For seven years, when wintertime was on its last legs, Chris Curry would report to spring training, and during six seasons as a catcher for the Chicago Cubs’ farm system and one season with the Gary SouthShore RailCats of the independent leagues, Curry would meet the baseball men.
The men were supposed to know everything. Underneath their ubiquitous wraparound shades, their eyes sparkled with experience and knowledge reaped from decades in the game. It was with those eyes that they would decide everything from what pitch to call to the very trajectory of Curry’s playing career. The names on the back of their jerseys may as well have been “ORACLE.”
Chris Curry
“The coaches used to be former major leaguers,” said Curry, now in his 12th season as baseball coach at the University of Arkansas at Little Rock. “The only feedback you got was from their eyes and their information.”
Today’s athletic landscape could not be more different. Coaches from all sports and all levels are embracing vast amounts of analytical data in the name of player development and in-game decisions. As the games themselves change, better technology becoming available and artificial intelligence looming, so have the lives of data-minded coaches whose jobs are nothing like what has come before.
“I have always thought of myself as a pure baseball guy,” said DJ Baxendale, director of analytics for the University of Arkansas baseball program. “I kind of ran into analytics by default.”
DJ Baxendale
Baxendale is not exaggerating. He grew up in a baseball family with his father, Greg, heading up the Arkansas Express, one of the state’s premier showcase teams. Younger brother Blake was a Division 1 college player who went on to coach and would be inducted into the Rogers Heritage High School Athletic Hall of Fame. Between the two was DJ, who worked as a conference starter and high-leverage reliever for Dave Van Horn at the University of Arkansas in Fayetteville before embarking on a minor league path similar to Curry’s.
Baxendale joined the Minnesota Twins organization right when analytics were taking off in professional baseball, but he was not prepared for the sea change also taking place at the college and high school levels.
“I mean, it was light years,” Baxendale said.
Baxendale is not the only one carrying the odd sensation of working a job that, until recently, did not exist. Further south, in central Arkansas, Charles Gravett handles the data for Curry’s Trojans program. Gravett’s preparation for a three-game weekend series usually involves analyzing footage of the upcoming opponent’s starting pitchers plus heavy-use relievers, taking note of arm angle, times to the plate and pickoff moves. Then he moves on to the offensive side, logging everything from bunt situations to what teams like to do with runners on first and third.
Charles Gravett
“It’s like a matchup matrix,” said Gravett, a former pitcher at Xavier University in Cincinnati. “I upload all the players from our team and all the players from their team, and it’s a computer program that spits out what matchup is best for each guy.”
Gravett’s job also has an inherently physical aspect, as well. He trains his eyes on video to find kinetic tidbits such as the position of a pitcher’s landing leg, his spinal alignment during the windup and the amount of energy exerted with every pitch.
All Baxendale and Gravett have to do is make a couple keyboard strokes, hit “print” and hand it over to their players, right? Not exactly. One thing Baxendale quickly learned on the job is that handing over all of the data all of the time is tantamount to trying to get a sip of water from an open fire hydrant. Choosing what to communicate matters, but so does choosing how to communicate.
“It’s always going to look overwhelming,” Baxendale said. “A typical Trackman (video data feedback equipment) report after a nine-inning baseball game, for us, I think it’s roughly 320 pitches. That’s about a standard, so that’s 320 rows and approximately 120 columns — so it’s going to look like a lot.”
Baseball, with its focus on numbers and trends and percentages, may have been the first to embrace big data, but these days, everyone is invited to the party.
“I used artificial intelligence a lot last year,” said Matt Middleton, head football coach at NAIA Kansas Wesleyan University.
Matt Middleton
Before taking his first head job, Middleton coordinated record-setting passing offenses at the University of Arkansas at Monticello and University of Central Arkansas in Conway. Yet his biggest coaching moment came while working with Harding University in Searcy’s 2024 national championship-winning offense, a run-first and run-second scheme centered on traditional option concepts and straight-ahead rushing.
Middleton’s foray into data began with a chance phone call from Tom Woods. Woods was not a football coach. He was an aerospace engineer who saw opportunity in the lack of analytics in the game at the time. Middleton, at the time trying to jump-start his offense at UAM, saw opportunity in being able to think differently than his play-calling peers.
