Sports have always been obsessed with trying to predict the next superstar. Scouts have spent generations trying to determine which athletes will become stars, and which ones will fail to make the cut. Previously, scouts measured height, weight, speed, strength, arm velocity, vertical jump, and countless other physical characteristics. Next, scouts began looking at advanced statistics beyond simple measurables. Today, scouts can analyze launch angle, spin rate, exit velocity, acceleration, movement patterns, workload, and thousands of other data points that previous generations of scouts could only imagine. All of this begs the question: Will we eventually be able to predict an athlete’s psychological performance with the same precision we attempt to predict physical performance?
Scouting the perfect athlete
When assessing athletes many analyses take place, but perhaps there is one big question that supersedes the others: Will this athlete excel in the clutch?
Some athletes thrive when the pressure increases. Conversely, other athletes perform brilliantly when little is at stake, but struggle terribly when the game is on the line. Sport psychologists have spent decades examining anxiety, confidence, attention, emotional regulation, personality, coping skills, and responses to pressure. So far, no reliable test can tell us who will thrive under pressure and who will struggle. But what if someday a biology test could yield those results?
As artificial intelligence continues to develop, we may soon have new ways to assess the relationship between personality and sport performance. For example, imagine the scouting combine of the future. An athlete enters a testing room and provides a saliva sample. That sample is combined with genetic information, hormone levels, heart-rate variability, neurological testing, personality assessments, reaction-time measurements, and years of performance data analyzed by artificial intelligence. Minutes later, a computer generates a new statistic called a Pressure Performance Score. Clutch potential: 38th percentile. The scout looks at the number and immediately crosses the player’s name off the list based on his low score.
As of today, we are not able to assess athletes using a pressure scale like just described. Genetic testing cannot reliably tell us who will become an elite athlete, much less who will deliver the game-winning hit in the ninth inning. Human performance is far too complicated for that kind of specificity today, but technology rarely stops because something is complicated.
Professional sports organizations already use sophisticated psychological assessments, biometric measurements, wearable technology, medical imaging, and enormous databases to evaluate athletes. As artificial intelligence becomes better at finding relationships among seemingly unrelated variables, it isn’t difficult to imagine teams searching for biological markers associated with performance under pressure.
Suppose researchers eventually develop a biological and psychological profile that predicts performance under pressure at a highly accurate rate. Would teams use it? After investing hundreds of millions of dollars in a pro sport franchise, it’s hard to imagine teams not going all-in with new testing that reveals the biology behind great clutch athletes.
While finding athletes that stack up biologically is great, what happens to the athletes the algorithm gets wrong? An otherwise talented athlete might lose an opportunity, not because he failed under pressure, but because a computer predicted that someday he might.
There is also the possibility of a self-fulfilling prophecy. Imagine telling a young athlete that his biological profile indicates he performs poorly under pressure. What happens the next time he steps to the plate with two outs in the ninth inning? In all likelihood based on his test results, he will perform poorly because he expected to perform poorly.
While it is a worthwhile pursuit to find biomarkers that yield important information relating to future sport (or life) success, it is also important that we understand the effects of self-fulfilling prophecies, as well as confidence, attention, and motivation. In other words, mentally tough athletes may outperform what their profiles predict, while athletes with less mental toughness may struggle under pressure despite profiles suggesting they should succeed.
Sports analytics began by helping us understand performance. Eventually, they may attempt something much more ambitious: predicting human potential before that potential has an opportunity to reveal itself. If science someday claims it can tell us who will become clutch and who won’t, should we use that information to decide who gets the opportunity to find out?
Final thoughts
Perhaps the greatest danger isn’t that technology will someday become good enough to predict human behavior, but that we will become too confident in believing that it can. Athletes are not static collections of genes, biomarkers, personality scores, and statistics. Analytics can undoubtedly help us understand athletes better, but there are important aspects of the psyche that are incredibly difficult to measure, including focus, motivation, and the will to win. In the future, there may be a point where prediction begins replacing opportunity. If a number someday tells us that an athlete isn’t built to be clutch, will we still give that athlete the chance to prove the number wrong?
drstankovich.com
