Down 24-10 to the Washington Commanders in the fourth quarter on Sunday, the New York Giants scored a touchdown on a great 2-yard run by Daniel Jones with 9:25 left in the fourth quarter to close the gap to 24-16.
The question was then whether the Giants would go for 1 or 2. "Analytics" says to go for 2. It's counter-intuitive at first, but statistics bear it out. In any case, the Giants failed on the two-point conversion when Jones kept the ball and was tackled short of the goal line. It became something of a moot point when Washington added a field goal to make it 27-16, but when the Giants scored again, they had no choice but to go for 2 again to try to cut the margin to a field goal. They once again did not make it, sending only three receivers into the end zone and good coverage giving Jones no place to throw.
Daboll explained the reasoning in his press conference after the game:
As the tweet above suggests, there's math behind it. As the article by Seth Walder of ESPN referenced in the tweet shows, the odds favor you winning if you go for 2 after the first TD more than if you kick. Here's a summary of Walder's demonstration of that using the Jaguars as an example:
Assume that 94% of extra points (XP) and a bit less than half (48%) of 2-point conversion attempts are converted, as league-wide statistics over a long period have shown. Then the four possible outcomes and their probabilities are:
There's just one problem: 2-point conversion success is way down this year: Heading into this weekend, teams were only 18-for-58 on 2-point conversions, or 31%. Repeating Walder's math, that leads to the following (assuming the team scores 2 TDs in regulation):
Convert 2-pt conversion and XP: Probability = .31 x .94 = 29% chance to win in regulation
Convert 2-pt conversion and miss XP: Probability = .31 x .06 = 2% chance to go to OT
Fail on 1st 2-pt conversion, succeed on 2nd: Probability = .69 x .31 = 21% chance to go to OT
Fail on both 2-pt attempts: Probability = .69 x .69 = 48% chance to lose in regulation
With a 2% + 21% = 23% chance to go to OT, and assuming a 50-50 chance to win in OT (i.e., 11.5%), that makes the odds of winning 29% + 11.5% = 40.5%. That's much lower than 50-50 and much lower than the 59% odds of winning in Walder's example.
Why are teams having so much trouble converting 2-pt attempts this season? No one seems to have an explanation. Perhaps analytics departments have done so much research by now that answers have been found to most of the limited options that offenses have in tight quarters. It could instead just be a small sample, but there's now about half a season in the books.
For the Giants specifically, with their two misses today, they are now 0-for-6 on the season in 2-pt attempts. You don't need to do much math to project what that implies for their chances of success if they go for two late in the game. The Giants don't have a road-grading offensive line that can usually make 2 yards on the ground near the goal line like the Eagles' tush-pushers. They don't really have a big X receiver that they play much who can just run a fade and grab a 50-50 ball in the corner of the end zone. There are only so many times they can go to the well of Jones zone-read keepers. That leaves them with few options for success.
Maybe next time, Daboll should just kick the extra points, get to OT, and hope he gets the chance to turn to his latest placekicker and say, "Hey Jude, don't let me down."