If I administered a test to you, scored the test, then gave you the raw score you received, what would you know? For example, if you complete an IQ test and score a 105, what does the number 105 mean? Furthermore, assuming a score of 105 is slightly above average (it is), what does the score allow us to better understand about you, or predict about your likelihood to experience future success? Scoring a specific number is one thing, but if we don’t know how to understand and properly interpret the number, what do we really know? Today, sport analytics present the same challenge — sure, we acquire specific data supposedly relating to future sport success, but do most people know how to use this information responsibly?
Understanding sport statistics
Sports analytics are everywhere these days, and there is little doubt they have made coaches and athletes more informed than ever before. Radar guns measure pitching velocity, cameras calculate spin rate and exit velocity, wearable technology tracks speed and fatigue, and sophisticated software can compare athletes against thousands of others across the country. The collection of data itself is not the problem. The more interesting question is what happens after the data are collected.
In psychology, one of the first questions we ask about any assessment tool is whether it actually predicts what we think it predicts. This concept is known as predictive validity. Simply because a test measures something reliably does not mean it accurately forecasts future success. The same question should be asked of sports analytics. Does a high exit velocity at age 12 predict becoming a successful high school hitter? Does a high spin rate on a baseball predict future pitching success? Sometimes the answer is yes—to a degree—but rarely is the relationship as straightforward as many people assume.
Today, sport analytics present the same challenge — sure, we acquire specific data supposedly relating to future sport success, but do most people know how to use this information responsibly?
Another important concept is external validity. Just because an athlete performs well in a controlled training environment does not necessarily mean those same results will transfer to meaningful competition. A player may produce outstanding numbers in a batting cage, but can those same mechanics be repeated against a live pitcher with two strikes, runners on base, and a championship game on the line? Context matters, and numbers collected in isolation do not always capture the realities of competition.
Perhaps the biggest mistake people make is assuming that because two things are related, one must cause the other (when in fact the variables correlate with one another). For example, many elite baseball players produce high exit velocities. That does not necessarily mean increasing exit velocity alone will create an elite hitter. It may simply be that better hitters naturally generate higher exit velocities because they possess superior mechanics, timing, vision, and decision-making. In this case, exit velocity may be a byproduct of great hitting rather than the cause of it. The same reasoning applies to pitching velocity, spin rate, sprint speed, and countless other measurable statistics.
There is also a psychological cost to consider. Young athletes today increasingly define themselves by their measurable statistics rather than their overall development. Instead of asking, “Am I becoming a better player?” they ask, “Did my exit velocity improve?” While analytics can provide valuable feedback, they should never become the sole measure of an athlete’s worth or potential.
Analytics have undoubtedly improved sports, and they are here to stay. However, numbers should inform coaching—not replace it. Great coaches understand that data represent one piece of a much larger puzzle that also includes confidence, decision-making, resilience, work ethic, leadership, adaptability, and countless other qualities that no radar gun or computer can fully measure. The challenge moving forward is not collecting more data. It is developing the wisdom to understand what those data truly mean—and just as importantly, what they do not.
Final thoughts
When using statistics, it is important that we understand exactly what we are measuring, as well as how useful the data will be in terms of predicting things in the future. Without understanding statistics, it is very easy to misunderstand data, including what is being measured, the importance of that information, and how useful it will be to apply the information in the future. Additionally, correlation (when two variables change or occur together) should not be confused with causation (when one variable directly causes change in another variable), yet this happens all the time with people not trained in data analysis. Use data as a part of evaluation and future planning, but make sure to employ a holistic approach that also includes examining confidence, motivation, and resiliency — constructs not easily measured in traditional testing.
drstankovich.com
