Science is a bad word to basketball trainers and coaches. Coaches and trainers tend to believe experience or other coaches and trainers. However, we are in a statistical revolution in basketball, and numbers are gaining a prominent place in coaching.
The problem with relying on experience is known in statistics as a problem with generalizability. I watched a trainer once whose entire belief system and methodology derived from what he had done as a player to ascend to a pretty decent playing career. Of course, this means that the trainer used a sample size of one to derive his methodology.
As I prepare my dissertation studies, I have to account for power and effect sizes and ensure statistical significance. As an example, to have statistical significance, I have to be 95% sure that my results that I find with my sample population are generalizable to my target population. If I lack this statistical significance, my results might be due to the testing procedures, the specific participants, or other errors.
If I conduct a study using 12 high-school freshmen from Salt Lake City, my goal is not to find out information about those 12 freshmen. Instead, I want to generalize my results to a wider population: preferably all high-school basketball players, but more likely all freshmen basketball players within the United States. However, with only 12 participants, it is unlikely that I can prove that my results are in fact generalizable. With a sample size of one, I would find only what works for one person in one specific set of circumstances.
What works for one person may be generalizable to a large population. However, the trainer attempted to generalize his results to players ranging from beginner to professionals, small to tall, boy and girl. One size does not fit all.
In more concrete terms, imagine that you and I wanted to learn about shooting. You decided to watch films of Reggie Miller, while I decided to watch films of Steve Nash. We would draw very different conclusions about proper shooting technique. Both are great shooters. However, would you teach a young player to shoot just like Miller? It worked for him in his Hall-of-Fame career, but would it work for everyone (or 95% of the population)?
When a coach coaches based solely on his experience, he uses a sample size of one. These results are not generalizable. Just because a system of training worked for one person does not mean that it will work for everyone. This is one problem faced by great players when they attempt to coach or train: They attribute their success to their training methods or the way that they were coached, but their success may have been due to genetics, height, luck, skill, work ethic, talent, etc. Reggie Miller’s shooting success may not have been due to his shooting technique but his practice habits and confidence. If a player attempts to emulate Miller’s technique, but does not have the same habits, confidence, size, etc., the technique is unlikely to work as well for the player.
In coaching or training, most of our approaches should be generalizable to a larger population. Of course, a coach or trainer should teach each player like an individual as well. There is not a one-size-fits-all approach to training or coaching. Therefore, a coach or trainer cannot treat each player like a mini him or her.
I have a method for teaching shooting that I believe to be generalizable. It is not based on one great shooter, but many great players and coaches, and many parts of the approach are supported by research. For instance, when I write that all players should shoot like Steve Nash, my belief is based on his shooting success, but is supported by motor-learning theory.
Each player starts from a different spot. Some players have bad habits to address or movement abnormalities to train. Identifying these differences and creating individual plans are important elements of coaching and training. However, the method starts with an approach culled from many sources, and not just copycatting one person’s success. One’s personal experience is valuable, but if one believes too strongly in a sample size of one, the experience can be limiting and lead to errors of attribution and ultimately less than optimal training or coaching.