“We played Henderson State [in Arkadelphia], who we had no business being on the field with during that time,” Middleton said. “It was the middle of the game, I want to say Quarter 2. I made up my mind that I was going to do exactly what the data said, and I went full fledged, all in. We made a comeback and made it super close. We didn’t win the game, but I realized the importance of it.”
Not all analytics require high-tech and high-cost video analysis equipment. While at Harding, Middleton found himself hung up on the idea of efficiency and how it relates to field position. Starting on your own 10-yard line, Middleton learned, presents a 1-in-30 chance of scoring. Get to the 50-yard line, and the chance becomes 1 in 5.
After running the numbers of his own offense and the numbers from opposing defenses, Middleton shared his conclusions with Harding’s offensive players. To swing the odds in their favor, the Bisons needed to complete at least 65 percent of their passes for the rare times they did throw or, ideally, pick up at least 4 yards on the ground on first down. Therein lies the beauty of data, since it can reduce an overwhelming task such as winning a football game to a series of attainable goals like getting 4 yards on first down.
To structure the information, Middleton counts on assistant coach Ben Blahauvietz, a spreadsheet whiz with an undergraduate degree in sports analytics (yes, that is a thing now) from Syracuse University in New York. As the use of artificial intelligence continues to rise, though, less time is needed pecking away at a keyboard. In fact, less time may be needed for actual coaching.
“My assistant head coach and I ran all of our stuff through AI,” said Sheridan High School football coach Kevin Kelley. “We drew up some parameters and said, ‘Based on the information, what do you [AI] think?’ and it’s important when you talk to AI to say ‘What do you think?’ Otherwise, it’ll find something out there that somebody else has put on the web and kind of assign it that way.”
Kevin Kelley
Kelley would know. A member of the Arkansas Sports Hall of Fame and nine-time state champion at Pulaski Academy in Little Rock, Kelley is a data-driven early adopter best known for his predilection to never punt on fourth down. Beyond that, Kelley maintains a deep interest in the financial markets and also studies both human kinetics and pedagogical methods, so much so that he helped found and operate Little Rock daycare center Kid Champion.
Data is not one size fits all. Whereas a sizable chunk of baseball analytics is devoted to the science of movement and using software to evaluate swing paths, launch angles, windups, arm angles and more, football tends to lean more toward play tendencies, scenarios and formations. That, too, is changing.
“The NFL is already on that track,” Kelley said. “They’ve got a GPS sensor in the center of everybody’s shoulder pads and on the ball. They’re tracking.”
Examples of tracking movement data in football include obvious data such as speed and acceleration but also subtleties such as whether an offensive lineman continues to dip his hips and strike outward as the offense moves down the field.
As tends to be the case with AI, there remains a sense of foreboding. What is stopping a coach, for example, from standing on the sideline and simply asking ChatGPT for an effective third-and-seven play out of 11 personnel against an odd front and Cover 3 defense?
“Right off the bat, I’ll tell you that you’re not allowed to use AI on the sideline,” Kelley said. “You can use it before, you can use it after, but you can’t use it during.”
That in itself may not be so bad. High school coaches are largely dependent on the influx of students living nearby and whether or not they have been exposed to the game early, the problem being both Arkansas and the rest of the country are three quarts low on qualified youth coaches thanks to the time commitment and verbal — and sometimes physical — abuse from parents.
A well-meaning but inexperienced youth football coach can likely absorb the basics of what is being taught at the local high school and learn how to apply it, all through AI, but that still leaves it to the coach in question to distill what could be a towering amount of numbers into bite-sized, learnable sets of practical data.
Baxendale, Curry, Gravett, Middleton and Kelley all separately maintained that analytics are only as helpful as a coach’s ability to communicate and teach, and while data can now explain an awful lot in sports, the things that cannot be seen or quantified still matter.
“I will retire in 20 years continuing to believe in heart and toughness and guts,” Curry said.
And those baseball men Curry knew, the ones with the ever-active eyes? They still exist, but they are now armed with so much more.
“All good baseball coaches can still see,” Baxendale said. “They know what’s going on, but now I can back that up with cold, hard data.”
READ ALSO: Past Meets Future — Digs of the Deal: AP&L Building Loaded with Untapped Potential
