Inspired by Jon Bois’ “The search for the saddest punt in the world” (2019)
In the 2023 football season, Cornell ran 833 offensive plays. While each of these plays were impactful to each game, this was the best one:
This play had the highest EPA, or Expected Points Added, which is a standardized metric to measure how good a particular play is. Essentially, EPA is a model that takes into account the time, field position, down, and distance of a particular play, and compares what should have happened to what actually happened to measure a play’s success.
This particular play, a 2nd and 10 from about midfield, had an expected point value of 1.73 points (any given play run from this particular moment should result in an average of 1.73 points). Since a touchdown was awarded at the end of the play, the actual point value of the play was 7 (including the extra point). Hence, the EPA was 7-1.73 = 5.27 points.
But 5.27 expected points added doesn’t really mean much without context. With this in mind, here’s a visualization of the EPA for every play from Cornell’s 2023 season, with the play we just discussed labeled.
Beyond the numbers, the play was beautifully set up. Wide receiver Doryn Smith ‘26 motions into the slot without a defender truly matching his movements, indicating to quarterback Jameson Wang ‘25, that the defense is likely in zone coverage.
Indeed, they were; when the ball was snapped, the two outside corners dropped back, and the linebackers spread out over the middle of the field in quarters coverage.
The routes on the right side of the field were specifically designed to target this zone: as Smith commits to a long dig route, receiver Davon Kiser ‘26 runs a deep post, a combination of routes that baits the right-middle safety to cover the intermediate route, ultimately giving up the deep shot.
Wang likely knew this when he took the ball, as his eyes stayed shallow so as to not alert the safety to the post. However, as soon as Kiser breaks away, Wang steps up in the pocket to avoid the edge pressure and heaves the ball towards the end zone. It’s now a footrace between Kiser and his defender.
With the ball in the air, let’s now talk about who the ball is going to: Davon Kiser. With a 4.32 40-yard dash, Kiser is the fastest football player in Cornell’s history, enabling deep passes. Outside of his speed, Kiser’s receiving is also noteworthy. Compared to his teammates, including WR1 Nicholas Laboy, Kiser has the most passes taken for over 25 yards on the season (seven). Knowing this throw is going more than 25 yards, with Kiser’s historic speed and receiving prowess, if you had to choose any one person to be at the receiving end of this pass (across the history of Cornell football), it would most definitely be Kiser.
Luckily for Cornell, it is, and the play goes down in the record books as the longest touchdown of the 2023 season.
Bus is the 48-yard bomb to a speed demon of a receiver resulting in the season’s longest touchdown pass actually Cornell’s best play of this past fall? Well, not quite; take a look at the scorebug.
In the game of football, with so many moving parts, sometimes even the “best” plays get overshadowed. This one play that I described might be the best individual play of the season, but in the bigger picture, with Cornell being down two scores when the play happened and the fact that the team would end up losing the game by 15, means this play isn’t quite the best moment of the season - its impact was limited.
For that, we’d have to look past the shortcomings of EPA per play, which lacks vital in-game context. For instance, a 99-yard touchdown would have an incredibly high EPA, but if the scoring team was down by 50, would that moment be as grand as its EPA indicates? With this in mind, I’m introducing a new statistic: Clutch Index.
Essentially, this combines a number of elements that make a play “clutch” and outputs out a number that quantifies just how clutch a given play was and is inspired by Jon Bois’ Surrender Index.
Let’s go through this part by part. Win probability added acts as an estimate for how meaningful a play was so that we weigh more toward plays that contribute to greater chance of winning, excluding meaningless stats. For a play to be considered “clutch” it must significantly help the team’s chances to win the game. We then multiply this by the amount of yards gained on the play. Simply put, the longer the play is, the more clutch it is. A 1-yard touchdown to win the Super Bowl might be the best play in that game, sure, but a 99-yard Hail Mary to win it might just be the best play in football history. The function 1+0.01(3600-Seconds Remaining))3 is best explained using the graph below: the less time on the clock, the more clutch a play should be.
