5. Utilizing weight-lifting as an adjunct for building muscular endurance

   ---

Over the last six weeks, I began training in earnest for the Vermont City Marathon, held on Memorial Day weekend in Burlington. This is where I plan to run my fastest marathon yet, averaging a pace of at least 7:21/mile (about 4:35/km).

My 20-week training block is periodized into four components—a six-week “base” phase to strengthen aerobic endurance, a ten-week “build” phase to promote adaptations to improve race performance, a two-week “peak” phase with more race-specific workouts, and finally a two-week taper with reduced effort directly preceding Race Day.

Target Heart Rate Range > Target Pace Range

When preparing for previous races, I almost always had a specific pace goal in mind for training sessions—say, a five-mile run at “easy pace” or a workout of 8 x 400 meters at “5K pace.”

COROS, a fitness technology company, disabused me of this mindset. Sessions within their free training plans exclusively employ heart rates zones based on one’s estimated lactate threshold heart rate (LTHR).

As the topic requires more exploration of physiology than is necessary here, see the Appendix for further discussion of lactate threshold.

Similar to the use of functional threshold power (FTP) zones in cycling, the use of prescriptive heart rate zones allows for more precise, personalized training parameters. Constructing sessions around these zones might also permit easier monitoring of progress, faster training adaptations, and a higher likelihood of achieving goals.

COROS uses a six-zone model for threshold heart rate ranges, although several other paradigms exist.

I spend the majority of my training in what COROS defines as “aerobic endurance,” a heart rate range between 80 to 90% of LTHR.

Over the last six weeks, I have been pleasantly surprised to see my average pace during aerobic endurance runs steadily decrease, a likely indication that my base aerobic fitness has increased.

A similar increase in base fitness could be appreciated from lower average heart rates during “easy pace” runs, but observing this pattern provides less insight to promote further improvements in fitness. Sure, one’s “easy pace” could be made faster, though it would not be immediately clear what the pace is being titrated to—a specific average heart rate? RPE?

Aiming to run within the bounds of a heart rate range takes some of the guesswork out of the equation. When more fatigue is present (observed either subjectively or by alterations in heart rate variability or resting heart rate) or hydration status is suboptimal, pace may need to slow to maintain the same heart rate zone, preventing one from over-reaching and risking injury.

Conversely, when one is more recovered and primed for effort, faster paces might be achievable within the parameters of the same heart rate range, allowing one to more sustainably build aerobic fitness without accumulating excessive stress.

These proclamations notably take for granted that heart rate will be accurately measured. Optical heart rate monitors embedded in GPS watches tend to be less accurate than optical sensors worn on the arm or electrodes in a chest strap.

(Anecdotally, I’ve found optical HR measurement to be atrocious in COROS and Garmin watches, but fairly accurate, albeit spotty, in different versions of the Apple Watch. I currently use an arm-band heart rate monitor from COROS, which tends to provide very accurate readings.)

Note that while I am currently a true believer in COROS’s system and products, I hold no financial stake. There are no affiliate links found anywhere on this platform.

80-20 Training

Exercise physiologist Stephen Seiler is credited with popularizing the 80-20 training principle within endurance sports. His observations of elite athletes suggest that about 80% of their sessions occur at a low intensity while the remainder are at moderate to high intensity.

The definition of “low intensity” widely varies in the literature but is generally held to be an aerobic effort, similar to our working definition of Zone 2. In other words, one’s rate of perceived exertion (RPE) is about 3 to 4 out of 10, one could sing to the letter “G” in the alphabet, and heart rate remains in a range roughly 65 to 80% of maximal heart rate.

Again, I use COROS’s “aerobic endurance” zone for low-intensity efforts—80 to 90% of my LTHR.

Importantly, the 80-20 principle applies to the number of sessions in a given week rather than the volume of time spent in either low-intensity or moderate- to high-intensity efforts.

For instance, this week of training contained five sessions—four sessions (80%) were run in the aerobic endurance zone (excluding a warm-up and cool-down) while my fifth session (20%) included intervals at threshold intensity (96-102% of LTHR). In terms of actual distribution, however, about 85-90% of my overall time spent running was in the aerobic endurance zone.

Prioritizing lower-intensity efforts might help augment base cardiopulmonary efficiency, reducing effort during higher-intensity sessions. Low-intensity efforts also tend to be less fatiguing, permitting higher performance during more vigorous workouts. Finally, there is generally less injury risk with lower intensity exertion.

Indoor Sessions

Jakob Ingebrigtsen is perhaps the most elite middle-distance runner in the world, currently laying claim to the world record at the 1500m, 2000m, and 3000m distance. Ingebrigtsen continues to mainly live and train in Sandnes, Norway, not exactly a Mecca for elite-level training. (Sandnes is almost an anagram for “sadness.”)

Now I know a lot of people are beginning to compare me to Jakob, and I’m flattered—I really am—but I can’t quite hold a candle to him yet.

Where we are similar, though, is in prioritizing indoor sessions during the cold, dark winter months, be they in Norway or New England. Knowing myself, I would have reneged from several planned runs if my only option was training outdoors in sub-freezing temperatures at 5 AM. The treadmill allows me to remain consistent throughout my training block and continue building fitness.

Given the high occurrence of inclement weather conditions, the treadmill also facilitates controlled sessions. “I can set a pace and forget it, without having to deal with wind, corners, or any other conditions,” agrees Ingebrigtsen.

The extender feature on my COROS device has been central to my treadmill efforts. During my run, my phone displays important metrics from my watch, including heart rate, pace, and distance, without the need to look down at my wrist constantly. (This feature is also found on the Apple Watch.)

