Daily Blog

Magness Speaks — Uncertainty, Randomness, and Over-Control in Training

PTF 2015 Womens 5000_Courtesy of Track and Field Photo.jpg

Control the Controllables

It’s what we preach in coaching and in life. Don’t sweat the small things that don’t have an actual impact on our well being or life. It’s one of the simplest and most profound lessons one needs to learn. And it makes perfect sense.

But like most things, can we take this concept too far? In the world of sport, we like to live in a world where we try to control every variable we can. We go out to Stanford to run in perfect weather, we set up rabbits so that the pace variance is minimal, and we write out months of plans trying to design the perfect training plan to reach our goals. Coaching is a game of control.

We ask our athletes to perform repeats at certain speeds with exact recoveries on a measured course. It’s all about control. Which seemingly eliminates chance, decreases the amount of uncertainty, and gives us the best opportunity to succeed.

All of those are valid points, and I would argue are true.

But have we gone too far?

A World of Uncertainty

What we do when we control is set up defined constraints. Our athletes are allowed to function within this box of constraints that we create. It’s a great situation when trying to elicit some physiologic adaptation, but the reality is this: when we create constraints we also restrict autonomy.

With heavily constrained programs, we’re reducing the athletes’ autonomy in practice to virtually nothing. There are few decisions to be made. It’s simply run X workout as assigned. Because practice is not truly a competition, and if paces are assigned, one doesn’t have to worry about what their teammates are doing around them. They can feel relatively certain that no one is going to throw in a hard surge midway through the mile repeats. They can feel pretty good that plus or minus a few seconds people will run the prescribed time until perhaps the last reps.

The problem is that races, games, and sports aren’t like this. Yes, they have their own constraints in terms of the race distance, game rules, and so forth, but there’s a degree of variability that one can’t control. Even with rabbits, there’s no knowledge of what the other competitors might do, how the pace plays out, or the variations that occur within.

Furthermore, we have to make decisions based on not only the race itself, but also what those are doing around us. When do these decisions occur? During the middle to latter portion of the race when stress levels and fatigue are at their highest.

The contrast is paramount. In practice, we are required to process different information and make different decisions than in race situations.

In the world of ecological psychology, this is referred to as making happening versus doing decisions. We can look at a traditional laboratory based run to exhaustion test where you might do a VO2max test where you run on the treadmill at ever increasing speeds until you decide to stop. This is a happening decision. There’s no active adjustment, it’s one choice you make. On the flip side, races are generally doing decisions. We decide whether to follow the leader, surge when someone picks up the pace, speed up or slow down once fatigue goes, or when to unleash our final kick. We are actively making a choice to do something.

If you don’t think this matters, consider the research on fatigue in exercise science. For years, we had a simple paradigm of how fatigue worked because it was based on tasks to failure studies. It wasn’t until relatively recently that our brilliant researchers realized that things might be just a bit different if we took into account the fact that races are paced and self-controlled. The model for fatigue has entirely shifted.


What we are doing is training for a different test. The test we are taking doesn’t entirely match up with the studying we are doing for it. Perhaps from a physiological standpoint it does, but not from a decision-making and challenge standpoint. In school terms, it’s like we spent our time memorizing definitions for our history exam, when the exam was going to be an essay. Yes, we were studying and gaining the knowledge and there is an indirect transfer, but it wasn’t in the direction in which we were going to be graded in our exam.

While not trying to get too deep here, Dr. Jim Denison, who has done some work with the Canadian Athletics teams, uses the ideas of a powerful thinker named Foucault to look at the mismatch of training demands and racing. In this model, we have an interaction of time, space, and power.

Briefly, these concepts act as constraints to our athletes. So where we train creates natural constraints. For instance, training on a track creates the constraint of running around in a circle at a certain distance where we have norms of running in lanes or doing intervals at set known markers. We have certain starting and stopping points that are well established. From a power standpoint, when we train in groups, we have designations that naturally occur. We have leaders in our group, who if someone “below” them took the lead, there would be an instant upset in power. There’s also the power relationship with the coach in the way that they dictate workouts.

