What kind of an entity is basketball? And based on its essence, how should it be practiced and coached? Based on the analysis, should changes be made to the prevailing procedures? If so, what type of changes?
These are the questions we are dealing with here. In other words, we are trying to build a solid basis for discussing and improving basketball coaching. To do that, we are defining the main concepts and their relationships because that is an elementary prerequisite for a meaningful discussion.
I am writing and posting this piece by piece, chapter by chapter. Also, I am revising the blog whenever I find it necessary. So do not wonder if you revisit the piece and find it different from what you thought it was.
I last edited this blog on Wednesday July 17 when I put in Chapter 30: Four Classes of Heuristics.
Chapter 1: How Basketball Is a Complex System
What type of an entity is basketball? Answering this question is important in order to define basketball. That definition will help us to define coaching and practicing basketball. Eventually, those definitions allow us to meaningfully discuss how the efficiency of practicing and coaching the sport may be enhanced.
In the contemporary research, invasion team sports such as basketball are often viewed as complex systems. Let us make that our starting point and see if that point of view is valid.
- A set of parts are “interconnected so that changes in some [parts] or their relations produce changes in other parts of the system”. That is bottom-up causation: changes at the lower scale of a system cause changes at the higher scale.
- The whole “exhibits properties and behaviors that are different from those of the parts.”
- They have autonomous parts whose interaction produces emergence. Thus based on the input, one can not exactly know what the output of a complex system will be.
- They themselves are parts of hierarchies of complex systems. This means that in addition to bottom-up causation, there is top-down causation: changes at the higher scales of the system hierarchy cause changes at the lower scales.
Given these definitions, a basketball team is a complex system. Most of the researchers also treat invasion team sport games as complex systems. This has been questioned. Lebed argues that rather than a system, a game is “a conflict of — two complex dynamical systems”.
However, there is no reason why a conflict like that could not be a complex system, too. Given the definitions above, a basketball game does meet the two conditions of a system and the additional two conditions of a complex system.
- The two teams involved in a game are “interconnected so that changes in some [parts] or their relations produce changes” in the game. The defense played by Team A affects the offense by team B.
- A game “exhibits properties and behaviors that are different from those” of the teams involved. Team A alone can not play a game but Team B is needed, too.
- The teams are autonomous parts whose interaction produces emergence. Thus based on the teams, one can not exactly know what the outcome of the game will be like. Based on information regarding Teams A and B, it is impossible to know in advance how a game between the two teams will unfold. This is the case regardless of the quantity and the quality of the information.
- Games are parts of hierarchies of complex systems. Consequently they e.g. adapt to changes in the environment. A game between Team A and Team B is a scale in a hierarchy of complex systems. A scale one step down is the scale of Team A’s individual players. A scale one step up is the scale of the league where the game is played.
Chapter 2: Defining Basketball
Based on the above, we look to next define basketball. Then, based on the definition, we move forward towards the practical applications regarding practicing and coaching.
Generally, “a formal definition corresponds to the formula an X is a Y + distinguishing characteristic, where Y is a class word or superordinate term”. Thus, we may start the definition by stating that basketball is an invasion team sport.
What then distinguishes basketball from other invasion team sports? Put in the terms of complex systems, we can say that the main distinguishing factor is that in basketball, the interaction between the two teams is confined by the basketball rules.
So, our definition comes to be:
- “Basketball is an invasion team sport where the interaction between the two teams is confined by the basketball rules.”
There are other definitions, perhaps better ones. However, for this definition to serve its purpose it only needs to be acceptable to everyone. In other words, we need to agree that what is says about basketball is true and essential. Given that, we can use the definition as a part of the basis for the current discussion. The same goes for all definitions presented here.
Chapter 3: Definitions of Practicing and Coaching
In our definition of basketball, rules of the sport have an important role: “Basketball is an invasion team sport where the interaction between the two teams is confined by the basketball rules.”
In fact, rules are the only thing mentioned to distinguish basketball from other invasion team sports. Consequently, any invasion team sport can be defined using the same formula: “X is an invasion team sport where the interaction between the two teams is confined by the X rules.”
In the FIBA rules, the most elementary confinements on the team’s interaction are stated in the articles 1.1 and 1.3:
- “The aim of each team is to score in the opponents’ basket and to prevent the other team from scoring.”
- “The team that has scored the greater number of points at the end of playing time shall be the winner.”
In other words, the goal of the teams is to manipulate the interaction so that they beat the other team.
To enhance their chances, teams practice. Practice has been defined as “repeated exercise in or performance of an activity or skill so as to acquire or maintain proficiency in it”. Given this and the viewpoint of complex systems, we can formulate this definition:
- “Basketball practice is activity that aims to optimise the proficiency of a team at manipulating the interaction in basketball games in their favour.”
In the context of sports, coaching may refer to training, instructing, and teaching an individual or a team. So, we may formulate this definition:
- “Basketball coaching is activity where a team is trained, instructed, and taught in order to optimise its proficiency (at manipulating the interaction in basketball games in their favour).”
Then a basketball coach is someone who does basketball coaching.
Remember, we are not yet trying to determine what basketball practice and coaching should be like. Rather, we are defining the terms in the current context. Then based on these definitions, we will later try to draw conclusions about the “should be” part.
Chapter 4: On Underlying Assumptions in Coaching
In the next two chapters we look into two different coaching approaches and some of their underlying assumptions.
Here we use the term “coaching approach” to refer to a set of fundamental principles that underlie and guide the coaching process. This set is to be comprehensive, and the principles are to be related and compatible. They may overlap but they must not contradict each other.
In other words, the coaching approach is not merely a set of coaching methods and exercises. Rather it underlies and guides the selection of those methods and exercises.
Some assumptions of this kind necessarily underlie all actions in coaching. There is always a rationale behind them. Sometimes the assumptions and the rationale get articulated, but most often they are applied implicitly, maybe even unknowingly.
That is because “coaches often see little value in a philosophy as they attempt to cope with more tangible aspects of coaching practice, such as session content and organisation”.
Coaches may consider their philosophical assumptions so “common-sense”, “taken-for-granted” and “normal” that there is no need to articulate, let alone examine them.
However, when trying to improve coaching, we are better off if we articulate the underlying assumptions. That makes it possible to assess the assumptions and the methods critically and to correct and improve them.
Say a basketball coach runs exclusively constant and blocked shooting drills. Obviously then he believes that constant and blocked practice is the most efficient method to learn shooting. That belief is there no matter whether the coach himself articulates it or not.
If he does articulate his belief, it becomes possible to have an intelligent and productive discussion on e.g. the pros and cons of constant and blocked practice versus those of varied and random. Without proper articulation, this type of a discussion is not possible.
Chapter 5: On the Positivistic Coaching Approach
Historically, the dominant approach in coaching and coaching research has been the positivistic approach. Using that particular adjective in the term is a matter of choice. The same basic approach towards coaching is reflected by these overlapping adjectives:
As discussed above, often coaches do not articulate or even realise the assumptions that underlie their coaching. Hence, not too many coaches will explicitly acknowledge that their coaching approach is positivistic.
Also, even if a coach’s general approach is to be considered positivistic, he may not subscribe to all positivistic claims. And the terms mentioned here have many definitions – i.e. not all forms of positivism are the same.
However, generally speaking it is justified to call the dominant approach in the coaching field positivistic. The argument here is that this is how most coaches in praxis coach most of the time. This claim grains credence from the following list. It connects underlying assumptions of a positivistic nature, general coaching principles, and basketball-related examples.
