How NFL coaches try (and often fail) to master the end-of-game chaos

“Everybody’s effort is on your shoulders,” Harbaugh said. “It’s not so much about what the smart decision or what the numbers decision is. It’s what’s going to be successful. It’s what’s going to help you win. That’s never certain.”

Maske’s NFL reset: Inside the league’s scoring decrease

In a league designed for parity, any decision made in a split-second from the sideline could determine the result of 60 minutes of physical combat on the field — and game management has grown more crucial this season than any in recent memory. Twenty-five games have been decided by six points or less, and 18 of those had a margin of three or less, both the highest in NFL history through three weeks.

During the opening week, Denver Broncos Coach Nathaniel Hackett’s doomed choice to engineer a 64-yard field goal in the final seconds of an anticipated “Monday Night Football” showdown created instant calamity and prompted a deluge of criticism. The 1-2 Detroit Lions’ record might be transposed had Coach Dan Campbell gone for a fourth and 4 last Sunday rather than try a 54-yard field goal with 1:14 left, a decision he said he “hated” in the immediate aftermath. Following kicker Austin Seibert’s miss, the Minnesota Vikings stormed for a go-ahead touchdown and victory.

“If you look at the NFL over the last number of years, two-thirds of the games are within one score in the fourth quarter,” said former Dallas Cowboys Coach Jason Garrett, now a “Sunday Night Football” analyst. “The teams that go to the playoffs, they play the same number of those games. They just win more of them.”

In the abstract, any teenager who’s played enough Madden can optimize timeouts and knows when to go for it on fourth down. In practice, amid the chaotic duress of an NFL sideline, simple choices become complicated. At pivotal moments, without the appropriate preparation or required nerve, an NFL head coach can turn into a stressed dad trying to tie his crying toddler’s shoe while the toast is burning and he’s about to be late for work.

“You got to be calm in a s— storm,” said former New York Jets Coach Rex Ryan, now an ESPN analyst. “It’s not on anybody else. It’s on you. You got to be calm, and you got to be thinking ahead of the game. It’s like playing chess.”

Head coaches operate with more information than ever before, as publicly available statistical analysis has drastically shifted how they operate. Coaches must synthesize all the insights gleaned from beefed-up analytics departments, and though the spread of information has allowed for smarter decisions, it’s also made those decisions more scrutinized and, at times, difficult to act upon.

If coaches get it wrong, a sea of analytically driven websites, Twitter accounts and podcasts are lying in wait to amplify their mistakes. But that understanding of analytics goes both ways. When Washington Commanders Coach Rivera went for two after a touchdown brought his team to within eight two weeks ago, many fans nodded with approval at the once-unconventional, now-routine call meant to boost a team’s win probability by avoiding overtime (even if it somehow still mystifies broadcasters).

Asked Wednesday how he approaches endgame scenarios, Harbaugh at first provided a deadpan look. “Flip a coin, usually,” he said, breaking into a grin.

Harbaugh has developed a progressive reputation among analytics devotees. In 2019, he pushed other coaches to eschew punts because of his extreme aggression — and frequent success — going for it on fourth down. Still, Harbaugh navigates the endgame far more through feel for the moment than dogmatic faith in probability.

“I know the analytics people would like to tell you they can measure everything, but they can’t measure exactly how the game is going, or the circumstances, or the emotions, or the feel of the stadium or just luck,” Harbaugh said. “In the end, you just try to make the decision that gives yourself the best chance to win. I always go back to this: Do I think it’s going to work?”

Even analytically inclined coaches must operate without black-and-white assurance. The NFL bans all devices that can transmit data — cellphones, FitBits, Apple Watches and more — on the sideline and in coaches’ booths. Coaches can study two-point conversion charts and win probability models during the week, but neither they nor the staffers they’re in contact with can consult the real-time models fans peruse online.

“They’re flying a little bit blind relative to what those of us on the outside are doing when second-guessing,” football analytics expert and the Athletic contributor Ben Baldwin said.

The prohibitions make preparation paramount. Twice a week, Rivera meets with Commanders analyst Doug Drewry to review situations that arose in the previous game: timeout usage, close fourth down decisions, two-point conversion opportunities. One morning every week, Rivera also meets with three members of his coaching staff to discuss the same topics and look ahead at how they could handle those moments on game day.

“There is a lot of that you try to retain and remember,” Rivera said. “When all heck is breaking loose — you got a couple players missing, momentum swinging, all that kind of stuff — you’ve got to do it.”

During games, Rivera can press an intercom button on his headset to access those coaches and Drewry for input on whether to punt or go for it, to kick or go for two, to call timeout or save it. (With a tone that suggested he’s delved for an answer, Rivera noted that analytics have provided no insight on the best time to call for a fake punt or fake field goal.)

