Computational Marathoning

a great article brought to my attention by a fellow marathoner and RBS'er, Laurie B...thanks Laurie!

Oct. 21: A New Science Is Born–Computational Marathoning
10/21/2010 5:01 PM
If the New York City Marathon offered its top prize money to the smartest guy in the field, I'd be betting on Benjamin Isaac Rapoport on Nov. 7. He's not likely to be on the podium, however. He's hoping for a PR around 2:50.Rapoport notched a recent PR in last April's Boston Marathon, clocking a 2:55:11. Most runners follow up their Boston Marathon finishes with a few celebratory beverages and a nice restaurant meal. Not Rapoport. He had an afternoon appointment in his calendar last April: After finishing the Marathon, he had to teach an endurance physiology class at Harvard Medical School.Rapoport has taught the same class after his Boston Marathon finishes for a handful of years. Last April, a lightbulb went off: He ought to share his research and conclusions with a wider audience. The result is the cover story in the new issue of PLoS Computational Biology. It's titled "Metabolic Factors Limiting Performance in Marathon Runners." Which means, in brief, that it's a "computational study" of how to optimize your marathon pace and the carbohydrate consumption necessary to hold that pace for 26.2 miles.http://www.ploscompbiol.org/home.actionMost runners recognize that these two issues–pace and carbs–are crucial to successful marathons. Everyone struggles to find their perfect pace, and everyone packs away carbs before and during the marathon. However, not everyone succeeds. Sometimes we crash and burn. Or, as we runners often say it, we "hit The Wall."Rapoport, 29, who will run his 19th marathon in a week in New York City, says he has only hit The Wall all hard once, in the 2005 NYC race. But the experience was bad enough, and his curiosity strong enough, to put him on a quest to eliminate The Wall through precise calculations. An MD student at Harvard Medical School, as well as a PhD student in electrical engineering at MIT, Rapoport hopes to become a neurosurgeon who implants electric circuits in damaged human brains.http://www.rle.mit.edu/avbs/whoweare_ben.htmIn his new paper, Rapoport sticks to the more mundane but still tantalizing challenges of the marathon. First he points out that carbohydrate loading makes a big difference. A 154-lb man who doesn't load effectively will find his glycogen supply stuck at about: 20 calories in the blood, 350 in the liver, and 1250 in the leg muscles. That totals 1620 calories. Unfortunately, the same guy needs about 2900 calories to finish the marathon. Oh, oh, big trouble ahead.Yet if this runner loads to the max, he will have: 20 calories in the blood, 650 in the liver, and 2270 in the leg muscles. That totals 2940 total calories. Now we're talking.But glycogen supplies are only half the game. The other half is your fuel efficiency, AKA your marathon pace. If you run too hard, your body selectively chooses carbohydrates as fuel and you risk depletion. This is known as going out too fast and bonking. If you slow a little, your body can burn more fat, sparing some of the glycogen, and you can reach the finish line with no Wall in sight.There's a third half too. If you consume some carbs during the marathon, you can modestly extend your range. That's why we have sports drinks, gels, and bars. They can help us avoid The Wall.The mathematics in Rapoport's paper are, frankly, beyond most of us (me, anyway). But he hopes to develop an online calculator any day now to simplify the process. After all, the main goal behind his work is to make "computational marathoning" (my words, not his) accessible to real-world marathoners. He wants runners to understand that The Wall is not an inevitability; it only results from inexact science. (As soon as he tells us about the calculator, we'll post a link here.)Below, Rapoport answers a few questions about his paper and his personal approach to training, carbohydrate loading, and pacing the marathon.Q: Do you follow the classic 7-day depletion and reloading carb diet, or do you just load the last several days?Benjamin Rapoport: I haven't found any good evidence that favors the full-depletion approach. I think the key issue is to eat mostly carbohydrates (I happen to prefer rice) and very little fat and protein the last three days before a marathon, and to take your big meals shortly after a workout. That way you get the highest percent of glycogen loading in your leg muscles. There's also evidence that you can increase your leg glycogen supplies if your final workouts include some high-intensity running. I do just 30 or 40 minutes easy the last three days before a marathon, but I finish up with a couple of minutes quite hard.Q: If runners with big muscles can store more glycogen, why don't big runners race marathons faster than smaller runners?BR: Big runners are like heavy trucks that have large gas tanks, but have low fuel economy (five miles per gallon). Elite runners are like good motorcycles that have small gas tanks but high fuel economy, so they get 60 miles per gallon. Because of this high fuel economy, the elites can run fast over a given distance without depleting their carbohydrate reserves. But bigger guys risk running low on fuel.Q: You acknowledge that midrace fueling doesn't have a large effect like effective carbohydrate-loading, but you still believe it important. Do you have a personal preference during your races–drinks, bars, or gels?BR: I like to take a gel pack–I've got no preferences among the various brands–about every 5 miles. I have a pet peeve about the sports drinks at marathons. The drinks have undergone years of validation, and we know they work at the precise intended concentrations. But when they're mixed up from dry mixes by marathon volunteers, you don't really know what you're getting. They could be too dilute or too strong, and neither is good. That's why I like gels–you know precisely what you're getting.Q: Caffeine?BR: Sure, it's been pretty well shown to help in endurance events like the marathon. There are basically two issues: What's your tolerance? And how much should you be taking for your body mass? Someone who consumes coffee on a regular basis might need more caffeine to get a performance effect. And of course no one really knows how much caffeine they're getting from coffee and other drinks. That's why I think caffeine tablets are better. At least they work for me. Again, you know exactly how much you're getting. There's some evidence that adding niacin to the caffeine makes it more effective more quickly.Q: How has your training evolved over the years, and brought you to faster marathon times?BR: I've learned more physiology. I've trained harder. And I think I've trained much smarter. I used to hate speedwork–I considered it drudgery. Then one day I just decided to start liking it. And ever since then, about three years ago, I have. I'm pressed for time like everyone else. Most days I only have about 45 minutes to work out. So I'll do several workouts at my marathon pace or faster, and then longer runs on the weekend. My speed workouts tend to be three cycles of two-mile repeats including a half-mile in 6:30, my goal marathon pace; a mile at 6:00; and a half-mile at 5:30 pace. On the weekends, I'll start from my base long run mileage of 11-12, with increases of 1-2 miles a week, and a maximum length of 24-25 miles. Then I taper for two to three weeks, decreasing mileage significantly, but continuing the speedwork. On my recovery days between runs, I do 45 to 60 minutes of riding on an indoor bike while also reading and studying.Q: A lot of what's in your paper is already known to marathoners–they have to carb load and run the right pace. Is your major contribution that you're providing more "numbers" to make the process more exact?A: I certainly hope so. That has been my goal from the start. I've heard that others beside you think the math is too hard, so that's why I'm working on the online calculator. The thing is, so many runners don't get the marathon right because they don't know how to make the individual adjustments. In the marathon where I bonked badly, New York in 2005, I had trained really well, but I got the carb loading all wrong. I want people to know that indiviIf the New York City Marathon offered its top prize money to the smartest guy in the field, I'd be betting on Benjamin Isaac Rapoport on Nov. 7. He's not likely to be on the podium, however. He's hoping for a PR around 2:50. Then it's back to his MD studies at Harvard and his PhD engineering classes at MIT.
If the New York City Marathon offered its top prize money to the smartest guy in the field, I'd be betting on Benjamin Isaac Rapoport on Nov. 7. He's not likely to be on the podium, however. He's hoping for a personal record of around 2:50. Then it's back to his M.D. studies at Harvard and his Ph.D. engineering classes at MIT. [Photo right: Rapoport running in Cambridge, Mass. Photo courtesy Patrick Gillooly/MIT.]