Lastly, from a time standpoint, we are encapsulated by time in our sport. We are inundated with pace information from splits, Garmin data, and so forth. This creates a natural constraint as we have norms and rely on external information to judge most of our progress and effort levels. One could argue that this constraint, which has gotten worse through technology, has reduced the ability to have the internal clock or the ability to read ones own internal sensation of effort.

The basis is similar to the concepts discussed here. We are constraining and controlling our athletes, not giving them the ability to grow and adapt in a variety of situations. We are creating athletes without the autonomy to grow, develop, and make a decision on their own.

Picking Up Clues

The way we make decisions in sport, whether it’s to speed up or slow down in running or to pass the ball in soccer, is by the perception of action possibilities in the environment. Our ability to discern what opportunities are available matters. So for a football star running back, he’s able to notice several different routes to take and almost instantly perceive and act upon the route that provides the best “hole” for him to run through. While our sub-par running back takes longer to discern the possibility or may even miss that hole entirely.

This ability to pick up “clues” from the environment matters. When we train in constrained and controlled environments, we’re not working on our ability to discern and bias ourselves towards the best decision to be made. In practice in a controlled environment, we might miss the repercussions of a subtly varied pace or the ability to recognize and anticipate someone else’s final drive to the finish.

Uncertainty, randomness, and so forth, therefore, should not be seen as things to eliminate, but instead as an additional way to stress the athlete to adapt. They put us in a situation not only physiologically but also psychologically to be better prepared for a variety of situations we will encounter during a performance.

We can see the repercussions in life and sport. In sport, we’ve all dealt with the athlete who can perform perfectly under controlled situations like set up time trials but would fail once they got to the unfamiliar situation of a tactical trial. In life, many diseases we have now are diseases of modernity, where it is the lack of stress that causes them. We don’t have enough variation, randomness, and stress from which to adapt off of. Or as Nassem Taleb points out:

“It is the systematic removal of uncertainty and randomness from things to make matters highly predictable in their smallest details. All that for the sake of comfort, convenience, and efficiency.”

This eternal search for efficiency has consequences. In the realm of health and in life it can create a complacency and lack of ability to adapt in disparate situations.

So What?

Don’t get me wrong; I love planning out the physiological adaptations in a systematic way. It’s a cornerstone to coaching, but perhaps we need just enough randomness, uncertainty, and flat out not knowing to challenge our athletes in a different way. After all, as was the theme in a wonderful conference I participated in put on by the Seattle Sounders, we’re trying to build Anti-Fragile athletes. Based on Nassim Taleb’s book, the idea is to build athletes who thrive off of uncertainty.

Therefore, while we shouldn’t overreact and get rid of the physiological approach to training, and we shouldn’t give up the wonderfully useful concept of controlling the controllables, perhaps we could build better prepared athletes and people with a degree of randomness. Don’t throw the baby out with the bathtub, but instead interject a degree of uncertainty to your training.

Early in my career, I made the mistake of trying to create athletes who were physiologically well adapted but could not thrive in randomness. They could kill the time trial but would crater when it came to the unfamiliar territory of championship racing. This isn’t a problem in certain circles of elite performance where athletes and coaches will contrive to set up races in the exact way they want them, but in the world of tactics, uncertainty, and stress.

But in the real world, it is an issue. And it stems from the dynamics of the athlete and the training we set up. As Denison alludes to in his work on Foucault, is it no wonder that athletes sometimes falter when they are used to turning their brains off and being challenged in only the direction of effort, where they are left with “happening” decisions instead of “doing” ones?

Instead of constraining our athletes, challenge them in ways to grow and develop. There’s an unending amount of ways to change the power dynamics during training. Open up the constraints and challenge athletes to make decisions. Then perhaps we can create athletes who thrive in situations ranging from tactical affairs to one-off time trials. Coaching is all about manipulating the constraints you put on the athlete, to adapt and grow in the direction you desire.


Any questions? You can send me a Direct Message on Twitter.  Thank you for reading.| SM

*note: a previous version of his post was originally published by Steve Magness on The Science of Running on Dec. 19, 2014. It has since been revised and updated.

Steve Magness