Table 1. The quote is from here.
The positivistic coaching approach has produced a lot of champions. So in a way, the problem is not that it doesn’t work. It does work – up to a point. But it doesn’t work optimally, or all the way through.
For example, the effectiveness of a team does correlate with the effectiveness of its players. In pick-up basketball, if you put the five best players on one team and the five worst players on the other one, the five best players’ team will win the game. And if you replace your center with a more effective one, you very possibly improve the team effectiveness.
So why does the positivistic coaching approach will eventually run short in basketball? Considering that basketball is a complex system, there are at least three explanations.
1) The Specificity of Learning Principle
- The specificity of learning is a most elementary, well-proven principle of sports training. What it means is that “improvement is observed only in the trained task, with little to no transfer of learning being observed even for very similar untrained tasks”.
- Previously we have defined basketball as “an invasion team sport where the interaction between the two teams is confined by the basketball rules”. Thus, according to the specificity of learning principle, the interaction between the two teams should the primary trained task in the sport.
- The positivistic coaching approach fails to comply with this conclusion. In the list above there are several points that imply that the emphasis of training may be put on other tasks than the primary one. The points in question are the ones 4-to-8.
2) Emergence of Interaction
- In complex systems, the output emerges from the interaction between the parts of the system. Because of this emergence, even if we know the the input, we can not know what the output will be. Hence, we can’t know what the input should be in order for us to get the optimal output.
- The same in basketball terms: We don’t know exactly what will happen in a game – or how the interaction with the other team will function. Because of this emergence, we don’t know how the team and its individual players should practice or play in order for the team to get the optimal result.
- This defies points 1-to-4 and 7-to-10 of the list above. That is because enhancing pre-determined tactics and techniques is not the optimal solution when the primary goal is not going to be executing those tactics and techniques. Rather, in a basketball game the primary goal is to adapt to the emerging interaction and to manipulate it the best that you can in any way that you can.
3) Top-down Causation
- The positivistic coaching approach relies on bottom-up causation. The idea is that through training, positive changes are caused in parts of a system. Then in turn, those changes will cause positive changes in the functioning of the system.
- Or in basketball terms, the players practice and develop. This development will help the team to play more effectively.
- But what is ignored in the positivistic approach is that in complex systems also top-down causation takes place. In other words, changes in functioning of the system will also cause changes in the functioning of its parts.
- Or in basketball terms:l It’s not just players affecting how the team plays, but also the team affecting how the players play. And as the teams affect how the game unfolds, the game affects how the teams play.
- This feedback loop defies reductionism, or points 3-to-6 in the list above. Because of the top-down causation, the functioning of a system can not be explained merely through the functioning of its own part.
Chapter 6: On the Holistic Coaching Approach.
The positivistic coaching approach has been challenged by the holistic coaching approach. The latter is based on two principles.
- The emphasis is on coaching individual athletes.
- Coaching is viewed “as a complex social process, which involves a myriad of interacting variables”.
Holistic coaching is related to athlete-centered coaching and humanistic coaching. Even if the terms are not synonyms, they do share the two principles mentioned. Because of that, the criticism presented here regarding the holistic coaching approach applies to its two relatives as well.
The holistic coaching approach can be viewed as a counterattack against the positivistic coaching approach. The holistic approach is based on complex systems thinking.
Yet, when it is basketball coaching that we view from the complex systems viewpoint, there is a fatal problem with holistic coaching: it puts the emphasis on a wrong scale of the complex system.
This is because of two things:
- Above, we defined basketball practice as “activity that aims to optimise the proficiency of a team at manipulating the interaction in basketball games in their favour.”
- The specificity of learning principle says that “improvement is observed only in the trained task, with little to no transfer of learning being observed even for very similar untrained tasks”.
So, to optimally improve the proficiency of a basketball team, the emphasis of practice must be on the interaction with the other team. This contradicts the fact that the holistic approach puts the emphasis on individual athletes.
Chapter 7: Kuhn’s Anomalies
The critique above on the positivistic and holistic coaching approach is coherent and logical. However, it is also mostly a priori or theoretical knowledge whereas practicing and coaching are very practical deeds.
Because of that, the critique above does not suffice to dismiss the two traditional coaching approaches when it comes to coaching basketball. More and different kind of evidence is needed – a posteriori or empirical evidence that is.
The method we will use to look for the evidence is this:
- Based on the positivistic and/or holistic coaching approach, we make predictions regarding research results.
- We see if there are research results that contradict the predictions.
- We assume that if the approaches lead to false predictions, there may be something wrong with the assumptions.
- We consider if there are so many anomalies that outlining a novel coaching approach is in place.
The contradictions between the research results and the predictions derived from the traditional coaching approaches are referred to as anomalies. The term comes from Thomas Kuhn: “Discovery commences with the awareness of anomaly, i.e., with the recognition that nature has somehow violated the paradigm-induced expectations that govern normal science.”
There will be no – and there can be no – definite point where we can say that we have accumulated enough anomalies and that we can now start outlining a novel coaching approach. That is because by definition complex system will inevitably produce emergent, unexpected output – anomalies, that is.
Instead, what we should look for is a point where there are so many critical anomalies accumulated that it becomes rational to assume that another type of an approach might help to coach basketball more effectively.
The methodological idea applied here was inspired by Richard H. Thaler. In the 1990’s he promoted then-novel behavioral economics “by writing about anomalies in people’s behavior that could not be explained by standard economic theory”.
Chapter 8: Anomalies Unveiled.
The positivistic coaching approach and the holistic one are based on very different assumptions. As mentioned above, the holistic coaching approach can be viewed as “a counterattack against the positivistic coaching approach”.
However, the approaches do share an assumption: the effectiveness of a team depends linearly on how effective its individual player are. This assumption is reached in different ways.
A part of the positivistic approach is the reductionistic belief that “the functioning of a system can be explained and predicted through the functioning of its parts”. This predicts that the more effective the individual players are, the more effective the team will be.
In the holistic coaching approach, the one primary principle is to improve the individual athletes. Now, we have defined basketball practice as “activity that aims to optimise the proficiency of a team”. Improving individual players may be an optimal way to do this only if there a linear correlation between the cumulative effectiveness of the individual players and the effectiveness of the team.
So, if we find research results where the correlation described is not linear, we also find anomalies relative both to the positivistic and the holistic coaching approach. That is why we here concentrate on that type of anomalies. It keeps this analysis compact.
However, it can be assumed that also different types of anomalies could be found regarding the positivistic approach. That is because the approach has multiple underlying assumptions (see the list above).
Importantly, an anomaly here is “a result inconsistent with” the expectations. Or, a result is anomalous if “it is difficult to ‘rationalise,’ or if implausible assumptions are necessary to explain it”.
In other words, in order to be an anomaly the results need not show that there is no correlation between the cumulative effectiveness of the players and the collective effectiveness of the team. Rather, an anomaly is a result where this correlation is inconsistent, i.e. non-linear.
The research referred to here is all NBA-related.
Anomaly 1: The too-much-talent effect
Swaab et al (2014): The too-much-talent effect. Team interdependence determines when more talent is too much or not enough.
Results: “First, the actual marginal benefit of more talent decreased at a much faster rate than people believed it would. Second — the relationship between talent and performance eventually turned negative.”
The anomaly: Adding to the cumulative effectiveness of the players may harm the effectiveness of the team.
Anomaly 2: The U-shaped effect of team familiarity
Sieweke and Zhao (2015): The impact of team familiarity and team leader experience on team coordination errors. A panel analysis of professional basketball teams.
