Last year after the Super Bowl I wrote about how Sean McVay’s critical and overdue decision to attempt a 4th down conversion helped the LA Rams win a Super Bowl. This year, 4th down decision making once again played a huge role in both who played in the Super Bowl as well as how the Super Bowl played. 4th down is one of my favorite things in the NFL so I figured we’d talk about it, and hopefully future seasons will provide me with yet more 4th down attempts in Super Bowls to discuss.
The big story here to me is that Philadelphia desperately needed both 4th down conversions they got, and on the last one they needed to convert to win the game, they blinked, punted, and the rest is history. Conversely, KC only really had one or two good opportunities to go for it on 4th down and took neither of them, but especially in the second half they were so dominant on offense that they never faced 4th down which is absolutely insane. The Chiefs ultimately won by three points, and in my opinion they likely would’ve won much more decisively had they converted the singular good 4th and Go situation they were offered. Conversely, if the Eagles had been more conservative on fourth down, it feels like they wouldn’t’ve been in the game at all, which I think shows the massive impact on winning that successful 4th down offense can have, but also the instability that being good on 4th down can have when you become complacent or reliant upon using 4th down to keep drives alive.
A good display of the benefits of being good on 4th down can be seen when we split up offensive efficiency (measured with EPA) based on downs. Successfully converting 4th downs adds tons of expected points because EPA pretty much only actually rewards plays that retain possession or score the ball. (This is what makes it different and more valuable than the traditional counting stat of yards and the slightly more advanced rate stat of yards/play, both of which fail to take into account whether the yards being gained are meaningful, rather than 16 yard passes on 27 yard downs.)
Plays that get you closer to retaining possession are typically assigned more or less neutral expected point values, and obviously plays that lose yards or don’t gain yards end up losing expected points. As a result, which plays add expected points correlates strongly to downs: most first downs do not convert possession, and traditionally are often not even called with that intention, as the boilerplate move on first and 10 is to run the ball a few yards and hope to set up a short pass or run to convert on 2nd or 3rd down. Just as an example: New England ranks as the best offense of the 21st century (1999-present, to be more specific, and yes, they are the best offense of both 1999–2010 and 2011–2022) by EPA/play according to rbsdm.com, and we can see what I’m saying play out by down.
Overall: .071/play
1st down: .047/play
2nd down: .086/play (note: obviously great offenses will have success on 2nd down that others will not)
3rd down: .082/play
4th down: .289/play (on 4th down offensive plays, no punts or FGs.)
So why are 4th down plays so impactful on expected points? (well, aside from the absurdly limited sample size of 4th down offensive plays even in contemporary football?) Because, as explained, if you are consistently getting closer to the 1st down marker but aren’t actually getting there, your expected point value is never going up until you successfully convert on 4th down, and the flood of expected points is then rewarded after the conversion. Here’s an example from the Super Bowl itself:
As you can see: the 4 yard play on 1st lost .01 expected points, and even a 5 yard run on 2nd down to set up 3rd and 1, a play that almost always converts, that actually lost .04 expected points. When Sanders finally gains zero yards on 3rd and 1, that loses expected points again. The 4th down play, even though it only gained two yards, gains back all of the expected points lost on third down as a result of retaining and refreshing possession of the ball.
As a result, if you are successful at converting 4th downs, it makes an incredible statistical difference. On the season, the Eagles were .100 EPA/play, ranking 3rd amongst all teams. On downs 1–3, they only gained .079 EPA/play (Pretty ho-hum, 5th amongst all teams, compare that to SB opponent KC’s .169/play on 1st-3rd down, they were literally twice as effective as Philly on what smart football people like to call “stable early downs”), being effectively neutral on first down with .001/play (that’s 1/100th of their overall offensive efficiency, abysmal, a shocking split to a more casual observer such as myself), highly effective with .141/play on second down, similarly effective at .147/play on 3rd. But they gained a monstrous .846/play on 4th down. Literally 20% of the Eagles EPA/play, meaning basically 20% of their offense, happened on 4th down. If the Eagles hadn’t discovered the effectiveness of the rugby maul on 4th down and were a team that punted on more of their 4th downs, I very seriously doubt they would have even been competitive in this game, or be playing in it to begin with. (It’s a “maul” by the way, “rugby scrum” has become the popular name for the Eagles 4th down play, and I had to have an actual UK rugby fan point it out to me, shoutouts to HDCantona from the Brickwallblitz/Deep Ball Project discord, and also shoutouts to Brickwall Blitz of the Deep Ball Project of course.)
