This season of the NFL could be the strangest since World War II, when NFL rosters were so heavily depleted that in 1943 the Pittsburgh Steelers and Philadelphia Eagles played as a single squad, the “Pittsburgh-Philadelphia Steagles.” The COVID-19 pandemic may push back the start of the season or cancel it altogether — and even when NFL football returns, big changes to the game will soon follow: an extra regular-season game, expanded and restructured playoffs and bigger rosters. So maybe now is the perfect time to look at other NFL rules that we could change.
Scoring in football has always seemed arbitrary, but one scoring play in particular takes the cake: The safety is the rarest, most difficult scoring play in the game. So why is it worth only 2 points?
Defensive touchdowns are far more common than safeties. There were 65 defensive touchdowns in 2019, but there were just 14 safeties forced on an offensive play. NFL placekickers made more than four out of every five field goals they tried — yet each of those 802 successful field-goal attempts was worth more than any safety.
As Alex Kirshner of SBNation’s Banner Society recently pointed out, Walter Camp’s foundational 1893 rulebook assigned 2 points to a safety without any real rationale. Unlike just about every other football rule (including the point values of touchdowns and field goals), the point value of a safety hasn’t been touched in the intervening 127 years. Kirshner argued that because harder, rarer sports feats are usually awarded more points, the safety should be worth a whopping 11.
As the old saw goes, “The most damaging phrase in the English language is ‘we’ve always done it that way.’”1 So setting Camp’s best first guess aside, how many points should a safety be worth?
First, we have to account for a key part of the safety’s value: the scoring team being awarded not just points but also possession of the ball. In the NFL, a safety free kick is set up from the kicking team’s 20-yard line, and the kicker can’t use a tee. In theory, this should give the safety-scoring team great field position. But in practice, that hasn’t been the case. Over the past five seasons, the average spot of the ball after a safety kick has been the scoring team’s own 36-yard line — less than 8 yards closer to the end zone than an average starting drive.
So if a safety kick is worth roughly the same as any possession, we can include that point total in a new safety valuation. As Kushner pointed out, the average drive is worth about 2.1 points, so we can say that “2 points and the ball” is roughly 4.1 points. To make sure our new safety value is worth at least what the old one was, we’ll round up to 5 points.
I tracked the change in win probability following safeties scored on an offensive play from scrimmage over the past two years. The highest-leverage safety came in Week 15, when New York Giants cornerback Sam Beal tackled Miami Dolphins tailback Patrick Laird in the end zone, expanding a 14-13 Giants lead to 16-13. That swung the Giants’ chance of winning from 63 percent to 81.4 percent, and they went on to win 36-20. Interestingly, that was one of the rare cases where getting the ball back did outweigh an extra three points. For the Giants, going up 19-13 and kicking off wouldn’t have been as advantageous as going up 16-13 and getting the ball back.2
Most safeties didn’t impact the game nearly as much. In fact, the average safety gave the scoring team a 7.1 percentage-point boost in win probability — and in almost every case, getting more points would have had more impact than getting the ball back.
If a drive is worth about 2 expected points, and a safety is worth 2 actual points, making the safety worth 5 points and giving the ball back should still slightly favor the scoring team. According to Pro-Football-Reference.com’s win-probability calculator, that’s how it works out: While a 2-points-and-the-ball safety boosted win probability by 7.1 percentage points, a 5-points-and-give-the-ball-back safety raised it by 9.7 percentage points.
On seven of the 21 safeties in 2018 and 2019, if the play had been worth 5 points and the ball went back to the team surrendering the safety, the change in win probability would have been less than 1 percentage point — usually because the game was all but won when the safety occurred. Of the 14 where the higher point-value and possession change would have made a bigger difference than 1 percentage point, 13 gave the advantage to the 5-point method — but only five of those 13 would have moved the win-probability needle by more than 4 percentage points over a traditional safety. So to give safeties significantly more oomph than they currently do, they have to be worth significantly more than 5 points.
Here’s the average increase in win probability for 5-, 7-, 11- and 15-point safeties over the past two seasons:
|Team giving up safety retains ball|
|Increase in win probability||7.13||9.68||15.39||23.86||31.73|
|Change over standard safety||—||2.55||8.26||16.73||24.60|
In case those numbers don’t seem all that large, remember that this is a proportional change in one team’s share of 100 percent. So if a hard-fought game is locked at 50-50, a 15-point safety would, on average, swing the probability split to 82-18.
It would almost be a Golden Snitch3 situation. Over the last two years of NFL football, 15-point safeties would have handed two-thirds of the scoring teams a 90-percent-or-better chance of winning that game.
So if 5 points is barely any better than the current rule, and 15 points is too severe, what’s the Goldilocks point where it’s just right?
The natural draw is to the 7-point mark. Besides being the middle option in our table above, it’s roughly twice as effective as the current rule at tilting the game toward the scoring team. A 7-point safety also doesn’t break the pleasing “threes and sevens” prime-number scoring Kirshner alluded to in his piece.
However, that still might not be the right value.
In 2019, there were 1,332 total touchdowns scored in 256 games. Making a safety worth 5 points, or even 7 points, changes offensive and defensive strategy. After all, offenses backed up against their own goal line already try hard to avoid a safety, and defenses that have pinned opposing offenses deep in their own end already try hard to score one. To really change how coaches approach the safety, it’s got to be worth more than you can score on one possession.
But if a safety is worth more than a touchdown — and, especially, worth more points than can be scored with one possession — then teams will have to anticipate a safety situation before it happens, accounting for the possibility of it in red-zone and third- and fourth-down play calls. It would certainly raise the entertainment value and add more late-game permutations.
Imagine a team getting the ball down by 9 points with a few minutes left, driving all the way to the one-yard line and then kneeling on fourth down. Deliberately surrendering the ball to give their defense a chance to win the game with a safety? That might sound far-fetched. But now that the NFL’s new, safety-minded onside kick rule has resulted in a recovery rate of just 10.4 percent in the last two seasons, a team down by 9 points or more in the game’s final minutes has precious little hope of getting back-to-back offensive scores. An 11-point safety would do the trick.
The NFL has so far been willing to alter some aspects of its games, but it hasn’t changed the value of a score since adopting the 2-point conversion in 1994. Would the competition committee ever consider such a drastic change? It’s impossible to speculate. That said, the NFL has been trying to minimize the role of the kickoff in recent years, sparking rumors that the league is considering eliminating the play altogether. If the NFL drastically changes the usual kickoff, standardizing what happens on a safety should be part and parcel of that reform.
Once you’re changing what safeties are worth, then, there’s no reason not to consider all the possibilities: weighting the safety in line with its rarity and difficulty, impacting overall game strategy and making it much more fun.
From that vantage point, 11 points looks like the sweet spot.