Stat Analysis

Advanced analytics on player and team performance

Lamar Jackson Part I: The Joy of Scrambles

Guest column by Keegan Abdoo

Lamar Jackson was frequently compared to Michael Vick in the pre-draft process, and for good reason. Both are quarterbacks from ACC schools who have strong arms and inconsistent accuracy, but the most distinct point of comparison that binds the two is that Jackson possesses the rare combination of breathtaking speed, elusiveness, and athleticism that has not been seen in an elite quarterback prospect since Vick.

While Vick's career ended up being quite the rollercoaster, it seems clear that some of his on-field struggles in his early seasons came from his coaching staff trying to force him into a more traditional offense. On the other hand, Vick's career year with the Eagles in 2010 came with the creative coaching of offensive coordinator Marty Mornhinweg and head coach Andy Reid, who adapted their scheme to take advantage of Vick's unique skill set. It will be paramount for Jackson's new coaching staff -- including Greg Roman (who coached Colin Kaepernick in San Francisco) and the aforementioned Morningwheg -- to craft an offensive scheme to cater to Jackson's electric athleticism. Part of this task will include understanding the schematic advantage that having a quarterback with rare agility can give an offense.

In the weeks leading up to the draft, Matt Bowen of ESPN brought up an anecdote from his career as a defensive back in the NFL during an appearance on the Bill Barnwell Podcast. When Bowen played for the Packers, they faced Vick twice one season, and their defensive gameplanning during the practice week leading up to those games was complicated by Vick's elite ability to make big plays with his legs in addition to his arm. "There are certain things against a quarterback with an electric skill set like Lamar or Michael Vick that impacts your defensive game plan so much you can't do certain things," Bowen said. For example, a defense "cannot play two-man. The cornerbacks are already taught to be in a trail position, looking at the receiver, they're not watching the quarterback. The safeties are already 15 yards deep at the snap to start and pedaling. So you have no one, if the quarterback gets out of the pocket, it's an automatic first down."

For the other two main man coverage schemes, Cover-1 and Cover-0, a defense can usually employ a spy who has enough speed to deter a mobile quarterback from breaking the pocket and scrambling. While this takes away a player in coverage or a pass-rusher, it's a better alternative to the quarterback tucking the ball and taking off for a first down. However, as Bowen explains, when it comes to a quarterback with Jackson's or Vick's exceptional athletic profile, Cover-1/-0 with a spy "looks great on the chalkboard, it looks awesome on Friday practice vs. the scout team. Guess what? You better have someone that can match a Michael Vick or Lamar Jackson athletically. And you don't."

Essentially, having a quarterback with that speed can force a defense to be more one-dimensional and use less man coverage and more zone coverage. Defenders in zone coverage will have their eyes in the backfield reading the quarterback instead of sticking with a receiver and thus will be able to react more quickly if the quarterback tucks the ball and takes off scrambling. While this theory makes plenty of intuitive sense, it is important to verify it through statistical analysis. Luckily, at Sports Info Solutions (SIS), we have the ability to do this, as we identify quality control information for every play in the NFL and college football's FBS. Utilizing our rich database of charting data that includes coverages and scramble types, we can test the conventional wisdom out in a few steps. These steps will be explored in three articles:

1. Examining league-wide scramble tendencies and efficiencies vs. different coverages to see how Man Cover-2 (or "two-man") compares to other coverages.

2. Looking at how Jackson's opponents changed their usage of Man Cover-2 against Jackson and see if there is a noticeable difference.

3. Analyzing Jackson's performance vs. Man Cover-2 to see if the coverage was warranted.

Nationwide Scrambling: Man Cover-2 vs. Other Coverages

During the 2017 season, FBS teams employed Man Cover-2 on 3,507 total dropbacks. That represented 6.1 percent of all dropbacks, making two-man the seventh-most frequently used coverage. Opposing quarterbacks scrambled 313 times vs. two-man, which gave the coverage a scramble rate of 8.9 percent. This was the highest rate of the seven high-volume coverages (over 1,000 total dropbacks). The only coverage overall that had a higher rate of scrambles per dropback was Tampa-2 at 10.8 percent, but that was on one-fifth the sample size (only 650 dropbacks).

