The art of sports science

There is a tendency in sports to treat athletes as machines, and to treat a team of athletes as a larger machine with each athlete acting like a cog in wheel. This isn’t unique to sports by any means and is common practice in many industries. However despite similarities, people are not machines and it is just as important to recognise the differences, which are pivotal to maximising human performance, as it is to accept the similarities.

Team Management: Are we creating a culture of helpless athletes?

Our desire to control the output (successful performance) by strictly controlling the parts (athletes/procedures) seems to now be permeating all areas of athletic development and team organisation. In many respects we are turning athletes into dependant robots. They rely on the professionals around them to tell them when to train, how long to train, how hard to train, when to eat, what to eat, how to recover, when to go to sleep and even what they should and shouldn’t be doing in their ‘free time’. The more control we give ourselves over the athletes inside our program, the less empowered they feel and the less responsible they feel for their own actions. By treating them too much as machines rather than as people  we are reducing and hindering the powerful aspect that makes us more than the sum of our parts, and in turn a team greater than the sum of the athletes that make it up. We reduce variability and small errors made by the athletes by strictly enforcing team rules or following rigorously detailed plans, but in doing so it’s very easy to lose the greater potential benefit which is intrinsic to people and extrinsic to machines, which when harnessed can result in enormous synergy and success to the individuals, team and organisation.

Reducing small variation, or errors, can have the unintended effect of creating much larger errors. For example, an athlete who’s spent the majority of a season feeling oppressed or controlled, treated more like a robot and less like a person, might choose to reclaim some of this independence in a big night out partying and drinking with friends. If this occurs around finals then this could be a big problem, but the larger problem is the headspace of the athlete who has now disengaged with the program. This is more like a chronic injury, such as a tendinopathy, compared to an acute quad tear. It took longer to develop, with no single cause identifiable but rather due to many contributing factors, and like a chronic injury once it presents as an issue it will take longer to resolve.

By reducing the control over athletes inside a program, treating them instead as intelligent people capable of making good decisions, you empower them and place the responsibility of their performance back onto the athlete. There may be smaller errors earlier in the season, perhaps with a player not preparing themselves as perfectly for a training session perhaps as we would like to see, but the way that situation is treated and worked through with the athlete can result in a positive growth of the individual, which actually reduces the chances of the same situation occurring or perhaps anything worse occurring later in the season. Coaches now shift from authoritarians ruling with an iron fist, to becoming recourses for athletes to use to further their own development and maximise their performance for the team. The athletes shift from being programmed machines, doing as they’re told, to engaged and motivated people who are constantly learning and striving to maximise their potential. This is the attitude Leicester FC has recently adopted and has seen enormous success with. The biggest advantage to this approach is that it is highly likely to lead to long term success. Essentially it is the old parable, give a man a fish and you feed him for a day, but teach him to fish and you feed him for a lifetime. The culture the club is creating is one that is antifragile. They will experience lots of small errors made by players but these errors will allow learning to occur and resilience as a result, making the group ever stronger. All the while they are reducing their vulnerability to catastrophic mistakes, such as the player going out and having a big night before finals. Mistakes and errors will always occur, the harder we try to prevent them and the bigger a deal we make when they are broken, the worse the errors seem and become. To take the opposite approach is a healthier, robust mindset. Assume that errors will occur, try to identify and mitigate the reasons behind why they occur and learn to grow from any that do so that they are less likely to repeat. Leicester FC is prioritising learning and personal development while empowering their athletes with a growth mindset, rather than a restrictive mindset which would cultivate fear of mistakes or broken rules. They are expanding their athletes choices rather than reducing their choices. Considering it is impossible to control athletes 24/7, and that it is impossible to rule out mistakes or broken rules, Leicester’s approach is far more antifragile and ties in with the synergistic approach to team development this article is presenting. Rather than restricting their athletes by treating them like machines, they have empowered them as people and have taken the cap off their athletes and teams potential performance.

