Which teams got it right on fourth-down and 2-point conversion calls in Week 2 of the 2021 NFL season -- and which teams got it wrong? As they did in Week 1, the Next Gen Stats analytics team uses the Next Gen Stats Decision Guide powered by AWS to break down the numbers behind the decisions that shaped the game:
Baltimore clinches victory on fourth-and-1
FOURTH QUARTER: With 1:05 remaining and the Ravens (leading 36-35) facing a fourth-and-1 from their own 43-yard line, Baltimore's Lamar Jackson rushes for 2 yards.
According to the Next Gen Stats Decision Guide, the Ravens' decision to go for it here was unequivocally the optimal call, by a difference of 24 percentage points in win-probability value. Converting meant the Ravens would win. If they hadn't converted? Their win probability would have dropped to 33 percent. And if they'd punted the ball back to the Chiefs, the Ravens' chance of winning would have been roughly 58 percent, assuming Patrick Mahomes and Co. would have gotten the ball back with under one minute left from deep inside their own territory.
But that only explains the outcomes. We also have to take into account how likely it was that Jackson and the Baltimore offense were going to convert on fourth-and-1 (the distance was 1.1 yards, to be exact, according to our ball-tracking data). In that situation, against the Chiefs' defense, the Ravens had a 75 percent chance of gaining enough yardage for a first down. Taking this into account, the expected win probability of going for it was 82 percent. The difference between 82 percent (the value of going for it) and 58 percent (the value of punting) represents the difference in win-probability value of the two decisions. In the end, the numbers say it was a no-brainer. Jackson's run kept the ball out of Mahomes' hands.
Ravens coach John Harbaugh is no stranger to going for it on fourth down when the analytics suggest doing so. Before the season, we ranked returning active head coaches by how often their decisions matched the numbers. Harbaugh ranked second, behind only the rate produced by Kevin Stefanski in his first season as Browns head coach -- which is less surprising the more you learn about how much the Ravens and Browns have built up their analytics departments.
Joe Judge held back by conservative approach?
The Giants' close defeat on Thursday Night Football was marked by untimely penalties, questionable game management -- and Judge's conservative decision-making on fourth downs. The Giants' head coach elected to kick a field goal four times when the Next Gen Stats Decision Guide showed value in going for a conversion on fourth down, and while Graham Gano made all four kicks, New York still lost the game. Notably, those decisions cost the Giants a net total of 4.9 percentage points in win probability.
FOURTH QUARTER: With 13:39 remaining and the Giants (winning 20-17) facing fourth-and-3 from the Washington 34-yard line, New York's Graham Gano attempts and makes a 52-yard field goal.
According to the Next Gen Stats Decision Guide, Judge's decision here cost the Giants 2.6 percentage points in net win probability. The odds of the Giants converting if they went for it in this situation were 46 percent, just a little smaller than the odds of making the 52-yard field goal (53 percent). With the made field goal, the Giants' win probability was 70 percent -- but a first-down conversion would have set them up with a 75 percent win probability.
FOURTH QUARTER: With 4:55 remaining and the Giants (winning 23-20) facing fourth-and-4 from the Washington 37-yard line, Gano attempts and makes a 55-yard field goal.
After trading field goals, Judge once again chose the more conservative option on fourth down. This decision was more of a toss-up (going for it carried a gain in win probability of 1.1 percentage points), accounting for Gano's chances of making a 55-yard field goal (44 percent) compared to a 41 percent chance of converting on fourth-and-4. Gano ended up making the field goal, increasing the Giants' probability of winning to 80 percent. However, if the Giants had gone for it and converted? The team's win probability would have jumped to 84 percent.
In a contest where win probability crossed the 50-percent mark 10 different times, controlling the clock and maintaining possession is pivotal when you have a lead. Washington ended up scoring a touchdown and taking a 27-26 lead in only 17 seconds of game clock following this made field goal.
Down 8, Philly leans on a familiar strategy
Even in a losing effort, the Eagles' staff proved why they've been among the savviest organizations when it comes to fourth-down and 2-point conversion decisions. The approach Philly took here -- going for 2 when attempting to come back from a 14-point deficit -- was, after all, pioneered by ex-Eagles coach Doug Pederson only three seasons ago.
FOURTH QUARTER: With 4:06 remaining, the Eagles (trailing 17-9) successfully convert a 2-point try on a Kenneth Gainwell run.
Under Pederson, the Eagles accounted for three of 11 "go for two when down eight" decisions made between the 2017 and 2020 seasons. Pederson's successor, Nick Sirianni, continued the trend here. And while the benefit was marginal, according to the Next Gen Stats Decision Guide, the decision was nonetheless the overall right call, given that a successful 2-point conversion would have brought their win probability to 9 percent, compared to just 7 percent after a successful extra-point kick.
Sirianni and the Eagles never got the ball back, as the 49ers converted a first down in three plays to set up three consecutive kneeldowns by Jimmy Garoppolo. But as many have explained before, the math behind the decision gave the Eagles the best opportunity to win. Having completed the 2-point conversion (46 percent probability), Philly would have been able to take the lead with an extra-point kick, should the team have gotten the ball back and scored a touchdown. Failing to convert the 2-point try would still have allowed for an opportunity to tie the game with another touchdown and a second 2-point attempt.
The later it is in a game, the more important it is to maximize each scoring opportunity. No matter what, facing a 14-point deficit, the Eagles needed to score two touchdowns to have a chance -- and kicking for the extra point after both hypothetical scores would have left them tied at best, meaning they'd still have roughly a 50 percent chance (or lower, in this particular matchup) of winning in overtime.
Brandon Staley gets bold in his own territory
THIRD QUARTER: With 13:41 remaining and the Chargers (trailing 14-11) facing a fourth-and-inches from their own 34-yard line, Los Angeles's Justin Herbert picks up 3 yards off a QB sneak.
Many coaches might choose to avoid risk in their own territory, but the Chargers' first-year head man stayed true to his desire to be at the "forefront" of the analytics movement and made the right call here. The Next Gen Stats Decision Guide strongly recommended going for it, with 6.2 percent difference in expected win probability. Thanks to ball-tracking data, we can precisely measure the actual yards to go -- and the Chargers had an 81 percent chance of converting the first down, given the inches-to-go situation. The QB sneak generated an 11 percent increase in win probability and extended a drive that would end with a game-tying field goal. The Chargers' comeback fell short, but Staley's aggressiveness helped significantly.
Kyler Murray airs it out
FOURTH QUARTER: With 6:12 remaining and the Cardinals (trailing 31-33) facing a fourth-and-5 from the Vikings' 41-yard line, Arizona's Murray connects with Christian Kirk on a 35-yard pass.
This play featured the perfect combination of data-driven decision-making and traditional scouting and coaching. From a data perspective, the Next Gen Stats Decision Guide recommended the Cardinals "go for it" by an advantage of 3.8 percent in expected win value. The Cardinals had a 44 percent chance of converting the first down, and they would have a 43 percent win probability if they just gained the 5 yards required. But the Vikings showed a Cover 0 look pre-snap, and Murray took advantage, using his speed to drift back away from the seven-man blitz and launch a bomb to Kirk.
Kingsbury worked with Murray on this strategy to buy time against Cover 0 during training camp, and Murray executed it perfectly with the game on the line. The deep completion gained 30 yards more than necessary for the first down, resulting in an increase in win probability of 30 percent. While the Cardinals would be forced to settle for a field goal (and a one-point lead) with time still left on the clock, the Vikings wound up missing on their own 37-yard field-goal try at the end of the game, and the Cards improved to 2-0.