GAINESVILLE, Fla. — Just steps from the second-story terrace that overlooks Florida’s practice court, Jonathan Safir plops down at his office desk and takes a timeout. The director of basketball strategy and analytics has been on the job for a hectic and pivotal two months, during which the Gators’ incoming staff gave the roster a bumper-to-bumper appraisal, prioritized which players to keep and decided which ones to pursue from the portal and the high school ranks.
The metrics that informed those decisions are Safir’s domain. At age 29, the former backup guard at Division III Vassar is adept at taking mounds of situational data and player tendencies and distilling them into “bite-sized chunks,” a skill that made him a confidante of head coach Todd Golden. Introduced years ago at a Jewish coaches breakfast, they synced up across four seasons at San Francisco, with Safir serving three of those as director of basketball operations before Golden promoted him to an assistant coaching role for last season’s NCAA Tournament team. When Florida made the bold move of hiring the 36-year-old Golden on March 18 — a deal finalized at 30,000 feet as the Dons flew home from Indianapolis after a first-round loss — Safir knew he’d be part of the package, even if it meant returning to an off-court position. Florida announced Safir’s hiring 11 days later.
“Todd and I have a lot of common traits — the way we think about things statistically and analytically, the way our brains work,” Safir said. “He has been like a big brother to me.”
A big brother who values deep-dive statistical analysis, invests in it and trusts the percentages, even when they sometimes fly in the face of conventional assumptions.
Take Jan. 25, 2020, a game the analytics crowd loves to reference: The Dons led BYU 79-77 with 22 seconds left. Cougars ball. Most coaches in that situation put their best defensive group on the floor and play out the possession. But in the timeout huddle, Safir reminded Golden that circumstances were right for fouling because:
• BYU led the nation in 3-point shooting that season (42 percent) and had made 14 of 25 that night. An offense so equipped at generating open 3s likely would be gunning for the lead.
• The Cougars’ All-WCC forward Yoeli Childs, while a dynamic inside-outside scorer, was only a 54 percent free-throw shooter that season. Childs also was BYU’s best rebounder, so putting him on the line diminished the Cougars’ chance at a putback.
• BYU was in the one-and-one.
Golden agreed with Safir, and with 17.6 seconds left, San Francisco reserve big man Taavi Jurkatamm intentionally fouled Childs. When Childs missed the front end, the Dons secured the rebound and went on to win 83-82.
A similar scenario unfolded in February when the Dons led Portland 72-70 with only six seconds remaining.
• Same two-point cushion as against BYU, only Portland wasn’t as lethal from 3-point range (35 percent, ranking 102nd nationally) and the Pilots had far less time to execute a half-court set.
• Portland was in the one-and-one, but the Pilots didn’t have a foul-shooting liability on the floor. The worst was Chris Austin, at 76 percent.
So the circumstances weren’t as advantageous as they were against BYU. In fact, the math was negligible as to whether fouling optimized San Francisco’s win probability. Other factors came into play, however, as the staff decided to foul Austin.
“If Portland had walked us off with a 3, in a game where we were 15-point favorites, a game we thought we needed to make the NCAA Tournament, I’m not sure we could’ve lived with ourselves,” Safir said. “So by fouling, the worst-case scenario was that Portland would make both of its free throws, then stop us from scoring at the end of regulation, and then they’d have to beat us in overtime when we had Jamaree Bouyea and Yauhen Massalski being unstoppable on offense. If Portland was able to do all that, we’d just have to tip our cap. Better to go that route, than giving them the chance to walk us off with a 3.”
USF’s foul paid off. Austin made the first free throw, clanked the second, and the Dons prevailed 74-71.
If the Portland or BYU games had turned out differently, Golden would’ve faced the kind of uncomfortable postgame questions that dog some coaches. Fearing the uncomfortable isn’t part of Golden’s DNA.
“Todd’s willing to take risks, to throw stuff out there, and that’s something I admire,” Safir said. “He’s willing to go for it.”
Some of that willingness stemmed from a 2019 Final Four meeting with Ken Pomeroy, the sage of basketball analytics. According to Safir, the engineer behind the advanced stats KenPom website typically finds no value in pitching end-of-half probabilities to coaches because they’re too risk-averse, too afraid of against-the-grain decisions backfiring. “But when he broke down the math to us, Todd said it made sense, and Ken could see we were actually going to implement it,” Safir said.
A year later, when COVID canceled March Madness, Safir approached Pomeroy about using the downtime to conduct deep-dive research. They wound up reviewing tens of thousands of previous games to better analyze the probabilities behind end-of-game fouling. They also computed appropriate circumstances for fouling with one-point leads and even with tied scores.