Results: “On the one hand, [team] familiarity is important for developing the mutual understanding that helps team members anticipate and adjust to each other’ s actions, thereby improving coordination. On the other hand, too much familiarity might negatively affect coordination —.”
The anomaly: The effectiveness of the team depends not just on the effectiveness of the individual players but also on how familiar they are with each other. Even if the current players improve their effectiveness, the effectiveness of the team may decrease because of the growing team familiarity.
Anomaly 3: How the players complement each other
Ayer (2011): Big 2’s and big 3’s. Analyzing how a team’s best players complement each other.
Results: “— high-scoring point guards don’t mesh well with high-scoring 2 guards. Talented, high-scoring centers fit well with more limited, defense-oriented power forwards who rebound very well, which also aligns with conventional wisdom. Unexpected results include the degree of fit when teams have two high-scoring 2 guards.”
Anomaly: “Constructing a team that can reach its full potential requires more than just acquiring talented players; these players have to fit well together.”
Anomaly 4: Synergies with the teammates and the opponents
Maymin et al (2013): NBA chemistry. Positive and negative synergies in basketball.
Results: “Because skills have different synergies with other skills, a player’s value depends on the other nine players on the court.”
Anomaly: The effectiveness of individual players will depend on the teammates and the opponents. Thus the players’ effectiveness is not cumulative in nature but rather situational and relative.
Chapter 9: Towards a Complex Coaching Approach
Since the positivistic and holistic coaching approaches do not seem optimal when it comes to coaching basketball, we need a novel coaching approach. What should it be like?
In philosophy of science, there’s a long line of discussion about how scientific theories and hypotheses are generated. A much-debated issue is the DJ distinction or the distinction between the context of discovery and the context of justification. Without getting into that debate, it seems rational to base the novel coaching approach on:
- The definitions of basketball, basketball practice, and basketball coaching.
- The critique and anomalies regarding the positivistic and holistic coaching approach.
We defined basketball “as an invasion team sport where the interaction between the two teams is confined by the basketball rules.” And coaching as activity that aims to optimise a team’s proficiency at manipulating that .
According to the specificity of learning principle learning almost exclusively happens in the trained task. So, in basketball practice the primary trained task should be the interaction between the two teams. This is where the positivistic and holistic approaches go wrong, as shown in chapters 5 and 6.
This leads us to the conclusion that in the novel coaching approach, the primary trained task should be the interaction between the two teams. All training, instructing, and teaching will be assessed based on how efficiently it advances that task. We call this approach the complex coaching approach.
Chapter 10: Diminishing Returns of Scrimmaging
Given the above chapter and the previous definition of basketball coaching, the primary principle of the complex coaching approach is to attempt to optimise the team’s proficiency at manipulating the game interaction.
Given this and the specificity of learning principle we may conclude that in order to improve the proficiency at game interaction, the team’s training task should be game interaction. This means 5v5 scrimmage using official basketball rules. In other words, scrimmaging is as basketball-specific as practice action.
Naming scrimmaging as the default practice task is not exceptional. Rather, it complies with recent calls for more game-like practices. And especially at the pro level, some coaches run practices that almost exclusively consist of 5v5 action.
Even if scrimmaging is the default practice action, it does not imply that teams should nothing but scrimmage. That is because of the Power Law, another well-established principle of sports training.
The Power Law says that the more you train something the more better you become at it, yet eventually the rate of learning will slow down. So, the more a team scrimmages, the better they learn to interact with each other and with the opponents. Yet as scrimmage hours accumulate, less and less learning happens per training hour.
So eventually it becomes more efficient to do other practice actions than scrimmaging. Because of the diminishing returns of scrimmaging, some other practice actions become more efficient. The variability they bring to the practice compensates for their relative lack of basketball-specificity.
Every step away from the maximum basketball-specificity of scrimmaging must be justified, yet eventually such steps must be taken.
Chapter 11: Defining Basketball-specificity
We now know that scrimmaging is as basketball-specific as a practice action can be. Next, we must define basketball-specificity. That is because otherwise it will not be possible to intelligently discuss and justify practice actions other than scrimmaging.
Above we defined basketball practice as “activity that aims to optimise the proficiency of a team at manipulating the interaction in basketball games in their favour”. Thus, a practice action can be called basketball-specific if it is justified to expect that it helps to advance this goal – i.e. to succeed in basketball games.
So, we define basketball-specificity as “a quality of a practice action that makes it justified to expect that executing the action helps the team to optimise their proficiency at manipulating the interaction in basketball games in their favour”.
This definition implies that in order to be basketball-specific, a practice action does not need to include any equipment or even movement patterns used in basketball games, not even the ball. Any practice action – say meditating – can be called basketball-specific for as long as it can be justifiably expected to help the team to optimise their proficiency at manipulating the interaction in basketball games.
This is important to point out because often the term “basketball-specific” seems to be used in more narrow a meaning. It seems to be assumed that in basketball-specific practice actions, some of the same equipment or at least movement patters are to be used as in games. Often this assumption becomes obvious by implication only since explicit definitions of “basketball-specificity” are rare.
Chapter 12: On the Tactical Periodisation Approach
The above remarks imply that there is no rationale for dismissing any practice methods right off. No matter far removed from actually playing basketball a method seems to be, it is basketball-specific if it helps to achieve the goals of practice. That is, if it helps to “optimise the proficiency of a team at manipulating the interaction in basketball games in their favour”.
How then can we assess the degree of basketball-specificity of a certain practice action? It has become clear that this is not to be done intuitively, or based on looks alone.
Rather, we should try to spot the mechanism behind the basketball-specificity. Meaning, we should look to find out what is required of a practice action in order for it to be basketball-specific. Or, how the effect of a practice action is mediated all the way to the game performance of a team.
To accomplish this, in this chapter and the following ones we look into concepts derived from the tactical periodisation approach. It was developed for soccer, but it can also be applied to other invasion team sports, such as basketball.
According to the tactical periodisation approach, the tactical dimension is the primary one. Meaning, it is tactical actions that determine the outcome of a basketball game.
However, the tactical dimension is always interwhined with the technical and the physiological dimension. That is, technical skills are needed to carry out tactical actions. Analogously, physiological capabilities are needed to execute technical skills.
In the tactical periodisation approach, it is claimed that – still analogously – the physiological dimension “is directed by volitional and emotional states” i.e. the psychological dimension. However, the psychological dimension is left out of this analysis for two reasons.
- The effect of the psychological dimension is not restricted to the physiological dimension. Rather, psychological factors directly affect the technical and the tactical dimension, too. In other words, psychological factors underlie the whole hierarchy of the tactical, technical, and physiological dimensions. Thus the tactical periodisation approach makes a questionable claim about the place and effect of the psychological dimension. That claim is not to be repeated here.
- The psychological dimension is fundamentally different from the the tactical, technical, and physiological dimensions. Unlike them, it is not a part of their concretely observable hierarchy. Thus the hierarchy mentioned can be discussed intelligently even if the psychological dimension is left out of the discussion.
This in no way implies that the psychological aspect is not important when it comes to basketball. It’s just that at this of point, the psychological aspect is defined out of the discussion.
Chapter 13: Causality in Two Different Approaches
In Table 2 (below) the column in the middle shows some complex system scales that are relevant to basketball. The column on the right hand side shows the equivalent dimensions of the tactical periodisation model.