Moreover, COROS allows users to calibrate the distance of indoors runs after completion, making pacing data more accurate.

Carbohydrate Intake

David Roche shocked the world with his course record at the Leadville ultramarathon in August 2024. Central to Roche’s success was his fueling strategy. Over the 100 mile race, he consumed about 120 grams of carbohydrates every hour, mainly in the form of lightly viscous gels containing maltodextrin and fructose.

Roche, who coaches elite athletes alongside his wife Megan, another ultra runner, advocates for “GI training” during endurance race preparation. Adapting the stomach and small intestine to tolerate progressively higher levels of carbohydrates may help optimize performance outcomes during endurance efforts.

At present, carbohydrate intake during training is more of an art than a science. For sessions lasting longer than an hour, Roche suggests starting with 60 to 75g of carbs per hour, with consumption occurring each half-hour. For efforts longer than 90 minutes, aiming for 75 to 90g of carbs per hour might be more helpful. Other recommendations suggest only exceeding 60g/hour for efforts longer than two hours.

In any case, consuming more carbs likely contributes to better performance and recovery.

This strategy, however, does not come cheap. Roche’s favored gels (SiS Beta Fuel, which contain 40 grams of carbs in a single serving and are perpetually out of stock) are $2.50 each. Maurten, another leading brand, charges about $5.20 for each 40-gram gel.

For reference, 40 grams of carbs in old-fashioned oats costs $0.07. A steaming bowl of oatmeal, of course, would not be fun to eat while running.

Always the frugal consumer, I’ve formulated a poor man’s version of GI training using unflavored maltodextrin powder dissolved in water. Each 40-gram serving costs about $0.45.

I typically mix in about one-third to one-half of an LMNT packet (one of the only use cases for the supplement in my opinion), which costs another $0.25 to $0.50, to improve taste and add much-needed electrolytes (roughly 300 to 500mg sodium). One 500ml serving, then, costs about $1, roughly 40% of the cost of a SiS Beta Fuel gel.

This method is certainly more feasible during indoor sessions, when I can easily reach for a water bottle while on the treadmill. Training outside using this approach would require me to carry multiple bottles or flasks, or routinely double-back to my vehicle to re-fuel.

I recognize that during later phases of my training, reliance on gels will be inevitable. Better start saving now!

Weight-Lifting

Historically, most long-distance runners shied away from resistance training. Ten years ago, I asked for strength training advice from my university’s cross country coach. He said that core-strengthening movements like planks and crunches were far more important than lifting weights. (The same coach was later profiled in the New York Times for promoting “a rampant culture of body shaming and eating disorders,” but that’s neither here nor there.)

To be sure, core strengthening has its place, but more deliberate resistance training is increasingly recognized as a vital component of race preparation. Past work shows that resistance exercise helps to enhance running economy and explosive strength while reducing injury risk. For their part, Clayton Young and Conner Mantz, who represented the U.S. in the 2024 Olympic marathon, regularly visit the weight room during training blocks.

Most running-specific strength routines predictably focus on lower-body movements, such as squat variations, deadlift variations, lunge variations, leg curls, and calf raises. Most recommendations utilize a higher repetition range (about 12-15) performed in 2-3 sets, often making no reference to RPE or a related concept called repetitions-in-reserve, or RIR. (See more here.)

Currently, I perform two strength-training sessions each week—one focused on upper-body movements and another for lower-body movements. The sessions take about 45 minutes each. I use a rep range between 8 to 15 reps to an RIR of 0 to 1 for each movement, usually for one or two sets. Admittedly, this component of my training is not as optimized as other components.

This is my current lower-body routine:

• 2x10 barbell Romanian deadlift (rest for 2 minutes between sets)

• 2x12 leg extension as a drop-set (second set is completed immediately after the first set at 50% less weight)

• 2x12 leg curl as a drop-set

• 1x10 dumbbell lunge

• 2x12 dumbbell Bulgarian split-squats (rest for 90 seconds between sets)

• 2x15 standing calf raise (rest for 30 seconds between sets)

Appendix

Recall from previous posts that lactate threshold refers to the point at which lactate builds up more quickly than it can be processed by mitochondria to create ATP.

As with VO2 max, the gold standard for lactate threshold measurement occurs in an exercise physiology laboratory. Blood samples are collected at regular intervals while the participant performs a graded exercise protocol, which gradually becomes more difficult.

COROS, as well as competitors Garmin and Suunto, provides estimations of LTHR through proprietary algorithms from wearables data. Each company offers users a “test” workout that requires an all-out effort or uses a graded protocol.

Joe Friel, an endurance coach who pioneered LTHR-based training, provides a lower-cost method for approximating LTHR: perform a 20-minute all-out effort and take 95% of the average heart rate over that period.

Again, heart rate needs to be measured accurately, ideally by a chest strap or arm band, for these estimates to be valid.

Paradigms for heart rate zones based on putative LTHR differ substantially. COROS, for instance, uses six heart rate zones, whereas Joe Friel uses seven. At present, there isn’t an exacting science for which framework might be more useful.

Personally, I’m not going to lose sleep over whether my aerobic endurance zone—where I spend the vast majority of my training, as noted above—lies between 80 to 90% of my LTHR, or more tightly between 85 to 89%.

As a final note, the aerobic endurance zone is also called “zone 2” under the COROS framework. Although this definition differs from the one I use to define Zone 2 in an energy utilization context (i.e., Zone 2 is when fat oxidation is maximized by mitochondria), it is likely true that “fat max” occurs within the aerobic endurance zone. Again, not something worth perseverating over.