So the first part of the theory bears itself out in the data. Quarterbacks do scramble against Man Cover-2 at a higher rate than any other high-volume coverage. The next question we should ask is, does this happen for a good reason? Our next chart shows the success rate and yards per play of all coverage schemes that saw over 50 scrambles last season.

Two-man allowed one of the highest efficiencies on scrambles of any coverage last season, the third-highest Success Rate (50.8 percent) and the second-highest yards per attempt (7.6). Another thing to note in this chart is that the top three coverages by Success Rate allowed were the three types of man coverage (Cover-0, Cover-1, and Man Cover-2). In 2016, two-man allowed the highest success rate (50.7 percent) by a full two percentage points and highest Y/A (7.95) by 0.9 yards (Cover-1 was second in both).

At SIS, we classify the primary reason for every scramble. There are three categories of scrambles: coverage, pressure, and open rushing lane. In the FBS last season, pressure scrambles were the most frequent (44.6 percent of all scrambles), followed by coverage scrambles (33.3 percent), and then open rush lane scrambles (22.1 percent). From an efficiency standpoint, when a quarterback scrambled because he saw an open lane, he was most efficient, averaging 8.7 Y/A and 59.6 percent success rate. Pressure scrambles were the second most explosive (6.4 Y/A compared to 5.8 Y/A for coverage scrambles), but their efficiency was almost identical to coverage scrambles (41.4 percent vs. 41.7 percent success rate).

Looking at the scramble type distribution between the nine main coverages (all had at least 70 scrambles), Man Cover-2 has the fourth-highest rate of pressure scrambles (45 percent), fifth-highest rate of coverage scrambles (33 percent), and sixth-highest open run lane scramble rate (22 percent). All nine coverages had at least 25 coverage scrambles, but Man Cover-2 really stood out as being especially susceptible to this scramble type on a large sample (104 rushes). It gave up a 52.9 percent success rate and 7.3 Y/A, which gave it a substantial lead over second place Cover-1 (47 percent success rate and 6.3 Y/A). This makes a lot of sense given Bowen's explanation earlier, as the cornerbacks are in trail technique and not looking at the quarterback. If they are locking down their matchups, the quarterback can pull the ball and break the pocket and will have a whole lot of open field in front of him. Cover-1 presents a similar issue, but often has at least one underneath zone or spy defender.

NFL Scrambling: Man Cover-2 vs. Other Coverages

Man Cover-2 was also the coverage most susceptible to scrambles in 2017 at the NFL level, where the coverage was even more popular. It was employed on 1,642 dropbacks last season, which represented 8.1 percent of dropbacks. This made it the fourth-most popular coverage (not including screens) in the NFL last year. Nine teams employed two-man on more than 10 percent of the dropbacks they faced. These teams include the Steelers (10.2 percent), Buccaneers (11.2 percent), Titans (11.4 percent), and Broncos (11.5 percent), all of whom the Ravens are scheduled to face in the 2018 regular season. Another AFC North rival, the Bengals, ranked 11th at 9.2 percent. Opposing quarterbacks scrambled 111 times vs. Man Cover-2, which was the highest rate (6.8 percent) of any coverage by a considerable amount.

Again, we should check the efficiency of scrambles vs. each type of coverage to see if quarterbacks are making rational decisions by scrambling so often vs. Man Cover-2. In the NFL, two-man was the worst defense in containing scrambles. Of the seven coverages that faced at least 25 scrambles last season, Man Cover-2 allowed the highest Y/A (8.5) and Success Rate (67.6 percent). (Note that Success Rate is calculated differently for college than the NFL, with a 50%/70%/100% threshold for yards to go for college instead of 45%/60%/100% for the NFL).

