This is a follow-up to last week’s article. This post will attempt to clarify the misconception that dynamical systems theory (DST) and nonlinear pedagogy (NLP), which for the sake of simplicity I will use interchangeably, are solely game-based approaches to coaching. To highlight this fallacy, we’ll define both open and closed skills - which is often a poorly understood topic in and of itself. Finally, we’ll take a look at the complexity of learning through the lens of fixed versus variable movement patterns - what they are, how they’re developed and why both are necessary qualities for skilled movement execution.
There isn’t a simple conclusion when it comes to the ideal approach when developing (and training) the elite tennis player. Motor learning in general is still a young field and we still have lots to learn (personally, I'm one of it's biggest students!) - but I'll try to present the information that’s available to us and attempt to provide further insight into this topic. If you haven't read the intro post, I suggest you do so before reading this article. You can review that post here.
Game-Based Coaching and Nonlinear Pedagogy are NOT the Same
As I previously mentioned, many associate DST and NLP with a strictly ‘game based approach to coaching’. Although game-based coaching plays a large role during NLP, they are by no means one and the same. If you’ve never heard of a game-based approach, it’s quite simple. It’s premise is based on emphasizing drills that are directly transferable to real-world tennis scenarios. In other words, the emphasis is on developing technique as it relates to a tactical intention. In contrast, a model-based approach to coaching would emphasize technique based on some sort of ideal WITHOUT a tactical outcome in mind. These would be considered the two extremes of each methodology.
A quick note, many tennis federations have adopted more of a game-based approach to their coaching development systems.
To highlight the distinction between game-based and model based coaching, let’s look at an example. A beginner wants to learn to play tennis and hopes to someday compete in club matches. Using a game-based approach, instead of hitting a basket of forehands from day one (a traditional, model-based approach), the coach attempts to rally with the student in a mini-tennis situation - a necessary skill in tennis. Right away, the coach would either directly or indirectly guide (key word here) the student to solve the problem of getting the ball over the net and landing inside the service line. This could be done in a variety of ways including noticing the height of the ball, the distance a player should stand to meet the ball with an appropriate impact point and so on. Let’s say the student is an adult and you began the mini-tennis game using real tennis balls. You then quickly notice your student is unable to handle the speed and bounce of the ball. You then replace it with a foam ball. As the student improves their ability to rally with the foam ball, you decide to make the task more difficult by changing the ball, adding movement, varying the spin and so on.
This example of game-based coaching sounds a lot like NLP doesn’t it? The coach is setting up constraints - the guiding principle of NLP. It’s true that game-based coaching is a big part of NLP, but they are NOT the same thing. Let's explore.
Open vs Closed Skills...It's Not Black & White
Before we highlight why NLP and game-based aren't exactly the same, it’s important to make the distinction between open and closed skills or situations. The following is a definition of the two (Bosch 2015):
“A closed skill is a movement pattern in which the movements to be made are predetermined because the environment in which they are made is unchanging. With open skills, the environment is not unchanging, and the movement must therefore be adapted (improvised) in response to the demands of the environment at that moment.”
With these definitions in mind, an example of a closed skill sport would be diving - the athlete knows what height they'll be diving from and what their diving sequence/routine will entail, all before they enter competition. Tennis, on the other hand, actually has characteristics of BOTH closed and open skills. Serving is usually considered a closed skill - the movement is predetermined - while rallying is considered an open skill - the environment is constantly changing and players are consequently always adapting. To illustrate the two skills, let's use our rally example from earlier. Because your student must first perceive the oncoming shot, how fast it's coming, what the spin looks like and so on, AND THEN make a decision about the execution of the shot, the situation would be classified as open. In contrast, ‘closed skill practice’ in tennis might be something like hitting crosscourt forehands from a hand-fed position or serving out of a basket - with no clear tactical intention.
But can skills be solely open or closed? The answer is an emphatic NO. Skills cannot be classified as either completely open or completely closed. For example, is tennis serve practice with no returner the same as serving against an opponent? Will the positioning of your opponent affect where you serve? Is serving against an opponent and no crowd the same as serving in the final of the US Open, into the wind and with the sun in your face? Not a chance. Gentile (2000), demonstrated that there’s a transition from extremely closed skills to completely open skills - he developed 16 categories to classify the range in available skills. Based on his classifications, shadow serving without a racquet would be completely closed while rally situations in a competitive match (where environmental factors are at play) would be completely open. The exact details of each category aren't necessarily critical at this point but as it relates to NLP, it's important to understand the distinction between open and closed skills and their various intricacies, as this will affect the organization of training.
Is Basket Feeding Useless?
At this point you may be thinking, how does NLP distinguish itself from game-based coaching? And how does the open and closed skill spectrum contribute to the overall picture? These are good questions. For starters, proponents of NLP are NOT totally against using a model-approach to learning, unlike some game-based coaches who often see little value in traditional coaching methods. NLP suggests that during the coordination stage of learning - considered a very early stage “internally focused instructions may be helpful in establishing a basic coordination pattern” (Peh et al 2011). This means that “closing” the skill for beginners and basket feeding may be completely appropriate when learning a new technique!