Restricted by classifications

Science loves classifications, this makes other people’s work more objective and relatable. It’s easier to interpret what other people are writing about or researching when you know what it is they are referring to when they use certain words which have an agreed upon definition. Very often this involves simplifying a more complex occurrence. We often sacrifice a degree of relevance, for specificity – or validity for reliability. One of the world’s most distinguished quantum physicists and a philosopher, Werner Heisenberg, warned: ‘…the scientific concepts are idealizations… But through this process of idealization and precise definition immediate connection with reality is lost’. This is a necessary part of language and communication, but problems arise when we start locking ourselves into these definitions which we have ourselves created. These definitions could be energy systems used (such as aerobic or anaerobic) or types of muscular performance (strength, speed, power, agility, stability). In reality, life, sports and the human body is incredibly complex. The more we simplify aspects the easier it is to forget how complex and interrelated things actually are, and problems arise when we prescribe training methods forgetting this. As was mentioned in part 1, we have a deep desire to understand things completely, we struggle with loose threads and chase closure. So we are motivated to believe that this simplified event actually behaves like we have defined it. In fact the more frequently a certain classification is seen to be correct, arguably the more dangerous it becomes. People expect to see what they expect to occur and evidence shows we do a very good, albeit irrational, job at explaining away anything contrary. What this means is that rather than treating athletes as individuals we can fall into the trap of treating them as research cohorts  and essentially trying to fit a square peg to a round hole, often at the expense of individuals athletes performance. Classifications and drawing upon others research is undoubtably a beneficial thing to do, but it is important to critically review these studies and closely monitor the responses of your own athletes being careful not to push too hard to make something work which might not be the best fit. We are in an age where we have more information than we know what to do with it, the challenge is now figuring out what is useful and what is not.

Since we seem to run into problems with relevance or transfer back to the real world when we get caught up in the details (reductive error) then perhaps a more appropriate means to view athletic performance is in terms of habits and trends. A better approach is to step back from the microscope and look at the movement habits and trends inside games. What are players doing consistently, what types of movements, when do they tend to rest and in what positions, how often do they tend to sprint? What are the specific players habits? When do they consistently get themselves into trouble, when do they consistently perform well by beating or stopping their opponent? Do their current habits suit what their teams specific game style requires? By viewing the trends and habits of the whole game, it might be harder to to identify whether specific moments are more or less dependant on the anaerobic system, the aerobic system, strength, power or stability, but despite lacking a detailed understanding or classification system, you actually have a very accurate view of the sports demands. In fact the view you now have is more accurate because we haven’t attempted to analyse all the specific details. The trend inside science is to understand all the details of these moments, which is fine and can provide much benefit, but problems arise when we get too caught up on the details and end up missing the forest for the trees. So there is a time and a place for digging into the details, but we must not get caught up on them to the extent that we lose sight of the actual event we are targeting improvements in. The ability to classify, quantify, measure, prescribe and produce a consistent expected response is useless if it isn’t directly relevant to that athletes competition performance.

Strength programs

Often this error is seen in strength programs inside the gym. All good strength and conditioning coaches would agree that the end result they’re developing is in-game performance, not necessarily what the athlete can squat or lift in any other gym exercise. Yet since on court progress is hard to measure and quantify, but the squat is very easy to measure and quantify, we tend to use that exercise and others to grade progress. The squat is the example we’ll use here but the concept is the same for any test or exercise removed from the competition setting. There is some good evidence that improvements in the squat do tend to transfer to game performance improvements, but it is still not the be all and end all. After all, there are also plenty of athletes who never squat and do very well at the international level. This over-valuing of squat performance is made worse if throughout the season the squat, or any other gym exercise, is presented to coaching staff as a mark of progress, perhaps with predicted 1RM values presented monthly in team meetings. But even the squat, as ‘functional’ or perhaps beneficial as it is, isn’t truly sport specific. The more that the load or weight on the bar is increased, the slower the movement becomes, and the exercise encourages eccentric action followed by concentric action acting at the knee. However if looking at basketball, or any acceleration or change of direction moment in other sports, a slow action at the knee, or an eccentric action on foot impact is actually a negative occurrence and increases loads on the body (muscle damage, inappropriate tendon loads) and slows players down. Think about a player planting a foot for a lateral change of direction. Better players will hit the floor and instantly be pushing off to go where they need while slower players will spend longer with their foot grounded and will see more eccentric action before pushing off since they require a longer time to reduce the force. Paradoxically, while the squat will improve muscular strength, it also encourages this slower inefficient movement habit. Similar problems occur when we look at specificity of action at the calf. Up to 6x body weight is measured as the force acting on the calf during running and jumping. If a 100kg athlete wanted to see this level of force at the calf when training a calf raise they would require 500kg on the bar, an incredibly unrealistic expectation. Not to mention that in competition the level of force acting on the muscle increases as speed of action increases (such as sprinting), yet as weight is increased in gym exercises the speed of action decreases (the same problem we saw in the squat and change of direction example). We gain similarity in force at the expense of similarity of action. In the case of the calves it becomes even less specific, in explosive actions the calves largely function in an elastic manner attempting to retain energy developed at the hip and transfer it into the ground for a spring type action. Avoiding excessive movement here is actually very important to prevent muscle damage from micro-tears. Yet in heavy calf raises we train the muscle to undergo slow eccentric-concentric actions which is the opposite of what we generally require in competition. Coaches such as Frans Bosch are creating a new culture in gym training which not only values the physical load on athletes and their muscles, but the nature of the loading as well as the motor learning and coordinative effects of strength training.