By contrast, the Chiefs have proven themselves capable on 4th down when they choose to go for it, but they’ve also been fairly conservative in terms of actually doing it, which seems to be one of the very few small troubles in the paradise of the current KC dynasty. In the second half the Chiefs simply never faced fourth down, which is a discussion for another time, but in the first half, they did pass up a decent opportunity: 1st quarter, tie game, 4th and 3 from the Philly 24. The Baldwin WP% model (aka The 4th Down Bot on Twitter) actually categorized this as a toss-up since it was .9% WPA instead of the full 1% that causes the model to say “go,” but in the mindset of modern, post-Madden, “analytics” football, this would be a routine go for it call, especially with how successful KC had been on fourth downs when they wanted to be. Taking a 7 point lead in the first quarter doesn’t necessarily mean much, but if the rest of the game had played out the same and KC scored on that drive (with that distance, let’s assume yes), they would’ve won by 6 or 10 instead of it being a close game at the end. They didn’t end up needing it in the long run, but that doesn’t make not going for it a good decision there by Andy Reid, although admittedly hindsight favors “Analytics” in this case since the FG kick from that distance was no good. They score three there and win the game, nobody’s talking in all likelihood . . . although again, assuming all’s the same, that would only give them a 6pt lead meaning any TD play from Philly would win the game.
Philly, on the other hand . . . I don’t like this term very much, but the thought that immediately comes to my mind when I think of Philly’s last punt in the Super Bowl is “choke.” The other is “Ball Don’t Lie.” (Another “Ball Don’t Lie” moment is, of course, kicking a FG on a routine 4th and Go and having the ball hit the upright, even with one of the most accurate kickers in the NFL, Harrison Butker.) 4th quarter, 10:33 to go, Philly leads 28–27, and with how KC’s offense is doing in the 2nd half, the expectation should be that if you don’t score on this drive, KC is going to score and gain the lead be it with 3, 6, 7, or 8 points. It’s 4th and 3, and, this is probably the reason they didn’t go for it: they were on their own 32. I say this with empathy: it got just a bit too scary for them. 3 yards was just a bit too much. Their own 32 yard line was, understandably, way too deep. If they don’t make it, KC will score, and with this field distance, probably 7 points, it’s just a matter of how fast they do it. I get it. I honestly kinda feel for these guys, being presented with a situation where if this was week 3 of the regular season and you’d just discovered the rugby maul you might just fuckin’ try it to see if it works. But this is the Super Bowl, of course . . .
. . Pause: Let’s recognize the primary error in thinking here: everybody in this sport understands that starting field position is tremendously impactful. However, because of offensive explosiveness in the modern game, the truth is that as soon as an opponent reaches roughly the opponent’s 40, they’re within scoring distance from either their offense or their placekicker. What does this mean? It means that in the long run you cannot afford to be thinking about how easy you’re making it for an opponent to score, especially in a playoff game. You’re in a playoff game, the opponent’s likelihood of scoring is high, because the best teams in football right now tend to be ones that score a lot of points. In other words, past a certain point, it doesn’t matter. The likelihood is that if you fail on 4th down, the opponent is going to score. The likelihood is also that if you punt, the opponent will score. You need to make the most out of every possession you get, especially late in games with possessions becoming a scarce resource.
As an addendum, consider this. What is a better situation to be in: down 14 with 10m to go in the 4th quarter? Or down 14 with less than 10m to go? The objectively and logically correct answer is the former, and you can generate that by choosing to go for it rather than punt and failing. This is why you see teams in late game situations actively trying to allow the opponent to score while the opponent is trying to score as slowly as possible, for clock management. Not to mention, remember me saying that as a game goes long, possessions become a scarce resource? Giving the opponent an artificially short field on which to score generates the extra possession you will need to match those seven points to begin with.
There’s a very simple reason the Eagles didn’t go for it, which is that they had the lead and didn’t want to sacrifice good field position. This is another understandable but objective error. If you have a one-point lead, or even a two point lead, you don’t really have a lead, what you have is a mildly higher win probability than if the game is tied because of the extremely unlikely chance that you’ll take a 1 or 2pt lead from the beginning of the 4th quarter to the end of the game. Three-point FG, or 6-pt+ touchdown, you’ve lost the lead. A safety would even tie the game, if the opposing punter boots you that deep.
As such, the correct decision is obvious: you go for it. And if you don’t get it, you live with the consequences and try to make some chicken salad, after all you do now have more chicken shit to make it out of. Is that really any worse than, I don’t know, punting and then giving up so many return yards on the punt return that the Chiefs score in 3 plays? As it was said, as it was written: The Ball Does Not Lie.
It probably comes off like I think the Eagles are really stupid. I really don’t. Going that 4th down and then losing would be an incredibly bold gamble based on where football discourse is right now. I really wouldn’t be surprised if a risky go for it call in the playoffs that leads to a loss resulted in a head coach losing his job, even considering the degree to which the NFL has adopted and even promoted advanced stats and analytics. That being said, they did definitely make an error, and I hope my reasoning has cleared up why to some degree. This year it’s a conservative call. But I’d say we’re in the last decade, maybe even half-decade, of a decision like this, in this game situation, not being openly lambasted for its cowardice.