The scale and the dimension in each row are parallel in the sense that they regard the same phenomena. For example, a game is the complex system scale where the two teams interact. Equivalently, the tactical dimension of the tactical periodisation model is the dimension where a teams tries to optimise its ability to manipulate “the interaction in basketball games in their favour”.
In spite of the parallelism, there is a fundamental difference between the two approaches. The complex system approach can be utilised to explain how the complex system of a basketball game functions. Contrarily, the tactical periodisation approach describes “the logical structure” of basketball. This structure is such that in order to be properly executed, the higher level dimension requires that the lower level dimension has certain properties
Because of this fundamental difference causality works differently in the two approaches. As discussed above, in complex systems the causality works two ways: bottom—up and top—down. Contrarily, in the tactical periodisation approach there is only bottom—up causality.
On the complex systems’ side, each players’ play affects the game. For example, the way a player passes the ball affects the way his team plays on offense (Row 2 affects Row 1). But also, the team’s play affects each player’s play. For example, the way the team spaces the floor affects the way that a player passes (Row 1 affects Row 2). So, there is causality both bottom—up and top—down.
On the tactical periodisation’s side, the technical dimension affects the tactical dimension. For example, the tactical choices are enabled and limited by a player’s passing skills (Row 2 affects Row 1). However, the tactical choices do not determine which technical skills a player has at the moment (Row 1 doesn’t affect Row 2). In other words, there is only bottom—up causality.
Chapter 14: Two Principles Regarding Practice
What does the above analysis of the tactical periodisation imply regarding practicing – in others words, improving the three dimensions mentioned?
The following two principles drawn from the analysis may seem self-evident and theoretical, in the bad sense of the word. They may seem far removed from the very practical question stated at the beginning of this blog: How should basketball be practiced and coached?
However, I believe that the three principles will be prove to be elementary to the complex basketball coaching approach. Remember, “there is nothing more practical than a good theory”.
Principle 1: Tactics rule. The tactical dimension is the only dimension where the improvement has intrinsic value to enhancing the game performance. Contrarily, improving the technical and the physical dimension is of instrumental value only. That is because the tactical dimension is by definition the only one that is involved in the actual interaction with the other team. Consequently, improving the technical or the physical dimension only enhances the game performance if it helps to improve the tactical dimension.
Principle 2: Tactics comply. The tactical dimension can be improved only as far as the quality of the technical dimension allows. And the technical dimension can be improved only as far as the quality of the physical dimension allows. Thus, improving the tactical dimension is limited by the quality of both the technical and physical dimension.
Chapter 15: Practical Implications of the Tactics Rule Principle
Let’s now look at what the Tactics Rule principle implies regarding the coaching praxis.
Generally speaking, tactics can be defined as the “skill of employing available means to accomplish an end. Here it should be noted that in the context of the tactical periodisation approach, the words “tactics” and “tactical” are very broad concepts. Thus the tactical dimension includes a variety of things:
- Pre-planning team actions in order to manipulate the interaction in games in the team’s favour.
- Practicing those team actions.
- Executing those team actions in games.
- Adapting those team actions to the emerging interaction with the other team.
So, the quality of the tactical dimension directly and inevitably determines the success of any basketball team. This implies at least these two most important and most practical things.
1) Improving the tactical dimension is the most direct way to enhance a team’s performance.
Say your full-court press does not work because the double team arrives too late. The quickest-to-implement method to enhance the press is through improving the tactical dimension. This can be done for example by practicing the timing of the double team or by replacing the designated double-teamer with a quicker one.
Alternative methods to get the double team to arrive earlier are:
- Improving the technical dimension. For example, improving the current double teamer’s defensive footwork technique in order to improve the quickness of his double-teaming.
- Improving the physical dimension. For example, improving the current double teamer’s lower-limb power in order to improve the quickness of his defensive footwork in order to improve the quickness of his double-teaming.
So, the latter two methods’ effectiveness involves more open questions about whether the improvement actually transfers to the game performance. This brings about the second practical implication of the Tactics Rule principle:
2) Improving the technical and the physical dimension only enhances the game performance if the improvement transfers to the tactical dimension.
Ideally, the improvement in the physical dimension transfers to the technical dimension, and the improvement in the technical dimension transfers to the tactical dimension – in other words, to the game performance.
For example, a player may improve his unilateral leg strength, an aspect of the technical dimension. That may help him to improve his change-of-direction speed and eventually his defensive footwork, or an aspect of the technical dimension. This improvement may transfer to the tactical dimension, provided that the player manages to use his improved footwork to improve the execution of double-teaming in game situations.
This chain of transference may go wrong in two ways.
- Improvement in the physical dimension doesn’t transfer to the technical dimension. For example, through improving his flexibility – or an aspect of the physical dimension – the player may look to improve his change-of-direction speed and thus his defensive footwork. However, it may turn out that this transference doesn’t happen because flexibility is not a limiting factor when it comes to the change-of-direction speed.
- Improvement in the technical dimension doesn’t transfer to the tactical dimension. For example, the player may not get to utilise the improved technical aspect. Say, even with his defensive footwork improved, it may not be quick enough to earn him the status of the designated double-teamer.
Chapter 16: Practical Implications of the Tactics Comply Principle
Above it was said that “the tactical dimension can be improved only as far as the quality of the technical dimension allows”. And that “the technical dimension can be improved only as far as the quality of the physical dimension allows”.
This is explained by the Power Law, presented in Chapter 10: “The more you train something the more better you become at it, yet eventually the rate of learning will slow down”. This phenomena in known as the diminishing returns.
For example, up to a point the press defense can be enhanced the most efficiently by working on tactical aspects, as was explained above. However, after a while this method pre-empts the improvement potential. This potential is – or is not – allowed by the technical dimension, say by the defenders’ defensive footwork technique.
That is why the tactical dimension doesn’t just rule the technical dimension but it also complies with it.
Parallelly, the same is true about the technical dimension and the physical dimension. In the hierarchy of the tactical periodisation approach, improving the technical dimension is more direct a way to enhance the game performance than improving the physical dimension is. In that sense, the technical dimension rules the physical dimension.
But the technical dimension also complies with the physical dimension. For example, once a player has worked on his defensive footwork technique for some time, the Power Law kicks in: the improvement potential allowed by the technical dimension has been pre-empted. Once a certain hard-to-define tipping point has been reached, it becomes more efficient to improve the technical dimension by improving the physical dimension than by improving the technical dimension directly.
One reservation: contrary to what I just wrote, the improvement potential of any dimension is never literally pre-empted. For example, there is always some room for improvement regarding the tactical execution of a certain press defense. But at a tipping point this room gets so small that it becomes more efficient to improve the technical dimension – say the footwork technique – in order to to enhance the game performance.
There is an elementary and necessary tool that coaches continuously use to create new improvement potential to the tactical dimension: they put in new tactics. The most common way of doing this in putting in new set plays. Or maybe in addition to the trapping man-to-man press, the coach may put in a 1-2-1-1 zone press.
While this is standard, the risk of over-emphasising the tactical dimension is ever-present. That is because coaches always have the pressure to the very next game. And as shown above, improving the tactical dimension is the most direct way to enhance the game performance.
If this risk is realised, there may be so little practice on the technical and physical dimensions, that they regress. That is because it takes much practice time to even maintain one’s technical skills and physical capabilities – let alone to improve them.
Given the above, in the short run it may be rational to practice just enough to maintain the technical and the physical dimensions and to concentrate on improving the technical dimension.