Ultimately we need to look at the moments in the game where players tend to beat other players, or get beat, and the moments when the most load is placed on the players body and the injury risk (both acute and chronic) is highest. These are the moments we need to improve performance in, and very often these moments identified for a performance benefit are also the ones where injury risk is highest. An example in basketball and many sports would be change of direction moments. These tends to be the times when players get beat (on defense) or beat someone else (offensively) and their injury risk/joint loads are highest. It’s not enough to simply take traditional strength exercises like the squat and assume they’re getting the job done. We need to look at the demands of the key game movements and look at the ways the body is loaded in them. We need to teach the body to express it’s strength, power, resilience in these moments, not just continue to build them with traditional gym movements while assuming transfer will occur without further prompting.

We tend to break up a gym program into various components, perhaps mobility, stability, power and strength or compound exercises to name but a few. But really this just makes the training program nice and neat or perhaps more measurable. If messing things up by moving exercises around helps improve coordinative benefit for example, you’re better off moving a power exercise inside the strength section, or perhaps to superset a stability exercise with a strength based exercise. This might make that strength session harder to measure or to compare to previous phases, but that’s not the end goal. The real reason we want to measure and compare is because we assume it will help us see what’s resulting in improvements in game performance through cause and effect, but if in attempting to do this too tightly (control and measure) we actually hurt our ability to influence game performance improvement (through reduced relevance of training), then we must definitely step back and embrace the ‘art’ of training and coaching rather than attempting to treat it as a pure science.


Physical testing is often performed with the same errors resulting in a simplification of game demands, sacrificing specificity for measurability. There is value in assessing various athletic capacities in order to either screen for imbalances (perhaps a left vs right hop for distance test) or to measure performance (such as a vertical jump test). However neither of these are directly specific to game moments and even the strength of their ability to predict injury is inconsistent. You might counter that plenty of vertical jumps occur inside a basketball game for example. But none occur in a setting where the athlete has all the time in the world to prepare themselves, gather themselves and complete the jump without needing to react to any external stimulus or zone out from any irrelevant distractions. Not to mention these tests tend to be performed while the athlete is ‘fresh’, very few jumps inside a game occur when an athlete is ‘fresh’ so their performance under fatigue is perhaps more relevant than the test measures. So while the vertical jump test is related to competition, the test is not truly relevant. Arguably the best value tests would be ones which create game settings that can be filmed and test how a player moves in response to a game moment, under all the distractions and pressures of a game. These are specific to games, because they are occurring in games, but they become harder to measure and quantify which also makes comparison later or during retesting harder. An example might be assessing how a player moves in response to a change of direction in play (such as a 180 degree turn and run). Do they appropriately lean into the acceleration and drive backwards with their first few sprint steps while retaining stiffness through the core to reduce energy leaks and enable the arms and legs to work together synergistically? Based on errors in this moment, running technique work can be implemented and adjustments to a strength program can be made to suitably address deficiencies in form. While measuring actual speed in this moment might be difficult to do, it is occurring in a game setting (factoring in reading the play and reaction speeds) rather than an artificial setting, perhaps involving timing gates, and is directly relevant to game performance. What’s more important, truly measuring what is important to that athletes performance, or having a neat test with a quantifiable numerical result?

