Clutch moments are the defining features of a great football game, the time when the stakes are the highest, the fans are their loudest and the margin for error is at its lowest. It is when the best players step up and put the team on their shoulders to make the plays that determine the outcome of the game. With the game and season on the line, who do you want leading your team in the 2 minute drill? In this article, I take a deep dive into how the fourteen different playoff quarterbacks have performed this year in these clutch moments.
The first step of this analysis was to define a "clutch play." It was decided that a clutch play occurs when the outcome of the game is still uncertain and the pressure is high for the team with the ball. These moments are at critical points where an offense needs to convert a 1st down to keep a drive going or when the offense has the ball at the end of a game when the score is close and the game is on the line. For these reasons, I defined a “clutch moment” as plays that occur:
In the 4th Quarter
In Overtime
On 4th downs
On 3rd downs longer than 5 yards
This weekend marks a significant milestone for five quarterbacks that take the field for their first career playoff game. But, just how have these players and the other nine starting quarterbacks fared in high-pressure situations throughout the season? To figure this out, I needed to delve into the data of all fourteen starting quarterbacks whose teams have made the playoffs and evaluate their ability to deliver in the clutch.
Using this code below in R, play by play data from nflfastR was filtered to only include passing plays among these fourteen quarterbacks in the defined “clutch moments.” Once this data cleaning and wrangling had been complete, it was time to visualize the data and figure out which QBs were “clutch” and which were not.
pbp <- load_pbp(season = 2022)
teamlist <- c("KC", "BUF", "CIN", "JAX", "LAC", "BAL", "MIA", "PHI", "SF", "MIN", "TB", "DAL", "NYG", "SEA")
df <- pbp %>%
filter(play_type == "pass") %>%
filter((qtr == 4 | qtr == 5) |
(down == 4) |
(down == 3 & ydstogo >= 5)) %>%
filter(posteam %in% teamlist) %>%
filter(passer_player_name %in% c("P.Mahomes", "J.Allen","J.Burrow", "T.Lawrence","J.Herbert","T.Huntley","S.Thompson", "J.Hurts","B.Purdy","K.Cousins", "T.Brady","D.Prescott", "D.Jones", "G.Smith")) %>%
group_by(passer_player_name,posteam) %>%
summarize(interception = sum(interception),
yards = sum(passing_yards, na.rm = TRUE),
pass_touchdown = sum(pass_touchdown),
cpoe = sum(cpoe, na.rm = TRUE),
epa = sum(epa),
wpa = sum(wpa),
n_chances = n(),
epa_play = epa/n_chances,
cpoe_play = cpoe/n_chances)Expected Points Added (EPA) is a statistic that measures the impact of a play on a team's expected point total for a drive. It assigns a point value to every play based on the likelihood of the drive resulting in a touchdown, field goal, or turnover. This statistic is useful in evaluating a quarterback's performance in clutch moments because it takes into account not only the outcome of the play, but also the situation and the field position. However, not all of the playoff QBs have played the same amount of games and played the same amount of snaps in “clutch moments.” For example, Brock Purdy, “Mr.Irrelevant” or the last pick in this year’s draft, has burst onto the scene after season-ending injuries to 49ers quarterbacks Trey Lance and Jimmy Garoppolo. However, he has only played in six game as compared to Patrick Mahomes playing in all seventeen. To ensure that the data would not be affected by the differences in the number of snaps, EPA/Play was used. This simply divides a QB’s EPA by the amount of plays he has played. This helps to control for the number of opportunities a quarterback has had and allows for a more accurate comparison of their performance in clutch moments.
The graph below represents the EPA/Play graph for the fourteen playoff QBs. Note: In this graph, the players are graphed along an x-axis that represents how many passing attempts they have each had on clutch plays, and the y-axis represents each player’s EPA per play. The dotted line represents the average EPA per play for all of the quarterbacks in the sample.
This graph has some very interesting trends. It shows that Tom Brady and Kirk Cousins have played the most snaps in clutch opportunities this season, as Brady leads all playoff QBs with nearly 350 passing attempts on clutch plays. This has not translated to points, however, as his EPA per play is below average and is 6th lowest among the fourteen quarterbacks. Superstars Joe Burrow and Patrick Mahomes have been the most clutch with the highest EPA/Play scores of the group. It is interesting to note that Josh Allen has been slightly above average and Jalen Hurts, the QB for the No. 1 seeded Eagles, has the second lowest EPA/Play in clutch moments. At the bottom, with a pathetic score of -0.50, Skylar Thompson, the replacement for Tua Tagovailoa, is by far the worst in clutch moments. In other words, when Thompson plays QB for the Dolphins, for every play on average, the team loses 0.50 expected points.
TD:INT ratio was also used to evaluate a quarterback's performance in clutch moments. It measures touchdowns to interceptions, indicating how well a quarterback takes care of the ball and makes good decisions under pressure. A high ratio means good decision making and points production while a low ratio means the opposite.
The next graph represents the ratio of touchdowns to interceptions on clutch plays. In this graph, the x-axis represents passing touchdowns in clutch moments, and the y-axis represents potential clutch moments lost to interceptions. A vertical dotted line representing the touchdown average and a horizontal dotted line representing the interception average intersect at a point, dividing the graph into four sections.
The higher up on the graph the player is, the more interceptions he had. The further to the right the player is, the more touchdowns he had. The optimal position to be in was the section on the bottom right, labeled in green, as it represents the ability to perform in big moments. Conversely, a player in the top left section of the graph was a player who folded under extra pressure and threw interceptions on key plays.
In the bottom right section, the most optimal, we see Jaguars’ QB Trevor Lawrence, Seahawks’ QB Geno Smith and, again, Bengals’ QB Joe Burrow. Burrow has thrown almost 20 TDs in clutch moments while only throwing 3 interceptions which is a remarkable statistic.
One noticeable data point absent from the graph is Ravens quarterback Tyler Huntley. Huntley has been horrific as a backup in Lamar Jackson’s absence, and his stats for both touchdowns and interceptions (1 and 1, respectively) were too low to be included on this graph.
Some of the QBs that throw a lot of touchdowns, but also a lot of interceptions are Kirk Cousins, Tom Brady and Josh Allen. These QBs are able to connect on scoring drives, but are more reckless in clutch moments and turn the ball over too much when the game matters most.
Justin Herbert of the Los Angeles Chargers has been the worst clutch QB in terms of TD:INT ratio. He has thrown the same amount of TDs as INTs in clutch moments and is the only one in the top left sector.
Brock Purdy and Skylar Thompson, who both land in the bottom right of the graph, have had few opportunities for clutch moments, as they have not played for the whole season. It is interesting, however, to see Jalen Hurts in this bottom left section. It is important to remember that Hurts might have low numbers because his team, with a league best record of 14-3, has usually dominated and thus hasn't had as many opportunities to be "clutch" this year.
In conclusion, my analysis of the performance of the fourteen NFL quarterbacks in this year’s playoffs has revealed some interesting trends; Joe Burrow and Patrick Mahomes have been the most clutch QBs based on their EPA/Play and TD:INT, while Jalen Hurts and Justin Herbert have not been as good. Brock Purdy has not had enough opportunities to be clutch for the 49ers, although he might be challenged to step up to the occasion when playoff games start. Skylar Thompson and Tyler Huntley have been disappointing, and I expect neither to have good games in the wildcard weekend.
We will have to see how these QBs perform on Saturday and Sunday, but expect the teams with the most clutch QBs to have the most success as we get closer to the Super Bowl.