Whenever Derrick Henry charges full speed down the gridiron, one may wonder how any defender is able to stop him. Even for NFL-level defenders, trying to stop a 247 pound human running at a top speed of 21 miles per hour seems like a task destined to result in a mouthful of turf. Yet this is the primary task NFL defenders are given every time someone holds the ball: get them down. The NFL tracks this accomplishment in terms of tackles, with a defender successfully downing a ball carrier earning them a binary 1, and in recent times companies like PFF have also tracked missed tackles as well. However, this system fails to account for the many nuances that go into a successful NFL tackle. This study will highlight one particularly important part of any successful tackle: getting the carrier to slow down. To find this, we used a combination of how the distance between a tackler and a carrier affects tackles and how the speed and weight of a carrier make it harder to take them down.
Effect of Distance:
Anyone who's ever watched a football game knows that the distance between the tackler and the ball carrier is incredibly significant in any tackle, as the day has not yet come where defenders can bring down carriers twenty yards down the field. However, how significant a 1 yard distance is versus a 2 or 3 yard distance is not nearly as intuitive as simply saying distance matters. We can get an understanding of the importance of distance by observing how far successful tacklers are from the carrier as opposed to defenders who did not record a tackle on the play. As predicted, tacklers end the play much closer to the ball carrier than other defenders, seeing a little over a yard of difference on average, or about 9 times closer than the average defender. The minimum distance between the two players occurs at around 5 frames, with a very slight increase afterwards, likely because of an event such as the ball carrier being flung to the ground. Using this information, we can create a metric to better understand close a defender actually has to be to make an impact.
Looking at the data, most tackles occur when the defender is within about a yard of the defender, with the number of tackles at other distances exponentially decreasing as the defender gets farther away. Based off this data, one can create a distance scalar that gives defenders more impact depending on their distance.
We take the average tackle distance, in our case 1.18 yards, and give anyone within it a distance scalar of 1, as they are deemed close enough to justify having a full impact on the ball carrier. After 1.18 yards, it fits an exponential graph to the tackling data where distance exceed 1.18 yards and accordingly assigns a distance multiplier between 0 and 1. Players with 0 have no impact and 1 have full impact. The result can be seen below:
When revisiting the original distance graph with the distance scalar, it shows a stronger distinction between tacklers and non-tacklers.
Speed Score:
Distance is obviously not everything when it comes to making a tackle however. Picture two scenarios where distance apart is held constant: one where the carrier is a full speed Derrick Henry and the other is a motionless Case Keenum. If NFL defenders had the choice of finding themselves in either of these scenarios, one can make a strong guess as to their preferred option.
To analyze this issue, we can apply Speed Score to tracking data. Speed Score is a metric calculated using the formula (200⋅𝑤𝑒𝑖𝑔ℎ𝑡)/(40-time)^4. First posited by Bill Barnwell as a way to measure running backs success, it has been used as a way to grade college prospects. However, speed score has potential applications to player tracking data as well, providing a way to account for the difficulty of tackling bigger, faster ball carriers. Before applying speed score to tacklers, we must first better understand the meaning behind a carrier's speed score. Using the speed score data for every frame, we can find how speed score changes in relation to the time of a tackle.
The graph demonstrates that the speed scores of the carrier increases until around the 1.5-2 second mark, at which point they begin to dramatically decrease until the time of the tackle. This makes intuitive sense, as it suggests that after getting the ball, carriers begin to accelerate until engaging with defenders, which results in them losing speed and having their speed score fall. This also reveals the insight that the average tackle begins to occur about 1.5 to 2 seconds before it occurs.
Looking at the differences between play times, it is noticeable that there is no increase in speed score for pass plays, likely due to most receivers having already reached their top speed while running their route. In addition, the average decline for receivers takes place with around 1 second remaining as opposed to run plays seeing the decline about 2 seconds in, which implies tackles take longer to complete on runs. While speed scores for carriers on run plays were actually slightly lower relative to carriers on pass plays, their slower decline may be a result of the increased difficulty of fully taking down large backs, especially in trenches where adjacent players can cause a carrier to remain upright longer than they would on their own. Scrambles have the highest speed score of the three play types, which reflects quarterbacks having the highest averages of any position group.