Golden described Safir’s work as the product of “a great basketball mind” and predicted “it probably won’t be very long before NBA teams come calling for his services in their respective front offices.”
The Gators’ online roster lists Trey Bonham at 6-foot. For a program that craves accurate data, that is, ahem, an imprecise number. The transfer guard from VMI met with reporters last Tuesday and smiled when asked if he’s even 5-11.
“We’ll say that,” he said. “Yeah, I’ll take it.”
Among a sought-after group of transfers headlined by St. Bonaventure point guard Kyle Lofton, Belmont wing Will Richard and LSU forward Alex Fudge, all of whom had their choice of power conference schools, Bonham is the pint-sized outlier who Safir calls “an awesome analytical find.”
He entered the portal April 4 and initially scheduled visits to Murray State and Appalachian State. Then Florida became involved and the landscape changed overnight. Programs including Purdue, Texas A&M, Iowa and Missouri followed suit, the sort of recruiting groupthink common when teams like a fringe player but don’t want to be the first to stick their neck out.
By the time those schools contacted Bonham, the Gators essentially had locked him up. On his Easter weekend recruiting visit to Gainesville, the staff dissected his strengths in such nuanced detail that Bonham felt as though he was being educated on himself. “They walked me through every metric,” he said. “Showed me that my pick-and-roll, finishing and rebounding for my size was off the charts. I really had no clue.”
For a kid who attended high school in Mobile, Ala., and dreamed of playing in the SEC, Florida’s pitch sounded perfect. “There wasn’t much they had to do to get me,” Bonham admitted. “I was already in as soon as they called.”
But why did the Gators call? What made them the first major-conference program to pursue Bonham?
Safir noticed Bonham’s advanced metrics at VMI were comparable to Wendell Green’s at Eastern Kentucky. Last season Green transferred to Auburn and became the sixth man on an SEC championship team, averaging 12 points per game with the nation’s 17th-highest assist rate.
What the staff saw from Bonham on film validated the data.
“The portal has been a great tool for finding undervalued players, and Trey’s a great example,” Safir said. “Everybody looks at his height, but the advanced numbers show he’s good finishing at the rim. And he was incredibly efficient aside from his wrist injury.”
Bonham played through the wrist sprain from late November until mid-January, sometimes avoiding using his dominant right hand. His sophomore season averages were 13.6 points. 4.0 assists, 4.4 rebounds and 35 percent shooting from 3, leading to third-team All-SoCon honors.
Over the final 16 games, with the wrist pain dissipated, his production improved to 17.8 points, 4.5 assists, 5.3 rebounds and 39 percent from deep.
After starting all 29 games at point guard for VMI last season, Bonham gives Florida a capable sidekick to the graduate transfer Lofton. The multiyear All-Atlantic 10 standout was hardly a sleeper find — choosing between Arizona, Purdue and Rutgers until a visit to Florida altered his course.
“We knew, and the whole world knew how good a player Kyle Lofton was,” Safir said. “But we still showed him analytics about why he could succeed here.” Things like how the staff charts 50 “hustle stats” during workouts and games — deflections, steals, bobbles, missed blockouts, defensive returns, virtual assists, etc. — in an effort to incentivize players toward becoming more complete.
“Coming here is like next-level,” Lofton said.
The goal under Golden is to ramp up the pace so that Lofton can be just as impactful playing 30 minutes per game as he was averaging 38-plus over four seasons at St. Bonaventure, where depth was lacking. His 5.9 assists last season ranked ninth nationally and would’ve been higher had the Bonnies been better than a 222nd-ranked 3-point team percentage, which sounds mediocre until it’s compared against Florida’s 321st ranking (30 percent). With the Gators restocking their wings, and Golden’s system emphasizing more spacing and 3s, the aim is that Lofton should be surrounded by better shooters who make him more productive on drive-and-kick situations.
The analytics on a player can be eye-opening. And sometimes, Safir admitted, the eye test is equally convincing. The 6-8 Fudge being a top-60 recruit a year ago and averaging 14 minutes for an LSU team that reached No. 12 in the AP poll made him a hot commodity in the portal. Richard’s athleticism, complete with an 83-inch wingspan on a 6-6 body, had big-time schools in pursuit and NBA scouts putting him on their radar for down-the-line drafts.
“The data can inform our decisions, then it becomes an amalgamation of roster fit, keeping everyone happy and trying to team-build,” Safir said. “Especially during the spring months, where you’re finding certain guys to buy in on certain roles on certain terms. The players we had at Florida were incredibly important to us, and we were trying to find pieces who best complemented them.”