The difference perhaps between traditional coaching and NLP is the emphasis on implicit instructions rather than explicit ones. For example, when attempting to improve a beginner’s forehand groundstroke depth, you may notice a lack of power because of a backswing that's just too short. Instead of instructing the beginner to bring the shoulder further back, bending the elbow a certain angle etc. (called explicit instructions), you may ask them to draw a circle with the tip of their racquet (implicit instructions). You can get further specific by asking them to make a large, medium or small circle depending on where they receive the ball and the tactical intention. What separates NLP from the model approach then, even with beginners, is that once an ‘approximation’ of the desired movement pattern is established, the focus of attention quickly shifts to one that is more external - or in the case for tennis training, more tactical and open (and less basket feeding).
Tennis Players Need Both Fixed and Variable Movements
Tennis training likely needs a combination of both coaching approaches. Let's illustrate why it's important for players to train both fixed (closed) and variable (open) skills. Every elite player possesses certain characteristics that distinguish them from their lesser counterparts. But when looking at technical training from a NLP perspective, at first glance it may seem like each individual should have completely different ways of solving movement (or mechanical) puzzles. But this isn’t entirely correct. According to Bosch, elite performers have what are called attractors and fluctuators in all movements. Attractors are fixed, stable and economical movement patterns while fluctuators are unstable, high-energy movements. Beginners tend to possess more fluctuators compared to elite performers. As learners become more skilled and attain more ‘stability’, they develop more attractors. However, movements cannot consist of only attractors, otherwise they would be performed very rigidly and any influence from the environment would throw the entire system off. For example, a player may be capable of hitting really nice groundstrokes off of balls that bounce just past the service line. But what about if the ball bounces just inside the baseline, making the shot more difficult? Without fluctuations in movement, the player will stand in the same position, won’t adjust their swing and will likely produce an error.
But this is why fluctuators continue to exist - even when performers enter the later stages of learning - to adapt the movements in the slightest way possible, in the most economical way possible in order to meet the situational demands. From the above example, if our player possesses fluctuators, then they’ll either shorten their swing or they’ll buy time by moving back to adapt to the deeper ball. And say what you want about Maria Sharapova, but she's a prime example of someone who either adjusts their positioning on the court, or gets lower & shortens the backswing when receiving deep, tough balls (just look at the 2 videos above). Another example would be the serve. It’s well accepted through research that the shoulder abducts to about 90 degrees at impact (Ellenbecker et al 1995). This is true for both beginners and advanced players, meaning that this aspect of the service motion is an attractor - it’s a stable movement pattern. Elite players, however, have more bending of the trunk during the serve, depending on the intended direction, spin, speed and so on. The varied movement of the trunk in elite players would therefore be classified as a fluctuator.
The image below is a theoretical look at the various stages of learning. There are 5 stages - stage 1 being the top of the image, classifying beginners and stage 5 being the bottom of the image, classifying elite performers. The valleys (indicated by arrows) represent attractors - the deeper the attractor, the more fixed a certain pattern of movement. Based on this theoretical model, as skills are mastered, we gain more fixed movement patterns AND these movements become less variable. But fluctuators never disappear, even at later stages.
It’s important to state that, when it comes to movement, there are many degrees of freedom at each joint. In other words, each joint will have differences in the number of ways in which it can move. The elbow joint for example, has been reported to have 7 degrees of freedom. So when performing a complex skill such as a forehand (which involves most major joints of the body including the ankle, knee, hip, wrist, elbow, shoulder, spine etc) there are thousands if not more permutations that could result in a successful movement outcome. Furthermore, each joint may have several muscles that act upon it, with each muscle being more involved or less involved depending on how the movement is performed...and as humans, the way in which our muscles act to produce movement is completely out of our control (i.e. we cannot specifically call upon the biceps brachii to activate at a certain percentage compared to the brachialis - both involved in elbow flexion - EVER, let alone when under time and space constraints, such as retrieving a hard hit forehand into the crosscourt corner).
That said, it’s very difficult for players to receive overly prescriptive instructions. Movement is just too complex. Although linear approaches are often very successful, in very specific situations - i.e. if I change A it’ll result in B - they do not take into account this complexity. Nonlinear approaches on the other hand, are far more elaborate and reflect the chaotic nature of movement. If I change A, it could result in B or C or any combination of many outcomes. Even a simple change in practice schedule from 3 practices a week to 4 could have large impacts. At first glance, you may think a player will absolutely improve with an increased time on court...but what if this player can't handle the added training load? And instead their performance drops. NLP approaches encourage us to look at learning and training from this perspective, one that’s constantly changing and forcing performers to continuously adapt. This, in my opinion, is how elite players are formed.