Yet in the long run this approach too may end up hurting the team. That is because eventually the improvement potential of the tactical dimension gets pre-empted, now matter how cleverly the coach utilises the improvement potential allowed by the technical and physical dimensions. The risk is high especially if the players are young and/or if the roster remains stable from season to season.
Chapter 17: A Misunderstanding of the Tactical Periodisation Approach
As shown above and as mentioned in Chapter 13, the tactical periodisation approach describes “the logical structure” of basketball. For a coach, it is important to analyse this structure – i.e. the combination of the tactical, technical, and physiological dimension. That is because this analysis helps to find out which dimension limits the performance of the team, and thus which dimension should first be improved through practice.
However, analysing what should be improved does not show how it should be improved. That is because in complex systems there is not only bottom-up causation but also top-down causation.
For example, say that the physiological dimension is the limiting one. The players’ conditioning is so poor that it severely hurts the technical and consequently tactical dimension of their game performance. By looking at the tactical periodisation structure alone, one might conclude that the players should now do physiologically-oriented training, i.e. conditioning.
However, it need not be so because basketball is a complex system. Because of the top-down causation, basketball players’ physiological dimension may be improved through tactically-oriented and technically-oriented training, too.
This is no news to proponents of the tactical periodisation approach. As a matter of fact, the realisation that there is top-down causation is one of the cornerstones of the approach.
This and the fact “the tactical dimension is the primary one”, seem to have led the tactical periodisation proponents to conclude this:
“— training should never separate the physical, tactical, technical and mental elements of preparation —. In particular, physical preparation should not be isolated and trained independently; it should integrate with the mental, technical and tactical training.”
However, this is a misunderstanding. The conclusion simply does not follow from its premises. Even though the tactical dimension is the primary one, there is no reason to assume training should always be tactically-oriented.
That is because – as mentioned above – tactics do not only rule but they also comply: “Improving the tactical dimension is limited by the quality of both the technical and physical dimension.” Thus, sometimes the most efficient way to improve the tactical dimension may be to work exclusively on the physiological dimension.
In the context of rugby, this is discussed by Tee al (2012): “Resistance training methods are indispensible in rugby union —, and it follows that training can therefore never be wholly tactical in nature.”
Chapter 18: Four Factors
At the beginning of this blog, we set out to figure out how basketball should be practiced. Based on the discussion above, we can now start putting together tools for systematically answering this question.
We defined basketball practice as “activity that aims to optimise the proficiency of a team at manipulating the interaction in basketball games in their favour”. When determining what the team should work on in practice, we must first determine what the possible answers are.
To do that, we use a method developed to analyse the overall efficiency of a team’s game performance. The rationale is that the factors included in this analysis are also the possible practice targets. In other words, the factors indicate where there is room for improvement.
The best known method of this kind is Four Factors. It was developed by Dean Oliver in his book Basketball on Paper (2004). Also well-known is the article on the subject by Kubatko, Pelton, Rosenbaum and Oliver himself (2007).
The four factors of the Four Factors are:
- Turnover percentage.
- Offensive rebounding percentage.
- Effective field goal percentage.
- Free throw rate.
The team’s offensive statistics are used to assess their offensive performance. And parallelly, the opponents’ statistics are used to assess the team’s defensive performance.
Oliver was confident that his method could be used to assess in which aspects the team did or didn’t do well. “There really is nothing else in the game”, he wrote. However, we must next take a critical look at Four Factors.
The method is based on analysing possessions. According to the definition by Kubatko et al (2007), a possession “starts when one team gains control (or possession) of the basketball and ends when that team gives up control of the basketball”. According to them, “an offensive rebound does not start a new possession”.
Here, however, we define a possession so that an offensive rebound does start a new possession. So, a possession “starts when one team gains control of the ball and ends when that team gives up control of the basketball with a turnover, field goal shot, or free throw.” This latter definition of a possession is the same that Kubatko et al use as the definition of play.
These two alternative definitions of a possessions may lead to confusions. For example, when discussing points per possession, it should be clear to everyone involved which definition has been used.
On one hand, defining a possession is a matter of choice. On the other hand, I do believe that having an offensive rebound start a new possession makes the analysis easier to do and to follow. That is for example because there is not going to be more than one field goal shot or free-throw set in one possession.
Chapter 19: Turnover Percentage
Turnover percentage or TOV% is “the share of possessions that end without the offensive team taking” a field goal shot or a free-throw attempt. The equation is:
- TOV% = Turnovers / Possessions
In praxis, it is often the total number of turnovers that is used to assess how well the teams manages to avoid turnovers. That is not an optimal choice: the number of turnovers depends not just on the quality of the team performances but also on the pace of the game.
The quicker the pace of the game is, the more possessions there will be. And the more possessions there are, the more turnovers there will likely be.
Given the above, TOV% as defined in the Four Factors method is a valid way to measure the amount of turnovers. Just as importantly, it is practical to use: easily calculated and interpreted.
One addition is needed, however. In the context of the Twelve Factors and the FIBA rules, an unsportmanlike foul is considered not just a foul but also a turnover. That is simply because after an unsportsmanlike foul and the last FT, the FT shooting team gets the ball back.
Chapter 20: Offensive Rebound Percentage
The equation of offensive rebound percentage is:
- ORB% = ORB / (ORB + Opponent DRB)
Here ORB=Offensive rebounds and DRB=Defensive rebounds, In other words, ORB% shows the share of available rebounds that the team took at their offensive end.
Often it is the total number of rebounds that is used to measure the success of offensive rebounding. That is not valid because the total number of offensive rebounds correlates not just with the quality of offensive rebounding but also with number of available rebounds, or the number of missed shots.
That number in turn depends on:
- The pace of the game i.e. the number of possessions in the game. “The quicker the pace, the more shots will be taken, the more shots will probably be missed, and the more total rebounds both teams are likely to get.”
- TOV%. The less turnovers there are, the more shots and thus misses there will be.
- The quality of shooting. The more field goal shots and free throws the team miss, the more opportunities for offensive rebounds they are going to have.
Conversely, ORB% is not affected by these factors. Thus it is valid for measuring the quality of offensive rebounding within the context of Four Factors.
Chapter 21: Effective Field Goal Percentage
Iin Four Factors, field goal shooting was “originally operationalised” as FG%, or FGM/FGA. The problem was that this equation treated two-pointers and three-pointer the same, even though 3-pointers produce 1.5 times as many points as 2-pointers.
Thus Kubatko et al replaced FG% with effective FG percentage. The equation is:
- eFG% = (2P + 1.5 x 3P) / FGA
This makes “the relative effect of 2P and 3P on eFG%” “the same as their direct effect on the point total”.
By definition, eFG% does not take into account all field goal shots. That is because in the stats, the missed shots where a shooting foul is called, do not count as FG attempts. This causes a bias: the more shooting fouls there are, the higher the eFG% will likely be.
The explanation is this. When it comes to eFG%, a field goal shot with a shooting foul can only be a make, never a miss. Reductio ad absurdum: if the opponents fouled the shooters on each and every field goal shot, eFG% would necessarily be a perfect hundred.
So, eFG% reflects not just the efficiency of field goal shooting but also the team’s ability to draw shooting fouls. This makes eFG% invalid for assessing the efficiency of field goal shooting.
This is not just a theoretical problem but also a practical one. In the context of Complex Basketball Coaching, we look to use the Four Factors to determine what the team should work on in practice. Using eFG% this is difficult to do. eFG% is affected by the efficiency of shooting and the frequency of shooting fouls, and eFG% does not tell where the room for improvement lies.