Now, let’s look at the multipliers. The index is doubled if the play results in a touchdown, if it was a go-ahead play, if it was fourth down, or if the game was tied. To heighten the stakes even more, if the team is down by 1 score, then the play matters four times as much. To explain this reasoning, I’ll turn to basketball. A game-winning shot means so much more if your team is down versus if you’re tied. The stakes are higher when you’re down: if you miss, you lose. If it's tied, a miss can still let you survive into overtime.
(For an example calculation of Clutch Index, a go-ahead 29-yard touchdown play with a Win Percentage Added WPA of 0.23 that occurred with 11 minutes and 18 seconds to go (not on fourth down) would have a Clutch index of 285.67.)
Using the formula for Clutch Index, here’s a plot of all of Cornell football’s plays throughout the season, sorted by when they occurred in the game and ordered by their “clutch” factor.
This yields exactly what we want, an extreme outlier and clear “moment of the year.” Here, I would’ve loved to show a visual of the play, but there unfortunately is no replay footage of this play. In a way, it’s poetic. The lack of accessible media surrounding Cornell sports has always made it difficult to highlight their successes: even the best moment of the season is a lost recording. Relating to football, the lack of a visual aid is representative of the ins and outs of the game. A play that was, statistically, the most important moment of the season for Cornell goes unnoticed to the untrained eye of the ESPN highlight curator. The only record of the play is a written account, featured in the Cornell Daily Sun newspaper on September 23rd:
“The offense stuck to its roots, mixing the rush and the pass as it marched down the field.
Wang connected with sophomore wide receiver Doryn Smith for 22 yards to start the
drive, and a 17-yard rush from Carothers had the…”
A 22-yard pass. The play itself is something that you might see on any given Saturday, but its context is what makes it important. Let’s go through the formula again. This play gave Cornell an added 19.8% to their win probability, went for 22 yards, and most importantly occurred with 189 seconds left while down one point. All of this results in a clutch index of about 71, which makes sense for an explosive pass play late in the fourth quarter when down by one score. The play was needed and valuable and thus clutch, as its index indicates.
To truly understand the importance of this play, however, we have to look beyond the play itself and position it in the larger surroundings of the season. This play happened on 3rd and 10, where, on the season, Cornell struggled, converting just 38% of their 3rd downs, and only 37% of 3rd downs in the 4th quarter.
Third and longs were even worse: they converted only 13% of 3rd and 10 or longer, with that number also dropping to 11% in the 4th.
As a result, the sheer unlikeliness of this play being a success makes its occurrence even more clutch.
Let’s now consider the consequences of the play. Despite the improbability of success, a first down was gained, and soon after, Cornell would find themselves in range for a game-winning field goal.
Going back to the start of this drive, at the 24-yard line, down by one, Cornell needed a score. With just three minutes left in the 4th quarter, down one point, this drive, as a whole, meant everything: do or die. On the season, with drives starting before the 25-yard line, Cornell football would only put up points 10% of the time. But here we are, seconds away from a potential game winning field goal against Yale.
The ball is snapped and held well, with the power to decide the game now solely resting upon the leg of Jackson Kennedy, the place kicker. Yet, so much more than just a victory was at stake here. Leading up to this moment, Cornell had lost the last five games against Yale, and eight of the last nine. Yale had beaten the team by a combined 118 points across these nine games, meaning that the games weren’t even close.
To even be within one score of Yale was somewhat unprecedented in recent years, with the average margin of victory for Yale being 13 points. Yet, here we are, riding off the momentum provided by the clutchest play of the season, the kick goes right down the middle.
Everything about this sequence wasn’t meant to happen. The clutchest play of the season was highly improbable, and the resulting field goal, statistically, should have never even have been attempted given the odds of Cornell scoring on long drives.
Still, everything that happened… happened. That’s the beauty of sports. I’ve always tended to believe in the idea of numbers being predictive, but as the team celebrates around the kicker, Kennedy, how much of that really matters?