Just as there is a difference between practicing and performing for players, there is a difference between planning and action for coaches. Some remain poised under pressure. Some choke.

When Baldwin studied coaches’ worst decisions in terms of win probability, a disproportionate number occurred in playoff games. Coaches are more likely to make a poor decision when it’s close and late, Baldwin said, and to be conservative when a bold-yet-optimal decision could lead to criticism.

“I’ve seen guys that couldn’t call a defense when a team went no-huddle,” Ryan said. “It’s the damndest thing you’ve ever seen. There are some guys that don’t do as well.”

NFL best bets for Week 4

When he coached the Cowboys, Garrett would visit with Navy SEALs in an effort to learn how to make crisis feel routine. Garrett constantly tested his players, his coaching staff and himself with different situations. During walk-throughs, he went through a nine-play script with different scenarios and variables — running the field goal team on the field with the clock ticking down, but with one of the members out with a pretend injury. The more he rehearsed, the more he could eliminate panic from any contingency.

“You’ll never duplicate each situation perfectly,” Garrett said. “But hopefully you’ve done it enough so those situations feel quiet and slow.”

Garrett created a situational football handbook each offseason, revising the last year’s edition based on new insights or changing personnel. He included best practices for two-minute drills when trailing and four-minute drills when ahead.

He also called his own offensive plays, which forced him to delegate. He put two staffers in charge of assisting him with challenges, one for offense and one for defense. His special teams coach walked through fourth-down calls with him. During games, he encouraged constant chatter on his headset. An assistant might point out he needed to prepare his two-point conversion play, and he might point out that the hands team needed to be ready — and not to forget that a key player was injured and his replacement needed coaching up.

But even the best-laid plans can implode. With the Jets, Ryan made it a priority not to leave Tom Brady time on the clock. In one close game against the New England Patriots, Ryan told his offensive players not to call timeouts, to let him manage the clock. “Sure as s—, boom, my quarterback calls timeout,” Ryan said. “I’m like, ‘What the hell are you doing?’ ”

The Jets kicked a field goal, and Brady answered with a winning touchdown in the waning seconds. Ryan’s ensuring cursing spree prompted a $50,000 fine.

Archives: As Green Bay’s offensive coordinator, Hackett was ‘not like the rest of them’

Ryan never relinquished calling his own defensive plays. “Why the hell would I turn it over when I got the job because of this?” he said. Hackett took the same approach, continuing to call offensive plays, as he had done as the Green Bay Packers’ offensive coordinator.

For an inexperienced head coach, the attempt at multitasking can lead to disaster.

Last week, in response to the failed 64-yard field goal attempt in Week 1 and his clock-management fiasco in Week 2, Hackett hired former Baltimore Ravens special teams coordinator Jerry Rosburg to assist with in-game decisions. On Sunday night, Hackett said, he learned on Rosburg’s counsel when he punted on fourth and inches from his team’s 34-yard line. Hackett initially had wanted to go for it, but Rosburg talked him into a punt, arguing that the defense was playing well and would win back possession with less risk than a fourth-down try. It unfolded ideally for the Broncos: They forced a three-and-out and went on to score the game-winning touchdown.

“All coaches, we want to grow,” Hackett told reporters. “[Rosburg is] a guy with a ton of experience that’s run game management before. And after talking with him, he was all for it, excited about it.’’

The funny thing about Rosburg’s advice: In a vacuum, he was wrong. According to the highly regarded Fourth Down Bot, which Baldwin designed, punting reduced the Broncos’ odds of winning by 4.6 percent compared to going for it, making it a “strong” recommendation to go.

Did Rosburg read the game perfectly, separating variables that a probability model built on reams of historical data couldn’t account for? Or did the Broncos’ players make their coaches lucky by overcoming their mistake?

“I would more lean on the latter,” Baldwin said. “… Your defense being good doesn’t only help you if punt the ball. If you trust your defense, you should trust your defense to prevent the opponent from scoring after you go for it. It’s not like the defense is only relevant in the situation where you’re punting.”

The Broncos’ victory underscored the vagaries of coaching. Coaches can make the wrong call and win or the right call and lose. They can botch the clock only for their players to save them, or they can plan for one precise moment for years only for a player’s error to cost them a victory. Their choices can mean everything, or they can mean nothing.

Wednesday afternoon, at the Ravens’ facility, after Harbaugh explained the process he’s honed for nearly 15 years, he paused and seemed to realize he may not have been completely joking after all.

“It’s almost like flipping a coin,” he said.