A solution might be “to split eFGS% into two variables”.The first one of two is Clean eFG%, It is the eFG% of shots where the shooter is not fouled in the act of shooting. FGA stands for field goal attempts. The equation is:
- Clean eFG% = (Clean 2P + 1.5 x Clean 3P) / Clean FGA
The other variable that the split brings forth is Foul eFGS%, or the eFGS% of shots where the shooter is fouled in the act of shooting. FGS stands for field goal shots. The equation is:
- Foul eFG% = (Foul 2P + 1.5 x Foul 3P) / Foul FGS
This split allows the coach to assess the efficiency of FG shooting without this assessment “being hampered by the bias caused by shooting personal fouls”.
Chapter 22: Free Throw Rate
In Four Factors, free throws are considered through the factor called free throw rate. The equation is:
- FT Rate = FTM / FGA
There are other problems with the equation as well, but regarding complex basketball coaching the main one is this: the rate is affected by both the free throw percentage and the number of free throws. And when it comes to coaching, enhancing those two numbers is two separate issues.
So, for the purposes of complex coaching, the free throw factor needs to be divided into two. The first one is simply FT%.
- FT% = FTM / FTA
The second one is the free throw frequency. It should designed to measure a team’s ability to get to the free throw line. That is best reflected by the number of free throw sets relative to the number of possessions. So the equation is:
- FT Freq = FT Sets / Possessions
Chapter 23: From Four to Twelve Factors
Based on the above, the original Four Factors becomes Six Factors:
- TOV% = Turnovers / Possessions
- ORB% = ORB / (ORB + Opponent DRB)
- Clean eFG% = (Clean 2P + 1.5 x Clean 3P) / Clean FGA
- Foul eFG% = (Foul 2P + 1.5 x Foul 3P) / Foul FGS
- FT% = FTM / FTA
- FT Freq = FT Sets / Possessions
Conversely, when using the opponent’s offensive statistics, these six factors can be used to assess the team’s defensive performance.
- Opp. TOV%
- Opp. ORB%
- Opp. clean eFG%
- Opp. foul eFG%
- Opp. FT%
- Opp. FT Freq
So, in a way Four Factors has now become Twelve Factors – six offensive factors plus six defensive ones. It is a valid and comprehensive method of assessing a team’s game performance – more so than the original Four Factors.
Chapter 24: Utilising the Twelve Factors
In Chapter 18, we started to look for a method for analysing “the overall efficiency of a team’s game performance”. In the context of complex basketball coaching, the rationale was to use this analysis method to spot practice targets, or to find out “where there is room for improvement”.
Remember, according to our definition, practice seeks “to optimise the proficiency of a team at manipulating the interaction in basketball games in their favour”.
Now we have outlined Twelve Factors as the sought-after method. The next question is how we can apply Twelve Factors in praxis to determine what to practice.
The procedure is seemingly simple:
- Calculate the Twelve Factors in your team’s games.
- Assess which factors shall be improved or maintained and which can be ignored.
- Plan the forthcoming practices accordingly.
- Execute the practice plans.
The next chapters are dedicated to this procedure. This procedure is the heart and soul of the complex coaching approach.
Chapter 25: Calculating the Twelve Factors
In utilising the Twelve Factors, the first step is coming up with the relevant statistics.
The process will be different depending on the context of the team. The box score and the play-by-play may be provided by the league, or you may need to keep all the stats yourself – or something in between.
Usually the official stats do not provide all information needed to calculate the Twelve Factors. This info tends to be missing:
- How many free throw sets have been taken.
- Which free throw sets have been due to shooting fouls and which due to team fouls.
- The number of possessions.
So, in order to have valid numbers, you probably need to do some data digging and calculations on your own in any case.
If you only use the box score, you can not have Twelve Factors but you have to settle for Four Factors – or eight factors. As shown above, that is not an ideal situation when it comes to using the method to enhance practical coaching. The following chapters are based on the assumption that you have exact, valid statistics from your games.
Perhaps some coaches – especially in youth basketball – will say that they do not need statistics. That is because supposedly they are able to intuitively assess where their team needs to improve and whether the improvement rate is acceptable.
These coaches may be right or they may be wrong. Without proper statistics it is hard to tell.
Chapter 26: Factors to Improve, Maintain, or Ignore
Once you have calculated the Twelve Factors, you may have a table like this. The data may include any number of the team’s previous games.
Next you divide these factors into three categories:
- Improve: Factors that need to be and can be improved through practice in order for you to optimise your proficiency at manipulating the interaction in games in your favour. The wording comes from the definition of practice first presented in Chapter 3.
- Maintain: Factors that need to be and can at least maintained at the current level through practice. Practice is needed because the performance level relevant here is relative. That is, the offensive performance is affected by the opponents’ defense, and the defensive performance by their offense. The opponents tend to improve their absolute performance during the season. Hence you must improve your absolute performance level in order to maintain the relative level.
- Ignore: Factors that can be ignored for the time being in practices.
This division is based on multiple criteria.
1) The factor’s potential importance
The Twelve Factors are potentially of different importance. That is shown by argumentum ad absurdum.
Try this thought experiment: What happens if the value of a factor is be the highest possible or the lowest possible? Does that alone determine the outcome of a game?
This thought experiment helps to divide the Twelve Factors into four categories. Only the offensive factors are mentioned here, but each equivalent defensive factor belongs to the same category.
a) If the factor has the highest possible value that will definitely determine the outcome of the game: TOV%.
If TOV% is 100%, the team will lose lose the game because it will not score a point.
b) If the factor has the highest or the lowest possible value, that will practically certainly determine the outcome of the game: Clean eFG%.
If the clean eFG% is 0%, you will lose. And if it’s 150%, you win.
c) If the factor has the highest possible value, that will practically certainly determine the outcome of the game: ORB% and FT Freq.
If ORB% is 100%, every time you miss a shot, you will get a FT set or the ball back. And if FT freq is 100%, you will get FT’s at the end of every possession and the opponents will eventually foul out.
d) Even if the factor has the highest or the lowest possible value, the team may very win or lose, regardless: FT% and Foul eFG%.
Even if you make all your FTA or all your FGA where you get fouled, you may very well end up losing the game. And vice versa.
2) The factor’s potential sensitivity to practice
If the team does all it can to push a particular factor towards the best possible value, how far can it go? This second thought experiment gives us a hint about how potentially sensitive each factor is to practice. In other words, how much it can be enhanced through practice provided that the previous practice has not pre-emptied this potential due to the Power Law (see Chapter 10).
This is a thought experiment out of necessity because in praxis team will not concentrate on enhancing just one of the Twelve Factors but rather on manipulating the whole interaction with the other teams. And each factor only represents one aspect of that interaction.
Here I list the Twelve Factors starting from the most sensitive ones at the top and moving towards the least sensitive ones at the bottom. The main criteria of the sensitivity are:
- The less opponent-dependent the factor is, the more sensitive it is. That is because the improvements made in such factors will not be neutralised by the opponents.
- The more tactically-oriented practice can enhance the factor, the more sensitive it is. That is because – as shown in Chapter 15 – “improving the tactical dimension is the most direct way to enhance a team’s performance.”
Please notice that even if the list based on these two criteria, the order of the factors is speculative and subjective. That can’t be helped because there is no objective way to weigh the criteria against each other. Also, besides tactically-oriented practice, technically-oriented and physically-oriented practice play important roles, too.
