In 1977 Jim’s work took the family to Indiana. Steve joined the middle school football team and found another valuable role model in its coach, Tom Saylor, who would have a lifelong influence on the undersized cornerback. Steve still remembers Coach Saylor’s favorite slogan: “Effort is everything.”
In 1980 Jim and Patricia divorced, and Steve moved back to St. Thomas with his mom. In high school Steve played soccer and ran track. He also took up the guitar and made a little money gigging in local bars with rock and country bands.
When Steve was 20 and in his first year at the University of Windsor, located near Detroit, Patricia died of brain cancer. Steve did his first triathlon soon thereafter, finding a healthy release for his grief in the prolonged suffering of endurance training. In 1989 he switched to bike racing, preferring it, as he would tell others, because the sport seemed even more sadomasochistic than triathlon.
By this time Steve had resumed his college studies, having transferred his credits to Tri-State University. He started a major in physical education and then moved to a dual major in biology and chemistry. He developed a passion for science and saw his future career in it, so he took the next step and entered a master’s program in microbiology at the University of Toledo. As these studies progressed, Steve’s passion for cycling continued to grow. After receiving his degree, he chose to defer life in the real world in favor of racing his bike full time, rambling across North America from race to race with a few fellow “bike bums,” as they called themselves. They slept on floors, packed light, ate cheap, and used whatever bit of money they had gotten from the last race to put gas in the car and get to the next one.
At Toledo Steve had fallen in love with a local girl named Christy Frankowski. There came a time when he had to choose between her and his bike, and he wisely chose her. But he did not want to leave the world of athletics entirely, so he returned to Toledo to study for a PhD in exercise physiology. Steve was fortunate enough to score a job at Eastern Michigan University while the ink of his dissertation on muscle injury was still drying, and he’s been there ever since.
In his first few years at EMU Steve concentrated on developing mathematical tools to quantify training workloads in cyclists, using data collected from bicycle power meters. He decided the logical next step would be to develop similar tools for runners, but he immediately hit an obstacle in the nonexistence of running power meters. He settled on accelerometers as the next best thing. An accelerometer measures changes in the speed and direction of its own movement (or in the movement of whatever it may be attached to). Steve’s efforts to quantify running workloads with accelerometer data were not unsuccessful, but after playing with the instruments for a while he became far more interested in using them to quantify various characteristics of the running stride and in using these measurements to answer such interesting questions as: How does the running stride change through years of training? What are the truly meaningful differences between the strides of gifted and average runners? What happens to the stride when a runner gets tired?
ACCELEROMETERS ARE BUILT into a variety of consumer speed-and-distance devices that runners affix to or insert in a shoe to track their pace and distance traveled in real time as they run. These simple accelerometers capture data in a single plane of movement: anterior-posterior, or forward-backward. Steve uses fancier accelerometers that capture data in all three planes of movement and at much higher resolution than the consumer units. They are able to measure the tiniest changes in a runner’s forward, vertical, and lateral movements.
One of several surprising discoveries Steve has made with the help of these devices is that lateral and vertical movements are not all that important. Although running coaches often instruct athletes to minimize “wasteful bouncing” in their strides, it turns out that the most efficient runners are not the least bouncy—rather, they fall somewhere in the middle of the vertical displacement spectrum. Less experienced runners are just as likely to bounce too little as too much, and in any case Steve’s research has demonstrated that overall stride efficiency is typically not much spoiled either by too much or by too little bouncing.
Only accelerations in the anterior-posterior (AP) plane are strongly associated with running economy. The runners with the smallest accelerations in this plane are the most efficient, regardless of their degree of bounciness or whatever else is going on with their stride. This makes sense when you consider that a runner who exhibits large accelerations in the AP plane is a runner who slows down and speeds up quite a bit from stride to stride, even while trying to hold a steady pace. The greatest source of energy waste in running is the braking that occurs when a foot makes contact with the ground. Those runners who brake the least when a foot lands spend the least energy trying to get back up to speed when pushing off the ground. Steve found that highly trained runners exhibit much smaller accelerations than nonrunners in the AP axis. He also found that, within a population of highly trained runners, those who exhibited the smallest accelerations in the AP plane were the most economical.
The most surprising discoveries in Steve’s research on the running stride came after he began to apply a more arcane mathematical analysis to his data with the help of Erik Bollt, an expert on nonlinear mathematics at Clarkson University. Erik worked out a way to measure the control entropy of a runner’s stride at any given moment. Control entropy refers to the variability or unpredictability of the behavior of a physical system. It can be any physical system, from something as simple as a few gas molecules bouncing around inside a sphere to—well, a runner’s stride. Each system is subject to certain constraints. The looser the constraints are, the more unpredictable the system’s behavior is and the more entropy is said to exist in the system. The tighter the constraints are, the more predictable the system’s behavior becomes and the less entropy is said to exist in the system. For example, suppose a sphere with gas molecules bouncing around inside it shrinks. In systemic terms, this means the constraint represented by the volume of space that the molecules have to move around in has been tightened. Consequently, the total number of possible states in which the system can exist—that is, the total number of different positions the gas molecules can occupy in a frozen moment—decreases. The result is that the system is now more predictable; it has less entropy.