4th down decision making is a lesson in crisis management. The crises can be small or big, you can be facing 4th down from your opponent’s 3 whilst up 14 in the 4th quarter, or you can be facing 4th and 3 from your own 38 up 1 point in the Super Bowl against one of the best offenses of all time. But regardless, you are here at 4th down because things have gone wrong in some form or another on at least one of the previous three plays. Nobody plays to be on 4th down. But when you get there, you have to do what you have to do to win the game. Everybody understands this with play calling. If it’s third and 27, the vast majority of football coaches, players, and fans, would agree that you need to call a long pass, or at least a play disguised as one, and we all agree on the inverse which is that doing anything else is the wrong call. We seem to have difficulty with that in regards to 4th down decision making.
People say there are no style points in football, but that’s not actually true. I could go over plenty of examples of this, but in football, people really care about the way you lose. Going back to the situation I was describing, which is that final Eagles punt of the game, everybody, including those entrenched in the football bubble, should know that the chances of the Eagles defending the Chiefs on the next drive are unfavorable. Even if you only reference the current game situation, in the first half, the Chiefs had four possessions, scored on one of them, and shoulda-woulda scored on another, so the game situation dictates that the Chiefs have a 50/50 chance of scoring the ball . . . which, in 2023, one should automatically assume is at least 51/49 since we’ve seen consistently the past few years that the best offenses beat the best defenses, and specifically in regards to this era of Chiefs teams, we’ve seen their passing offense be able to flat out embarrass those defenses. Nick Sirianni must’ve gotten it in his head that his defense was better than the Bills defense the Chiefs had faced in the previous years’ divisional round. He decided to trust his defense, and we are reaching the stage in this era of football where doing that in a playoff game, late in the game in close score situations is a flat out tactical error, a game-management mistake. And yet, trusting your defense is widely seen as the better way to lose, because at least if you lose that way you . . . accomplish what, exactly? I think most people would say you accomplished not giving away free points in a possibly game-deciding situation, but the people who would think that way aren’t even remotely considering that, just for example, if you punt it you could give up field position just as good on the punt return. (Here’s your reminder: special teams play is not automatic.) They’re also, of course, not considering the benefit of the conversion, which even if that drive didn’t end up scoring, retains possession for another minute or so and gives KC less time to score the ball when they get it back, which could’ve won them the game since KC had to kick a FG to take the lead at the end.
I said this before in regards to the Ravens/Buffalo regular game from this season: you have three options. Convert 4th down, fail to convert 4th down, or punt the ball to an offense extremely likely to score. Only one of those outcomes is a win. The other two options are both losses, but in the football zeitgeist currently we prefer that a team face their game-losing play as deep into the game as they can rather than make a game-losing play earlier by making a risky decision. They’re both losses! Having confidence on 4th down, especially in late game situations, means being okay with doing something that might directly lead to losing (rather than indirectly), but also gives you a chance to win the day that you otherwise have to pray arrives at the end of your opponent’s next drive. Even more importantly, as the Ravens’ John Harbaugh has demonstrated time and time again that making difficult and risky decisions as a coach, (and in Harbaugh’s case on 4th downs, almost always the right one in the game situation,) and then willingly taking the heat if it fails, is one of the things that makes you a great leader. In the aftermath of the Ravens/Bills regular season matchup, the vast majority of the commentary was directed towards Harbaugh’s decision making, and absolutely none of it went to the players who made various crucial mistakes in execution, namely on red zone plays that were absolute dime play calls from the mostly reviled (including by me) Greg Roman with no positive results, including the exact 4th down call that was so controversial.
Let’s finally start wrapping this up by looking at a fake stat I just made up: point differential on drives with at least one decent 4th down opportunity according to the Baldwin Win% model, or “4th and Go Drive Point Differential”, or “4GDP.” This is not a very good stat for multiple reasons, but in this case I am seeking to be descriptive and not predictive. The idea is simple: we take a look at how many points were scored (if any) on drives with at least one successful 4th down conversion, essentially assessing the rewards earned for the hard gamble of choosing to go for it. On the one drive where Kansas City had a chance to go for it, and instead kicked the field goal that missed, they left at least three points on the board, but we’ll go ahead and call it three. On their last punt the Eagles were outside of scoring range, so we’ll go ahead and treat that as +/- 0, just neutral. On the other two drives where they converted 4th down, they scored 10 points, making for a 4GDP of 13 points in a game that was decided by less than a touchdown. I feel like that makes my original point obvious: the Eagles bravery on 4th down kept them in this game. The failure of that bravery in the clutch, especially with how the Philly defense was playing in that second half, is one of the major reasons they lost.
Last year I ended by saying “Long live Brandon Staley” but unfortunately he became a coward. So I’ll give the credit where it probably should’ve gone in the first place: Long live future Hall of Fame head coach John Harbaugh.
