Story by
Take it from Jimmy Johnson: NFL coaches who rely too much on analytics play risky game
- Story by
- Take it from Jimmy Johnson: NFL coaches who rely too much on analytics play risky game
Let's call it what it is: Analytics represents the artificial intelligence of the NFL.
Go for two? Go for it on fourth-and-two?
Some coaches have to learn the hard way. The AI coming from the computer doesn't always know best.
Advertisement
"In some ways, coaches use analytics as a crutch," Jimmy Johnson, the Hall of Fame coach, told USA TODAY Sports. "There's more that goes into it than analytics give you."
Sure, next-level statistics can provide plenty of clues about tendencies and probabilities. But the numbers -- increasingly embedded into the decision-making that can determine games in the NFL -- shouldn't be confused as bottom-line answers that supplant common sense.
Take it from Johnson, the Fox Sports analyst who won two Super Bowls during the 1990s with the Dallas Cowboys . He was as puzzled as millions of others while watching the NFC championship game play out with a San Francisco 49ers victory.
The Detroit Lions had an opportunity to likely tie the game midway through the fourth quarter with a 48-yard field-goal attempt. Instead, Lions coach Dan Campbell opted to go for it on fourth-and-three from the 49ers' 30-yard line. Like a fourth-and-two gamble in the third quarter, the decision backfired, as a heavily pressured Jared Goff threw a pass that fell way short for Amon-Ra St. Brown .
Advertisement
"I love the job that Dan Campbell did this year," Johnson said. "But if you kick the field goal on that second fourth-down try, now San Francisco 's mentality is going to be different if it's a tied game, rather than being ahead. There's a lot of factors going in."
Campbell acknowledged afterward that he understood the criticism attached to his decision. Defenders of the coach pointed out that he merely acted as he had throughout the season, as the 40 fourth-down attempts by the Lions during the regular season were second-most in the league. They converted on 21 of those attempts, with the 52.5% success rate slightly better than the NFL average. There's something to be said for sticking with the aggressive approach that got them to the doorstep of the Super Bowl.
OPINION: It's so Detroit: Lions' first Super Bowl was in sight before meltdown for the ages
But with the season on the line, it was a bad move. The 49ers, having already rallied from a 17-point halftime deficit, took possession and drove 70 yards for a touchdown that extended their lead to 10 points.
Advertisement
In another sense, Campbell's gamble was a reflection of typical business in the NFL. Bolstered by analytics, teams are seemingly going for it on fourth downs and trying for two-point conversions more than ever -- and not just in crunch time.
The shortcomings shouldn't be ignored.
"It doesn't tell you the strength of your offense, the strength of the opponent," Johnson said. "It doesn't give you the weather. It doesn't give you a lot of factors. It doesn't give you momentum. It doesn't give you the risk and the reward. And that's what I always said. Yeah, it's a 70% positive to go for it on fourth down. But the risk is that you lose the game if you don't make it."
No, the analytics don't consider if your left guard has been getting destroyed throughout the game by the defensive tackle. It doesn't seem to compute that the No. 2 receiver has dropped a couple of passes.
Advertisement
"Exactly," Johnson agreed. "The analytics are probably good for Philadelphia going for it on fourth-and-one (with its signature "Tush Push" play). But it might not be quite as good for another team."
As the 49ers reveled in the victory that set up a Super Bowl 58 matchup against the Kansas City Chiefs on Feb. 11 in Las Vegas , San Francisco tight end George Kittle wondered, too, about the risky decisions.
"Why do analytics people say that momentum isn't a real thing?" Kittle pondered during his postgame news conference. "I had that conversation with Pat McAfee . That's just the biggest load of horse crap I've ever heard in my entire life."
Johnson also points to the human element. He said analytics don't take into account the potential psychological lift and psychological letdown attached to the decisions.
Advertisement
"You're going for it on fourth down, and if you don't make it, that opponent now has a tremendous psychological lift," Johnson said. "And the percentage of a team scoring after a turnover (on downs) is higher than getting the ball at the same spot through a kick. Because of the psychological lift."
During his heyday on the sidelines, Johnson probably gambled as much or more than any coach, especially when it came to fake punts, fake field goals and onside kicks. Yet he tried to account for failing to convert as part of the preparation.
"The one thing I would do is tell our team that we were going to do it, like on a Wednesday or Thursday," he recalled. "If you tell them, 'Hey, we're going to run this,' they're going to be ready for it. The other thing is, that prepared your defense. If it didn't work, they didn't have a psychological letdown. That's the thing analytics don't take into play."
Risky decisions, naturally, loom as a subplot for what might determine Super Bowl 58. Or maybe not. It's interesting to note that even though they have two of the most explosive and productive offenses in the league, the Chiefs and 49ers are on the conservative side when it comes to fourth-down attempts and two-point conversion tries.
Advertisement
Neither team converted a two-point attempt during the regular season; the Chiefs were 0-for-1 and the 49ers never went for two during the 2023 campaign. On fourth downs, the 49ers had the fewest attempts in the NFL with 13 (converting seven, for a 53.8 success rate). The Chiefs, meanwhile, converted on half of their 20 attempts.
Of course, with the Vince Lombardi Trophy on the line, such numbers may not matter. The flow of the game and situational factors, however, could hugely influence game-management decisions by Chiefs coach Andy Reid and his 49ers counterpart, Kyle Shanahan.
All things considered, does either coach have an advantage?
"Both of them are veteran coaches, been around a long time," Johnson said. "So they shouldn't have an advantage, one way or another."
Advertisement
Then again, Johnson wasn't too impressed with the manner in which San Francisco closed out the NFC title game. After taking possession with 56 seconds on the clock, the 49ers called consecutive runs to force Detroit to use its final two timeouts, then ended the 34-31 game with Brock Purdy taking a knee.
"I see so many screw-ups on time management," Johnson said. "I'm talking about veteran coaches. Even with Kyle the last game. Rather than run it when you're trying to kill the clock, then running it and then taking a knee, you're better off to take a knee, let them get a timeout, take another knee, let them run out of timeouts, then run it. Because if they fumble on the third run, they're out of time."
Now that's some old-school analytics at work.
"I see this time after time," Johnson added. "I don't understand it. So many times, coaches get caught up in the emotions and they really aren't thinking straight."
This article originally appeared on USA TODAY: NFL coaches who overuse analytics are playing a risky game