Also to be noticed is that the list is symmetrical. For example, Opp FT freq is the factor at the top of the list, and FT freq is the factor at the very bottom. And FT% is the second factor from the top, and Opp FT% is the second from the bottom.
This symmetry stems from the complementary essence of the game. When enhancing a factor, the degree of difficulty is by definition the same for both teams. Otherwise the odds would be biased one of the teams. Thus it must be as easy for you to enhance your ORB% as it is for the opponents to enhance theirs.
1) Opp FT freq. Can be effectively limited even to 0% through tactically-oriented practice alone. If taken to an extreme, this may mean e.g. never going within one meter of an opponent who is in the act of shooting. If less extreme tactics are applied, technically-oriented and physically-oriented practice may be helpful, too. This may include learning to use hands without fouling while blocking shots and improving the jumping ability in order to block shots.
2) FT%. Fairly independent of the opponents’ defense. However, they can affect your FT% through tactics: they can choose to risk fouling only players with a low FT%. But this effect is small, and can be neutralised by technically-oriented training – i.e. by improving your FT shooting.
3) Opp foul eFG%. Can be effectively lowered by making all your shooting fouls hard. Towards this end, a lot can be done through tactically-oriented practice. That is, by learning to avoid shooting fouls altogether unless you’re in a good position to force the opponent to miss his FGA. However, technically-oriented and physically-oriented practice may be needed, too. Examples include learning to use hands properly while fouling and adding muscle mass in order to make the body contacts more impactful.
4) ORB%. Up to a point, can be raised by having more players crash the offensive boards, i.e. through tactically-oriented training. In the long run, technically-oriented and and physically-oriented practice are needed, too. Examples include players learning to tip the ball more accurately and to jump higher. It is difficult for the opponents’ defense to completely neutralise these efforts.
5) TOV%. Dependent on the opponents’ defense. Can be affected through your tactically-, technically- and physically-oriented training. Examples include enhancing shot selection, the ability to catch the ball, and the quickness in spacing the floor.
6) Clean eFG%. Dependent on the opponents’ defense. Can be affected through your tactically-, technically- and physically-oriented training. Examples include improving the execution of set plays, the accuracy of shooting off the dribble, and the jumping ability in order to dunk the ball.
7) Opp clean eFG%. Dependent on your defense. The defense may be improved through tactically-, technically- and physically-oriented training. Examples include improving the coordination in pick-and-roll defense, defensive footwork of the on-the-ball defender, or the jumping ability in order to block shots.
8) Opp TOV%. Dependent on your defense. The defense may be improved through tactically-, technically- and physically-oriented training. As discussed in Chapter 15, examples include improving the timing of a double team, the double teamer’s defensive footwork , and the double teamer’s lower-limb power.
9) Opp ORB%. Up to a point, the opponents may enhance this factor by having more players crash the offensive boards. Also, they may improve their ability to e.g. tip the ball accurately and to jump. It will be difficult for your defense to completely neutralise these efforts.
10) Foul eFG%. Can be effectively lowered by the opponents if they make all their shooting fouls hard enough.
11) Opp FT%. Can be affected through tactically-oriented practice: you can choose to risk fouling only opponents with a low FT%. This effect is small, however, and the opponents can neutralise it through technically-oriented training – i.e. improving their FT shooting.
12) FT freq. The opponents can effectively limit this to 0% if they do their all to avoid shooting fouls. If taken to an extreme, this may mean e.g. never going within one meter of your player in the act of shooting.
3) The league averages
You can compare the Twelve Factors of your team against the league averages and use those averages as baseline values.
If the value of your team’s factor is clearly worse than the league average, there is probably room for improvement. In Table 3, the team’s TOV% is 19% and the league average is 15%. This implies that through relevant practice TOV% could probably be improved.
If the value of your team’s factor is clearly better than the league average, it probably makes sense to concentrate on maintaining that advantage. In Table 3, clean eFG% is 57% and the league average 52%. It seems unlikely that this difference could be made much wider.
4) The team goals and success so far
If the team goals have been realistically set, the success so far may help to assess which ones of the Twelve Factors should be improved, maintained, or ignored. If the success has been better than expected, maintaining the values of the factors might be enough. If the success has been worse than expected, the analysis should be used to spot at least one factor to be improved. If that doesn’t happen, you are settling for underachieving.
5) The team’s characteristics
As claimed in Chapter 1, complex systems “have autonomous parts whose interaction produces emergence.” Because of this, complex systems – such as basketball teams – are by definition unique. Thus, we can’t a priori decide what a team should be like and what its Twelve Factors should be like. Rather, we need to analyse the initial conditions of the team, the context, and the Twelve Factors. Based on this analysis we then decide which factors should be improved, maintained, or ignored.
For example, Table 3 shows that the team has not been rebounding well at either end of the floor. ORB% is 26% and Opp ORB 30% while the league average is 28%. Based on this, it would make sense to pick ORB% and Opp ORB% as targets to be developed through practice.
But what if the initial condition is that – for whatever the reason – the team is an underdog and a bad rebounding team. And that given the initial conditions, 26% and 30% are quite satisfactory numbers. In that case it might be wise to categorise both ORB% and Opp ORB as Maintain factors, instead of as Improve Factors.
Chapter 27: Interactions Between the Twelve Factors
Above we listed criteria for deciding which ones of the Twelve Factors should be developed or maintained through practices or ignored for now. Besides those criteria, it must be considered that the factors affect each other.
This is typical of complex systems. As mentioned in Chapter 1, “they have autonomous parts whose interaction produces emergence”.
Table 4 charts intense interactions between the Twelve Factors. Lines connect those pairs. The nature of each pair’s interaction is discussed below. Beforehand, please notice this about the table and the discussion.
- Even though the lines are similar in colour and thickness, the nature of interaction varies from interaction to interaction.
- The line or the interaction may refer to one-way of two way effect.
- The interaction here does not refer to a causal relationship between the two factors connected but rather to a tendency. In other words, when a certain change occurs in Factor A, a certain change tends to occur in Factor B.
- In the chart there are twelve interactions marked by lines, yet below there will be only six discussions. That is because each interaction has a complementary pair, and the same discussion goes for both interactions. In other words, when there’s an interaction between two offensive factors (Clean eFG% and FT freq), there is going to be a similar interaction between the corresponding defensive factors (Opp clean eFG% and Opp FT freq). And when there’s an interaction between an offensive factor (Clean eFG%) and a defensive one (Opp TOV%), there’s going to be similar interaction between the corresponding defensive–offensive pair (Opp clean eFG% and TOV%).
- The twelve lines only describe the main interactions between the Twelve Factors.
1) TOV% – Clean eFG%
Turnovers are usually committed while the team is trying to find a high-percentage shot. If not for that, why would the offensive team do anything but sling shots?
So an attempt to increase Clean eFG% will involve changing the quality and/or quantity passing and/or dribbling. This exposes the team a higher risk of TOV.
The interaction works the other way, too. Usually it is not that hard to lower TOV% per se. As implied, the team may simply start to sling shots more quickly – or before a large portion of the accustomed risk of TOV even occurs. What is difficult, however, is lowering TOV% while maintaining the current Clean eFG%.
2) Clean eFG% – Opp ORB%
If the opponents want to raise their ORB% the most usual thing to do is to have more players go after the offensive rebounds. This may work because at least up to a point, the number of players crashing the offensive boards correlates with ORB%.
However, there’s a potential downside to this tactic. As more players crash the boards, less players get back on defense to protect against the fast break. This may make Clean eFG% grow.