Rapoport notched a recent PR in last April's Boston Marathon, clocking a 2:55:11. Most runners follow up their Boston Marathon finishes with a few celebratory beverages and a nice restaurant meal. Not Rapoport. He had an afternoon appointment in his calendar last April: After finishing the marathon, he had to teach an endurance physiology class at Harvard Medical School.

Rapoport has taught the same class after his Boston Marathon finishes for a handful of years. Last April, a light bulb went off: He ought to share his research and conclusions with a wider audience. The result is the cover story in the new issue of PLoS Computational Biology (free, full text here). It's titled "Metabolic Factors Limiting Performance in Marathon Runners." Which means, in brief, that it's a "computational study" of how to optimize your carbohydrate consumption and marathon pace to achieve a maximal-for-you performance over the full 26.2 miles.

Most runners recognize that these two issues–carbs and pace–are crucial to successful marathons. Everyone packs away carbs before and during the marathon, and everyone struggles to find their perfect pace. However, not everyone succeeds. Sometimes we crash and burn. Or, as we runners often say it, we "hit The Wall."

Rapoport, 29, who will run his 19th marathon in New York City, says he has only hit The Wall hard once, in the 2005 NYC race. But the experience was bad enough, and his curiosity strong enough, to put him on a quest to eliminate The Wall through precise calculations. An M.D. student at Harvard Medical School, as well as a Ph.D. student in electrical engineering at MIT, Rapoport hopes to become a neurosurgeon who works with electronic circuit board implants in damaged brains.

In his new paper, Rapoport sticks to the more mundane but still tantalizing challenges of the marathon. First he points out that carbohydrate loading makes a big difference. A 154-pound man who doesn't load effectively will find his glycogen supply stuck at about: 20 calories in the blood, 350 in the liver, and 1,250 in the leg muscles. That totals 1,620 calories. Unfortunately, the same guy needs about 2,900 calories to finish the marathon. Uh oh, big trouble ahead.

Yet if this runner loads to the max, he will have: 20 calories in the blood, 650 in the liver, and 2,270 in the leg muscles. That totals 2,940 total calories. Now we're talking.