Retaining key players from Mike White’s NIT team was an integral part of recruiting. Colin Castleton landed an NIL deal, sure, but he’s spending a fifth year in college because Golden’s staff dug into ways he could become more efficient. Kowacie Reeves, a freshman wing who went from looking lost to playing with authority late last season, spent a few days in the portal but was coaxed out, energized by the tempo and spacing of Golden’s new system.
The staff also trusted that Myreon Jones is the player who shot 38 percent from 3 during his career at Penn State, not the guard who dipped to 32 percent last season at Florida and posted a career-high turnover rate.
“Myreon Jones was one of the best guards in the portal last year, but for whatever reason, he didn’t have the best year statistically at Florida,” Safir said. “If we can bring back ‘Penn State Myreon’, that’s a very good player, a potential all-league player. That played a big role in our desire to keep Myreon and we’re fortunate he wants to still be here, because he’s graduating and could’ve gone anywhere.”
The most curious analytical case study involves 6-11 backup center Jason Jitoboh. Receiving only spot usage over his first two college seasons, he played double-digit minutes as a junior and made four starts when Castleton was sidelined by a shoulder injury. Safir labeled Jitoboh a “difference-maker,” a description that goes beyond his individual averages of 4.1 points and 2.4 rebounds.
“Looking at Jason, his on-off splits are off the charts. With him on the court last year, Florida played ridiculous,” Safir said. “The team’s offensive and defensive efficiency numbers were strong with him on the court, and we’re trying to dig deeper and see exactly why. We’ll probably get a better understanding once we get more practice data here with us coaching.
“The macro level of deciding to invest in him, of letting him know we were pouring into him and that we wanted him to be around, a lot of it stems from his offensive/deficiency efficiency.”
Recovering from eye surgeries, Jitoboh has been cleared for non-contact drills. His presence — and whatever ancillary benefits it brings — is more important after Morehead State transfer forward Johni Broome picked Auburn over the Gators. Florida’s pursuit of another big ended when 6-11 transfer Mouhamed Gueye opted to return to Washington State, the program headed by Kyle Smith, a mentor to Golden and Safir. It’s unlikely at this late stage the Gators can find a game-ready big for that 13th scholarship, meaning it likely will go to a developmental player. That creates more responsibility and opportunity for Jitoboh.
“He has proven he can do it in smaller increments,” Safir said, “and now let’s see if we can push it even more.”
Safir began as an economics major at Vassar before switching to American studies. He didn’t take many college math classes and admits that when concepts become abstract, as with proofs and theorems, “I lose it a little bit.” On the applicable side of mathematics, however, he “loved stats and numbers” and that’s where basketball has become a science.
“I thought I was a smart player in college, but I think about the game much more cerebrally now,” he said, crediting his exposure to “innovative thinkers” like Golden and Smith.
“Now, the analytics factor into all of our decision-making, whether it’s scheduling, recruiting, on-court strategy or player development. We use it for everything. It’s one of the backbones of the program and how we operate.”
The Gators began summer workouts last week, with one session focusing on the installation of “penetration rules” — showing players where they need to move when a ballhandler drives inside the free-throw line. The oversimplified version of analytics is that teams should maximize offensive efficiency by shooting only 3s and layups. Safir said that’s not the case under Golden.
“We don’t just say ‘All midrange shots are bad, so don’t shoot them.’ Myreon Jones might want to work on his midrange, just like Jamaree Bouyea at San Francisco wanted to work on his midrange. We encourage them to work on that. It’s a great skill to have and it translates to the NBA,” Safir said. “Plus, it’s an important end-of-shot-clock skill, because there are times when ballhandlers are going to the basket and everyone knows they’re coming.”
In those late-clock situations, where defenses are prone to constrict, a midrange pull-up shooter has, say, a 45 percent chance of making it. In addition, there’s maybe a 30 to 35 percent chance of getting the offensive rebound depending on where the bigs are located.
“So those midranges become good shots because in end-of-clock situations, it’s a lot better to get a shot on the rim than to have a turnover,” Safir said. “You just want it to come from a good shooter who’s in rhythm.”
Insights from all this data, while valuable to coaches, ultimately must be embraced and trusted by players, who often hear competing voices outside the program. Safir lauds Golden for striking a tone that integrates analytics with a personal touch.
“Todd has a special talent for parsing it down and getting the players to buy into what the numbers are saying,” Safir said. “They see the benefits, they see how they can get better, and they see that he cares about them because at the end of the day, we want basketball players, not robots.”
(Top photo of Jonathan Safir: G. Allan Taylor / The Athletic)