In running, a stride pattern exhibits low entropy when each individual stride looks almost exactly like the preceding one and the next. A stride pattern exhibits more entropy when each stride looks a little different from the one before and the one after. These variations are not typically visible to the naked eye, but the accelerometers can pick up very subtle differences.
One of the first things Steve wanted to look at through the lens of control entropy was the relationship between entropy and fatigue. So he slapped accelerometers on a bunch of subjects and had them walk and then run at incrementally increasing speeds on a treadmill until they quit in exhaustion.
Steve expected to find that entropy increased near the point of exhaustion. This hypothesis was intuitive. Entropy can be thought of as disorder, or chaos. No great feat of imagination is required to conceptualize fatigue in running as a falling apart, or disintegration, of the stride. But Steve found the opposite of what he expected. Control entropy in the strides of his subjects decreased with fatigue. Instead of falling apart, their strides became more locked into a particular pattern—robotic, if you will.
In retrospect, Steve realized he should have anticipated this finding. A constraint suggests a limit. It is clear that exhaustion in running occurs when the body (which, of course, includes the brain) encounters a performance limit. Regardless of where that limit originates, it has to be mediated through the stride, because that’s all running is. An exhausted runner must, therefore, exhibit a constrained stride, or a stride with low entropy.
To give an unrealistically simple example, suppose a certain muscle in a runner’s leg gives out toward the end of one of Steve’s dreadful incremental treadmill runs to exhaus
tion. Unable to use this muscle any longer, the runner will have to adjust his stride, coming up with altered movement patterns that allow him to sustain the same speed. That one muscle, normally preferred but now useless, will act as a constraint on the stride going forward, making the adjusted stride more rigid and robotic—less entropic—than the runner’s normal, nonfatigued stride.
Steve was also interested in comparing control entropy in the strides of nonrunners and experienced runners. He found that entropy tended to be greater in the strides of advanced runners at all speeds. This finding was to be expected, as research in the field of motor learning has long shown that beginners always exhibit less variation in motor skills than practiced experts. Beginners are locked into a certain way of tossing a Frisbee or shuffling cards, whereas experts have acquired a certain amount of freedom, or play, in their movements. They are less constrained by limits on their coordination.
There’s a twist, however, where things get really interesting. While the strides of all runners, novice and advanced, become more robotic, or less entropic, as fatigue increases, and while the strides of experienced runners generally exhibit more entropy, or freedom, advanced runners are usually more constrained, more locked up, than beginners at the point of exhaustion. What does this mean?
The answer lies in a crucial difference between the running stride and our prior example of a constrained physical system—gas molecules bouncing around inside a sphere. Gas molecules move around constantly because that’s just what they do in obedience to the laws of physics. But the act of running is different; running happens because a mind wills it to happen. A fall in entropy therefore indicates more than just the constraining effect of fatigue on a runner’s stride. It also indicates a mental effort to push through that fatigue. As soon as a tiring runner gives up and slows down or stops, entropy skyrockets. The stride is no longer constrained because the runner is no longer pushing against a constraint—no longer trying.
It makes sense that the only way to become really, really fatigued in running is to refuse to slow down despite fatigue. You simply can’t experience extreme fatigue without having exerted extreme efforts to resist fatigue. Fatigue and the fight against it are two facets of a single thing, and control entropy highlights their connectedness. A runner who exhibits an extremely low level of entropy at the point of exhaustion is a runner who has pushed his body deeper into the misery of muscle fatigue than a runner who abandons the treadmill before becoming quite so robotic.
In a strange but real way, then, control entropy allows us to begin to quantify the mental toughness of a runner in action. When subjects raise the proverbial white flag in Steve’s terrible treadmill tests, they always look and feel exhausted. But the control entropy readings that emerge a day or two later, after the numbers have been crunched, lay bare the truth. They reveal who really carved himself hollow and who wimped out. The lower the entropy, the greater the mental toughness. And it happens that experienced runners are generally willing and able to suffer more—in a running task, anyway—than nonrunners. That’s why they exhibit lower levels of entropy at exhaustion even though, by dint of their experience, they start with higher levels of entropy.
It is interesting that low entropy is indicative of inexperience in running, on the one hand, and of extreme fatigue and extreme efforts to resist fatigue, on the other. The link between these two correlations becomes clear if you think of entropy as an indicator of how much a runner’s stride is being mentally forced. Nonrunners have to apply a lot of mental energy to their running, even before they get tired, because they have not practiced the skill much. They’re like a kid taking his first drum lesson, who’s mentally exhausted after two minutes spent trying to tap quarter notes with his left hand and half notes with his right. Experienced runners are able to run with a much quieter brain because repetition has made their stride movements almost automatic. They’re like a professional drummer who can play with all four limbs simultaneously while also singing and feel virtually no mental strain. Advanced runners don’t have to force their stride as much with their minds, and control entropy registers this quiescence of the brain as a certain looseness in the stride.