Looking at the top scorers at each position, speed score appears to do a strong job of recognizing difficult players to tackle. The top five running backs and tight ends were all recognized as strong performers in the first half of 2022. The top wide receivers in terms of speed scores were not just strong receivers in the first half of 2022 but players especially recognized for their physicality.
While initially surprising to see players such as Derrick Henry not in the top five, upon further examination this isn't shocking. Henry, who still finished an impressive 6th among eligible running backs, had the ball in his hands almost three times as much as running backs like Breece Hall and Cordarrelle Patterson. This larger sample size likely resulted in Henry's burst plays being weighed down, while Hall's massive runs in early 2022 had a proportionally larger impact due to an overall smaller sample size. It is for this reason that a frame limit was put on eligible players to avoid players with two big plays topping the charts.
Interestingly, despite being the heaviest group overall, running backs had the lowest speed score of any position group. Their larger sample size likely weighs it down for the same reasons as listed above, along with their need to accelerate after being handed the ball, while receivers and tight ends are already moving fast once they make a catch. Running backs also tend to have lower top speeds than other position groups. Quarterbacks surprisingly had the highest scorers, which may be a result of them having the choice of scrambling at opportune times, while other positions cannot control it when they get the ball with three defenders already surrounding them.
Distance Scalar Speed Score:
Now that the distance scalar allows us to better quantify how significant a player's distance is, we can pair it to Speed Score to better understand how much of an impact defenders have on carrier Speed Scores. To do so, we will create a metric called distance scalar speed score, or DS3. DS3 is calculated through the following equation:
DS3 give higher values to defenders when they are close to the ball carrier and the carrier has a lower Speed Score. The Speed Score of the carrier is added by 1 to prevent the result from going to infinity if their speed score is 0. Summing all the frames on a play gives the player's DS3 total for the play.
To examine how DS3 can be applied, we'll look at a play from Week 1 when the Buffalo Bills visited the LA Rams. There's 10:07 left in the third quarter, with the Bills electing to hand the ball off to Devin Singletary. After dodging an initial tackle attempt by A'Shawn Robinson, Troy Hill and Leonard Floyd took down Singletary for a 13 yard gain.
Looking at the play, we can see A'Shawn Robinson is the first player to have his scalar multiplier spike, but due to him missing the tackle he does not receive a high DS3 score. Singletary quickly runs past him and his scalar multiplier dips. Troy Hill engages him next and has more success, maintaining a scalar multiplier of 1 and increasing his DS3. Leonard Floyd joins Hill in tackling Singletary a half second later, resulting in Singletary being slowed down even more and seeing a greater increase in their DSS scores. Ultimatelty, Singletary is taken down thanks to the combined efforts of Hill and Floyd.
We can take the sum of a player's DS3 over the course of the play and multiply it by the percentage of frames they were at their highest distance scalar value. This benefits players who were able to maintain their tackles throughout the play.
The results of the Overall DS3 follow our analysis: Hill had the highest score overall, with Floyd in a close second and A'Shawn Robinson having minimal impact. Taylor Rapp, who arrived as Singletary was already about to go down, got a minimal amount of DS3. Since Troy Hill received the most DS3, he is credited as most responsible for the tackle.
We can calculate the percentage of DS3 a given defender earned on a play by taking their individual Overall DS3 score and dividing it by the sum of every defender's DS3 score. Doing this returns values with the same proportion as the values above, only now in a percentage form.
DS3 is not able to be summed in the same way speed score is, as differences in initial carrier speed score between plays, multiple tacklers vs 1, and other intricacies make it difficult to sum it in that manner. However, DS3 can be used to assign a penalty or bonus to players on a specific play. To do this, every time a player gets a distance scalar of 0.5, they are marked as having engaged with the defender. If the first player to engage failed to get a plurality of the DS3 on a play, they can be knocked as doing a poor job. These results can theoretically be summed across plays to get a better understanding of how defenders are performing.
DS3 goes beyond tackle numbers in assigning a percentage to players for how much they contributed to slowing down the carrier. Considering the difficulty of tackling ball carriers, being able to accurately measure the impact defenders have on getting ball carriers down is crucial to defensive success.
コメント