People are not machines: Physics and it’s limitations

Our attempts to mechanise people is also evident in the heavy reliance on physics in our strength training methods. Biomechanics presents the human body as a moving bundle of angles and force vectors relative to the forces of physics. Our gym programs are periodised to primarily manipulate volume and intensity, a very mechanistic approach to variation. Yes we might change exercises around every few weeks as well but really the variance isn’t huge, and the exercises we vary tend to be related to each other as gym based strength exercises rather than game moments or movements. For example we might alternate between back squat, front squat and box squats as our primary knee based strength exercise over 3 cycles. But the goal exercise remains the squat. We tend to over-prioritise load in the squat since it is arguably the best exercise to develop lower body strength in the gym and as a result we presume that it is the best way to develop lower body performance in competition. This would be true for machines, replace a weaker part with a stronger part and you will see improved load bearing or strength in all tasks. But this is because the machine has been programmed to complete a certain task with very specific movement program which remains the same despite external changes – all that has changed is a stronger part. An athletes performance isn’t anywhere near so simple and as this article has explored, is more than just the sum of it’s parts. Unlike machines or robots, athletes do not have programmed responses to game events. Yes they do train certain habits, but they are still vulnerable to changes in coordination or variation in motivation levels or a myriad of other factors which machines aren’t subject to. If machines have only one variable – quality of machinery, then athletes have many and they are all related and interdependent on one another. Yes we appreciate the performance improvements the squat can provide, but we must also be aware of it’s limitations and the negative effect a strong squat focus teaches. That it encourages slow eccentric-concentric action at the knee and hip and that the more the load is increased the less specific it becomes to game moments (higher load-lower speed in the squat vs higher load-higher speed in game moments).  In fact the number one focus of most gym programs to increase the amount of weight used in compound lifts, is actually reducing the specificity of the movements and having a negative coordinative effect.

So if we are limited in the specificity of force development by using a physics based approach to gym development, how should we view gym based training? Improving performance inside game specific movement patterns and habits would be key, then you’d want to build physical resilience in them. Again looking at the squat, an athlete arguably receives the greatest benefit to their performance in learning to sit backwards into a squat without their knee’s drifting forwards or falling into a valgus position while learning core stiffness. Once they become proficient in this they begin loading up the bar and increasing the amount of weight they are squatting but they are at a point of diminishing returns. The greatest benefit they received, and will receive, was in learning to move in a way which increased strength and power from the hip and stability at the knee and through the torso. This isn’t to say you never load athletes with weights, this would be a means of improving resilience and thus reducing how beat up they get from competition. But it does reduce the need to load them excessively and too frequently in the weight room. We have also discussed about a potential negative effect of chasing this resilience through increased load and that is a reduced specificity of the movement and often encouraging of movement patterns, or habits, that we in fact do not want to see in competition. The learning of the movement was the largest benefit, and that occurs at the brain. If we begin to look at the motor learning component of exercises it shifts the emphasis from loading the athlete, to providing the athlete opportunity to learn and refine efficient movements while building resilience in them. Step ups and their variations are not just to strengthen the quads and glutes (a muscular/physiology approach), nor to simply improve strength or force potential at the knee and hip (physics/biomechanics approach) but to allow the athlete opportunity to learn to improve their efficiency of movement of their ‘sprint steps’. Essentially we have isolated the most important part of an athletes acceleration and are now providing them the opportunity to get better and better at it without the distraction of other aspects of the game. Simply performing a step up with good form will provide much of the muscular stimulus, and strength/power improvements. In order to get the most out of that exercise, the athlete should be thinking about their first step on an acceleration, whether it be running, defensive or handling the ball. If they can keep this mental image while executing each repetition of the step up, and hold the idea in their mind that they are teaching their body that their first step in these moments is a powerful drive down into the ground, then they receive all the physiological benefits which allows them to be quicker on the court, but also the psychological habits and intent which means they will in fact be quicker on the court.