But glycogen supplies are only half the game. The other half is your fuel efficiency, AKA your marathon pace. If you run too hard, your body selectively chooses carbohydrates as fuel and you risk depletion. This is known as going out too fast and bonking. If you slow a little, your body can burn more fat, sparing some of the glycogen, and you can reach the finish line with no Wall in sight. Of course, before your next marathon, you can also train harder, or lose a few pounds (within reason), to raise your VO2 max, and hence your marathon efficiency.


There's a third half too. If you consume some carbs during the marathon, you can modestly extend your range before bonking. That's why we have sports drinks, gels, and bars. They can help us avoid The Wall.

The mathematics in Rapoport's paper are, frankly, beyond most of us (me, anyway). But he has developed an online calculator at endurancecalculator.com to simplify the process. [Ed: Your computer must have Java to view this calculator.] After all, the main goal behind his work is to make "computational marathoning" (my words, not his) accessible to real-world marathoners. He wants runners to understand that The Wall is not an inevitability; it only results from inexact science.

Below, Rapoport answers a few questions about his paper and his personal approach to training, carbohydrate loading, and pacing the marathon.

Q: First, you seem a pretty smart fellow. But did you pass the Boston Marathon's final-exam online registration test on Monday for getting into the 2011 race?

Benjamin Rapoport: Yes. Judging from my confirmation number, I think I was the 12th person to register. I started clicking at 8:59 Monday morning, was done five minutes later, and oblivious to the fiasco until someone mentioned it to me the next morning.

When tapering for a marathon, do you follow the old seven-day depletion and reloading carb diet, or do you just load the last several days?

I haven't found any good evidence that favors the full-depletion approach. I think the key issue is to eat mostly carbohydrates (I happen to prefer rice) and very little fat and protein the last three days before a marathon, and to take your big meals shortly after a workout. That way you get the highest percent of glycogen loading in your leg muscles. There's also evidence that you can increase your leg glycogen supplies if your final workouts include some high-intensity running. I do just 30 or 40 minutes of easy running the last three days before a marathon, but I finish up with a couple of minutes quite hard.

If runners with big muscles can store more glycogen, why don't big runners race marathons faster than smaller runners?

Big runners are like heavy trucks that have large gas tanks, but low fuel economy (five miles per gallon). Elite runners are like good motorcycles that have small gas tanks but high fuel economy, so they get 60 miles per gallon. Because of this high fuel economy, the elites can run fast over the marathon distance without depleting their carbohydrate reserves. But bigger guys risk running low on fuel if their efficiency is too low.


You acknowledge that midrace fueling doesn't have a large effect like optimal carbohydrate-loading, but you still believe it's important. Do you have a personal preference during your races–drinks, bars, or gels?

I like to take a gel pack about every five miles–I've got no preferences among the various brands. I have a pet peeve about the sports drinks at marathons. The drinks have undergone years of validation, and we know they work at the precise intended concentrations. But when they're mixed up from dry mixes by marathon volunteers, you don't really know what you're getting. They could be too dilute or too strong, and neither is good. That's why I like gels–you know precisely what you're getting.

Caffeine?

Sure. It's been pretty well shown to help in endurance events like the marathon. There are basically two issues: What's your tolerance? And how much should you be taking for your body mass? Someone who consumes coffee on a regular basis might need more caffeine to get a performance effect. And of course no one really knows how much caffeine they're getting from coffee and other drinks. That's why I think caffeine tablets are better. At least they work for me. Again, you know exactly how much you're getting. There's some evidence that adding niacin to the caffeine makes it more effective more quickly.

How has your training evolved over the years, and brought you to faster marathon times?

I've learned more physiology. I've trained harder. And I think I've trained much smarter. I used to hate speedwork–I considered it drudgery. Then one day I just decided to start liking it. And ever since then, about three years ago, I have. I'm pressed for time like everyone else. Most days I only have about 45 minutes to work out. So I'll do several workouts a week at my marathon pace or faster, and then longer runs on the weekend. My speed workouts tend to be three cycles of two-mile repeats including a half-mile in 6:30, my goal marathon pace; a mile at 6:00; and a half-mile at 5:30 pace. [Ed: This is roughly equivalent to marathon pace/lactate threshold pace/VO2 max pace; also, marathon pace/tempo pace/intervals pace.] On the weekends, I'll start from my base long-run mileage of 11-12 miles, with increases of 1-2 miles a week, and a maximum length of 24-25 miles. Then I taper for two to three weeks, decreasing mileage significantly, but continuing the speedwork. On my recovery days between runs, I do 45 to 60 minutes of riding on an indoor bike while also reading and studying.

A lot of what's in your paper is already known to marathoners–they know they should carbo load and run the right pace. Is your major contribution that you're providing more "numbers" to make the process more exact?

I certainly hope so. That has been my goal from the start. I've heard from a number of readers that they think the math is too complex, so that's why I developed the online calculator. The thing is, so many runners don't get the marathon right because they don't know how to make the individual adjustments. In the marathon where I bonked badly, New York in 2005, I had trained really well, but I got the carb loading all wrong. I want people to know that individual quantification is relatively easy to do, and that it can be essential for the individual runner. It's the only way to know exactly what you need to do.