3) TOV% – Opp clean eFG%
if the team’s TOV% becomes higher, Opp clean eFG% may change correspondingly. That is because the opponents may get to start more possessions off a steal, and such possessions are on the average more productive than other possessions.
A steal here means that there is a turnover and after it, the ball stays live. During the NBA regular season 2018–19, a possession that started with a steal produced 1.28 points per possession (PPP). PPP of all possessions in the league was 1.11.
Not all turnovers are created equal. If the team is to turn the ball over anyway, it’s best to throw it out of bounds or to commit a violation – that is, to cause a deadball. That type of turnovers may actually lower Opp clean eFG%. That is because the possessions starting with a deadball had the PPP of 1.08 – or lower than the total PPP.
4) Clean eFG% – FT freq
If the team has a high Clean eFG%, what is the defense to do? A common solution is starting to contest the FG shots more closely. Then the opponents take the risk of committing more shooting fouls – or adding to the team’s FT freq.
On a large scale, this happened in 3P shooting. As it has become more and more dangerous an offensive weapon in the NBA, the portion of 3PA that lead to a shooting foul has grow.
5) ORB% – FT freq
Offensive rebounds often lead to putbacks, or quick shots from under the basket. Thus, as the ORB% grows, so probably does the portion of close-range FGA of all FGA. This may raise FT freq because the closer to the basket an FGA is taken, the more probably there is going to lead to a shooting foul.
6) TOV% – Opp foul eFG%
If Opp foul eFG% is high, they earn a lot of and-ones by getting fouled in the act of shooting. When trying to lower Opp foul eFG%, the team may try to force more misses. This attempt may lead to the fouls being harder and subsequently to causing more unsportsmanlike foul calls. This will raise TOV%, because as mentioned in Chapter 19, unsportsmanlike fouls count as turnovers.
Chapter 28: On the Limitations of Sports Analytics
In the previous two chapters we have discusses the characteristics of the Twelve Factors (Chapter 26) and the interactions between the factors (Chapter 27). Our goal is to use this discussion as a basis for modelling how we should decide “which ones of the Twelve Factors should be developed or maintained through practices or ignored for now”.
We will proceed to that in the next chapter. For now, a word of caution is in place:
No matter how thorough and reliable our analysis of the Twelve Factors is, we cannot validly draw objective, detailed and unconditional conclusions regarding which factors should be developed, maintained, or ignored – i.e. what the team should work on in practices.
The reason is not the quality or the quantity of the data but rather the essence of the entities. As we noted already in Chapter 1, basketball game, teams and players are all complex systems. And “based on the input, one can not exactly know what the output of a complex system will be”.
In other words, we can not know in advance exactly how practice will affect a player, exactly how a player’s actions will affect the team performance, and exactly how the team performance will affect the game.
This very same limitation concerns not just the Twelve Factors but all sports analytics – simply because it all deals with complex systems. Keeping this in mind is elementary because sports analytics is becoming a more and more important part of sports and coaching.
The term “sports analytics” is often left undefined. Usually the underlying definition seems to comply with this one: “the use of data and advanced statistics to measure performance and make informed decisions, in order to gain a competitive sports advantage”.
The growing importance of sports analytics is not hard to understand: modern technology is able to tell in astounding detail what players and the team have done in past practices and games. This ability has tempted many into overestimating the analytics’ ability to tell what players teams and players should do in future practices and games.
Chapter 29: Heuristics as a Decision-Making Method
So, there is no way to draw objective, detailed and unconditional conclusions regarding what the team should work on in practices. What to do?
The paradox is simple yet unsolvable because of two things:
- In a complex system, there is no objective way to tell the optimal number of criteria to be used in a decision-making.
- “The number of criteria in a given decision-making problem is critical in deriving which method appears to be the best.”
Thus when we apply a method to choose and justify our training targets, the selection of the method will necessarily be subjective.
Some may question if such an explicit method is needed at all. That is because traditionally, coaches have made decisions subjectively and intuitively, without explicitly referring to any decision-making method at all.
However, there are advantages to having an explicitly defined decision-making method:
- It allows having an intelligent, relatively objective discussion about the decisions.
- It makes it possible to explicitly assess and improve the decision-making method and process based on the outcome.
In this context I choose to apply heuristics in the decision-making. As mentioned above, the choice is subjective up to a point, but I think it can be justified reasonably well.
“adaptive tools that ignore information to make fast and frugal decisions that are accurate and robust under conditions of uncertainty.”
My rationale is that in sports, heuristics are often used to make decisions. Sometimes heuristics are applied implicitly, sometimes explicitly, but they’re ubiquitous. And because they will be there no matter what, it makes sense to keep utilising them and to start assessing them critically.
For example, substitutions are often based on heuristics. Coach may have explicit rules of thumb, like “If the number of a player’s personal fouls exceeds the running number of the quarter, take him out.” Or the heuristics may be implicit: coach may take out any player who seems to be breathing heavily.
Chapter 30: Four Classes of Heuristics
Mousavi and Gigerenzer (2017) list four classes of heuristics. “Each class highlights a main characteristic of how humans deal with information.” The characteristics are:
- Recognition ability
- Equal weighting
- Sequential search
The classes based on Mousavi and Gigerenzer’s article are as follow. Each description is followed by an evaluation of how the class might fit our purpose – that is, forming heuristics to determine what the team should work on in practices. All quotes are from the same article.
When “one option is recognized but the other is not”, the recognised one is chosen. These heuristics assume that the more familiar one will be more valuable.
In our case this class does not appear to be useful. That is because the number of possible choices is quite limited. There are only twelve of the Twelve Factors, and they all should be recognised by coaches.
Equal weighing schemes
If there is an “absence of reliable assessment of the relative importance of different components”, then each component is given equal attention. In our context, this would imply that an equal amount of practice time was devoted to each one of the Twelve Factors.
This is does not appear rational because we have two reliable assessments of the Twelve Factors’ relative importance. The two assessments were presented in Chapter 26: the factor’s potential importance and the factor’s potential sensitivity to practice.
As discussed, these assessments do not put the factors in a definite order of importance. Yet they do show that the factors are not of equal importance.
Sequential consideration of reasons
In this class the reasons are put in order “based on their validity or some other ordering principle”. Then “information is sought piece-by-piece” until “goal-driven fulfilment” is achieved.
In our context, this would mean that the Twelve Factors were put in order of importance. The most important one is considered first. Then it is decided whether that factor should be improved or maintained in order for the team to reach its goals.
Next, the second most important factor is considered in the same manner. Next, the third most important one, and so on.
For our purposes, the sequential consideration heuristics appear useful for two reasons.
- As shown in Chapter 26, we have two justified ordering principles available: the factor’s potential importance and the factor’s potential sensitivity to practice.
- The number of relevant factors is limited. It is possible to consider all Twelve Factors – not just in theory but also in praxis.
Satisficing strategies much resemble the sequential consideration heuristics: here, too, the reasons are considered one by one. The search is stopped “when the first object is encountered that satisfies the aspiration level”.
The difference between the two heuristics is that in satisficing strategies “there is no prespecified order to the search and examination of information”. Also, here the goals are adjusted as “the examination of information proceeds”.
Because of this, if we were to use satisficing strategies, we would by definition neglect information provided by the Twelve Factors. This would be a loss, because we previously seen that this information is useful.
So, sequential consideration heuristics are better suited for our purposes than satisficing strategies or any other heuristic class.
Coming up Chapter 31: Twelve Factors and Sequential Consideration of Reasons.