Like the squat example, athletes will receive the most benefit to this process early on when the step up emphasis is introduced and refined. Following this, as weights are added, they do receive benefits but not at the same rate. Since ever increasing the load isn’t the best way to necessarily see improvements, given reduced relevance of coordinative patterns and muscle-force relationships seen with higher loads, what now becomes the focus? Building resilience inside this movement, reinforcing good habits to offset any bad habits that might creep in from sloppy movements in competition or training, and improving execution in a range of settings becomes key. Gym performance is improved when variation between exercises is reduced. We tend to complete one exercise then move onto the next, or might superset it, but tend to avoid circuit type training for athletes. After all we’re chasing strength improvements and this requires the athlete to be ‘fresh’ and prepared for the intended exercise set. However, that is relevant when increased load on the bar is our primary focus. If we adopt a coordinative approach, then by adding more exercises between sets, we increase muscular confusion. Making it harder for the athlete to ‘find’ the right movement pattern for that exercise. For example, Louis Simmonds from Westside Barbell, an elite powerlifting club, played around with explosive jump squats between squat sets as ‘primers’ to improve performance in their strength sets. After experimenting though, he decided that the speed of the jump squat actually hurt performance in the next set of weighted squats since the timing was so different. This is a negative if your focus is on performance in that set of heavy squats. But in a game of basketball you go from performing any number of movements, such as a slide or back-pedal, and then are required to accelerate quickly in a sprint step in response to a stimulus. Despite earlier mental and physical ‘distractions’ the athlete must now quickly find the most effective way to accelerate in their sprint step. In this game setting, it is irrelevant how much power the athlete can express in their acceleration steps if allowed appropriate time, here the limiting factor is their ability to ‘find’ the most appropriate movement pattern in a split second and execute it. Back to the gym and Simmonds jump squats. While the muscular confusion was a negative for that set of heavy squats where purely gym performance was the goal, the benefit of ‘finding’ the most effective way to execute a squat, or step up, having previously been performing unrelated or similar exercises, is actually of significant coordinative benefit. In powerlifting the squat is the end goal, this is the exercise and the test. In sports, the squat is just a tool to develop movement ability in related, but distinctly different movements, such as sprinting. Perhaps sacrificing a small percentage of weight used in the next squat set for the benefit of improving performance in times of muscular confusion is arguably, at least at times, a more effective tactic for improving transfer for field and court sports athletes.

As soon as we adopt an approach that prioritises learning as well as muscular changes, we break down the traditional strength template which 90% of sports program look like. No longer are exercises placed under restrictive classifications included as section headings inside the program, such as stability, power or strength. But they are now structured in such a way that achieves a coordinative benefit. Exercises are not completed in a linear manner, but are jumbled up to challenge the athletes ability to execute various movements one after another. These might be dissimilar, such as a squat and a bench press. Or similar, such as a squat and a plyometric jump with any number of other exercises placed between the first and second set. There may even be changes in how the movement is completed set to set. For example, the step up box height might change on each set. Or it might transition from a barbell step up, to dumbbell-goblet to plate-overhead. Adding resistance bands into the mix allows increased variety while enabling the coach to retain the relevance of game movements.

The rationale behind the exercises selected and the order used is passed onto the athlete. They are educated that the gym is not just about strengthening muscles, but is also an opportunity to refine key movements which occur in the game. If athletes can see the relationship between the gym and competition, you should see increased motivation and improved transfer from gym to competition improvements. Athletes benefit when events are simplified. They want to know how they can win each drill, and they are very good at learning the most efficient way to do so inside the rules they’re provided. If you can present a start position for a gym exercise, and an end position, then the athlete simply has to focus on getting from A to B as quickly, smoothly or efficiently as possible (depending on your individual focus for that athlete and movement). From this you can begin more detailed coaching of the exercise technique, but with this as a starting point you tend to see most athletes get 90% of the way there themselves without excessive coaching and verbal cues, both of which restrict performance transfer out of the gym.

Embracing the art of sports science to improve relevance

One of the strengths of science is the collective accumulation of knowledge, which largely occurs through measuring and recording. But since athletes will never be as controllable and consistent as machines, and then if sports can’t or shouldn’t be treated as a pure science which can be controlled, reduced and quantified, then it seems a compromise is needed. One that embraces rather than fights against the art of sports science while retaining the benefits of the scientific method. One that sacrifices an element of specificity for relevance.