The more fatigued the muscles become, however, the more mental energy nonrunners and advanced runners alike must apply to their strides. Tired muscles are less responsive to the brain’s whip, so the brain must whip the muscles harder to get the same amount of work from them. Also, the more fatigued a runner becomes, the more suffering he experiences. Thus, an additional amount of “forcing it” is required to resist the psychological temptation to quit—to shout down that old devil on the left shoulder. Control entropy captures this intense application of mental resources by marking an increased rigidity in the stride.
The upshot is that a runner’s brain must be as inactive as possible to turn out the best possible performance. The stride must not be forced any more than necessary. But when approaching the breaking point, a runner wants to have the capacity to force it more than the next guy.
THE FIRST QUESTION any runner will ask when presented with these ideas is how they might be exploited for practical benefit. If the critical characteristics of a superior stride are less braking when the foot hits the ground and looser, less forced movement patterns, then the performance-seeking runner will want to know the best way to develop a freer stride with less braking. A coach as well as a scientist, Stephen McGregor wants to answer this question too. He still has a lot of work to do before it’s been answered fully, but he’s making progress.
Actually, the question was halfway answered before Steve even started to address it formally. After all, one of his earliest studies with accelerometers showed that experienced runners exhibited more entropy and less braking than nonrunners, and the salient difference between non-runners and experienced runners is, obviously, running experience. That alone is pretty strong evidence that simply running a lot over a long period of time makes the stride more efficient. It is an unconscious, automatic evolution.
Some running coaches try to improve their runners’ strides the same way golf instructors teach swings and strokes: by defining correct technique and encouraging (or forcing) their athletes to consciously emulate it. In fact, technique instruction has become quite a vogue in running within the past decade. However, scientific testing of these techniques has consistently shown that making conscious changes to one’s natural stride actually reduces efficiency. It makes no difference what the specific change is. Steve’s work with control entropy explains why. When you make a conscious change to your stride, your brain becomes more actively focused on your running. Your body wants to do what’s natural, but your brain forces it to do otherwise. And forcing it always reduces control entropy.
Steve believes, on the basis of his work, that there is no such thing as correct running form. Yes, there are some general differences, visible to the naked eye, between the strides of nonrunners and those of advanced runners. “But within the population of trained runners,” he says, “there is nothing you can capture on a camera and put on a billboard and advertise as good running form. The runners on our team at Eastern Michigan that we’ve tested all train in a similar way, but they have different running styles. I look at some of them and say, ‘Wow, that person has really horrible form,’ and they actually do very well in our testing.”
The reason for the great variety of running styles observed in high-level runners is straightforward, according to Steve. It’s merely a reflection of the great variety in the structure of runners’ bodies. Two runners with disparate physiques cannot be expected to maximize their individual running economies with precisely the same stride. Efficient running is like a puzzle that each body must solve for itself. And that puzzle cannot be solved consciously. A runner cannot, for example, determine that if his height is X, his inseam Y, and his thigh circumference Z, his optimal stride cadence is therefore 168 steps per minute. The refinement of running form must instead be left to unfold through unconscious trial a
nd error. In much the same way that a species of life figures out how to survive in a changing environment by evolving blindly through random mutation and natural selection, the running stride evolves through repeatedly confronting speed and endurance limits—through crises that challenge the neuromuscular system to come up with novel movement patterns that yield more speed and endurance. You just have to run hard, without thinking about it, and let the process happen. Consciously fiddling with your stride in the hope of accelerating its evolution toward greater efficiency not only can’t help but is almost guaranteed to hurt.
Which is not to say the process can’t be accelerated by other means. Steve has seen some evidence to suggest that runners who train in groups and runners who train at relatively high intensities have better strides than runners of equal experience who train alone and runners who train at lower intensities. He thinks it’s possible that simply trying harder day after day may accelerate stride improvement. Runners who habitually push themselves to keep up with teammates or training partners, or who chase after challenging time standards, raise the stakes on their bodies, pressuring their bodies to figure things out faster. Just as species evolve fastest in a rapidly changing environment that threatens extinction, the running stride may improve most quickly when a runner exacerbates the cost of his stride’s current constraints by pushing against those limits over and over again.
THE EXERCISE SCIENCE laboratory at Eastern Michigan University is located in Room 248 of the Warner Building. It makes an intimidating first impression on the subjects of Stephen McGregor’s grueling running tests. The space looks more like a laboratory than a gym, and in this context the giant Woodward treadmill on which the accelerometer testing is performed looks more like a torture device than a piece of exercise machinery. About forty feet deep and fifteen feet wide, Steve’s lab is stuffed with all manner of equipment, some items identifiably exercise-related, others not. Half of the stuff appears to be in use and the other half in storage, yet it is difficult to discern which is which. A homemade sign is plainly visible on a wall near the treadmill:
Iron War Page 27