A solution in the gym would be to know what your general philosophy or goal is for a specific gym program, based on the team and individual requirements at that time of the season and the head coaches goals. This would be the idea or mind frame you hold while you build the gym program. While the exercises themselves will differ and target different physical and mental characteristics, they all are in there because of your general philosophy or goal which permeates all the exercises, sets and rep schemes and allows the collective synergistic effect of the program to have the desired effect on the athlete. While this broad, permeating idea is harder to measure or define, it is no less a part of the program than the specific exercises or set-rep schemes are. All good training programs will have this built into them, whether it is clear in the mind of the designer or perhaps more a subconscious influence. However it is extremely valuable and important for the S&C coach to take the time to specifically decide upon the general goal of the program before they begin designing it, and even to make this clear to the athletes. If the athletes can clearly see what you’re trying to do with the gym sessions and how that directly benefits their performance, or they can better see the link between a gym exercise they’re improving and game moments, then the transfer they see from improvements in the gym to performance in competition will sky rocket. Evidence of this sort of thing is seen in the placebo effect of drug trials, or even expectations of drugs for that matter. Evidence consistently shows that if people expect a thing to have a certain effect, then that’s the effect they tend to see. If players can’t see the relevance of something, they will only ever see a limited improvement or transfer from it. Sell your program in a way the athlete can relate to and even if your program isn’t ‘perfect’ you will get a better result from someone else who doesn’t sufficiently ‘sell’ their program to the athletes.

A solution to measuring changes in competition performance, such as speed, without isolating the trait with a 20m sprint or similar test, would be to use more subjective approaches. A subjective method might involve getting coaches to fill out a survey for each player at the end of each 2 week period. One of the questions might relate to players speed during training and competition, such as asking the coaches how fast they thought the player to be, perhaps with either a quantitative score (7/10) or qualitative score (very fast).  Of course you would need to determine whether you wanted this score to be compared to the players team mates or some other scale. Or perhaps you might use a score relevant to their prior performance average as held in that coaches mind (much slower, slower, no change, faster, much faster). You could also have a score for speed during offensive moments, and a speed for defensive moments. There would be value in getting the player to rate their own progress in these areas. Then you could average the scores, maybe weighting higher the head coach and players scores, to determine that players performance in the target areas. This is a form of testing which doesn’t increase risk of injury or training loads to the players (as throwing a 20m sprint test into a training session would), is directly relevant to the demands of competition (speed, reaction, reads, decision making etc) rather than being an isolated test with reduced relevance (20m sprint between timing gates), and encourages the athletes to focus on these key performance indicators throughout the week, increasing the athletes focus on improving in these areas during training and games – where it really matters.


  • Athletes and teams are frequently compared to machines which in and of itself is not a problem. However problems arise when we begin to think of and treat athletes and teams too much like machines.

  • There is an underlying motivation for people to consistently think of athletes and teams more like machines and less like the complex and unique humans they are in order to control performance. However this is a ultimately utopian dream.

  • It is encouraged by three related trends in modern science which is to view things in mechanistic, materialist and reductive manners. This encourages the view that wholes are explained entirely by the sum of their parts and that matter is the only reality.

  • An alternate view which seems to be easier to relate to in sports is that instead of wholes simply being a sum of their parts, that instead each part is itself a smaller whole. Each whole has an element which makes it greater than the sum of it’s parts. This might be represented by an individual athlete performing beyond their physical capacities, or a cohesive team beating a team comprised of better skilled individuals.

  • This alternative view encourages that there is a holistic element of athletes and teams which cannot be catered for simply by focusing on physical characteristics such as training, nutrition and recovery, but must involve psychological aspects and relational skills which can more readily be expressed or recognised as an art than a pure science.

  • To reduce athletes or teams to their parts is to cap individual or team performance and risk breakdowns in performance when individuals needs are not met.

  • To get caught up on the measurable characteristics of athletic or team performance and develop them outside of competition is to risk sacrificing relevance for specificity. Drills, strength programs and performance testing very frequently seem to make this error and priorities quantification over relevance. The dominance of physics in our strength programs and testing is an example of this error.

  • A solution to some of the limitations in the scientific approach to athletic development is to step back from the microscope and view the big picture, identifying traits, habits and underlying intent that involve the athlete or team as a whole and factor in their environment (competition).

  • Science itself is not the problem, science is a fantastic method of enquiry and should readily be utilised in athletic development. But it has it’s limitations and there are errors which can arise when science becomes a means of seeing the world entirely, rather than simply a tool used at moments of our choosing to interpret the world around us. .