Clearly, the league-wide data backs up our hypothesis: two-man is the coverage that leaves defenses most susceptible to success by scrambles. It repeats itself across both the NFL and FBS, and the good news for Jackson is that the Man Cover-2 is used in the NFL at 133 percent the rate that it is used in the FBS.

Scramble types follow a pretty similar distribution at the NFL level. Pressure scrambles once again are the most common (45.4 percent), followed by coverage scrambles (29.8 percent) and then open rushing lane scrambles (24.8 percent). When quarterbacks scramble with an open running lane, they have a ridiculous 69.3 percent success rate and 8.5 Y/A. Scrambles that are forced by pressure are more efficient (53.8 percent success rate and 7.3 Y/A) than those forced by coverage (48.3 percent success rate and 6.4 Y/A).

Man Cover-2 had the second-highest relative pressure scramble rate (49 percent), fourth-highest rate of coverage scrambles (29 percent), and sixth-highest open run lane scramble rate (22 percent) of any of the seven coverages that faced at least 25 scrambles last season. Just like in the FBS, coverage scrambles are especially effective against Man Cover-2, with by far the highest success rate (75 percent; next closest was Cover-4 at 52.9 percent) and Y/A (9.0; next closest was Cover-3 with 6.0) of the five coverages that saw at least 15 coverage scrambles. Combining this with the larger-sample FBS data, it is clear that coverage scrambles are a huge weakness that can be exploited in two-man coverage.


After exploring how scrambles perform against each coverage, we can reasonably conclude that Man Cover-2 is the coverage most susceptible to scrambles at the FBS level. This is based on the combination of two-man giving one of the highest scramble rates and highest efficiencies of all the coverages. At the NFL level, there's no question that scrambles are most successful against Man Cover-2; it allows the highest scrambles per dropback, yards per attempt, and success rate of any coverage. Two-man is used more frequently in the NFL, moving from the seventh-most frequent coverage called in the FBS up to the fourth in the NFL. Pressures cause the plurality of scrambles against Man Cover-2.

I've broken down the league-level efficiencies and scrambles against each coverage and shown that Man Cover-2 is the most vulnerable to being exploited by a scrambling quarterback. In the next article, later this week, I will focus on Lamar Jackson specifically, looking at his charting data to see how he performs vs. two-man and if defensive coordinators decreased their use of the coverage when they faced him.

Keegan Abdoo is a former Research and Development Intern at Sports Info Solutions. Interested in working for SIS and being able to perform this type of research with access to its rich charting database? Apply here for a R&D position.


6 comments, Last at 03 Sep 2018, 3:55pm

1 Re: Lamar Jackson Part I: The Joy of Scrambles

by BywaterBrat // Aug 27, 2018 - 3:32pm

Great stuff--- getting excited about what the SIS-FO brotherhood can do together!

Login or Register to post comments

2 Re: Lamar Jackson Part I: The Joy of Scrambles

by jtr // Aug 28, 2018 - 7:46am

Interesting stuff, I've never seen stats broken down by defensive coverage shell before. Very cool.

One little pet peeve, I think a bar would work much better than a line for the success rate graphs. The X axis doesn't actually mean anything on these charts, so the line doesn't mean anything either. Two bars (YPA and success rate) for each category would make more sense and make it look less like you're trying to find some sort of trend over the spread of the X axis (which is usually what line graphs are for).

Login or Register to post comments

6 Re: Lamar Jackson Part I: The Joy of Scrambles

by njgiants333 // Sep 03, 2018 - 3:55pm

Thanks for your feedback! I will avoid using a line graph in a double axis graph going forward, duly noted.

-Keegan Abdoo

Login or Register to post comments

3 Re: Lamar Jackson Part I: The Joy of Scrambles

by David // Aug 29, 2018 - 2:31am

Marty Mornhinweg

Login or Register to post comments

4 Re: Lamar Jackson Part I: The Joy of Scrambles

by Vincent Verhei // Aug 29, 2018 - 2:46am

Aw dammit. Fixed. Thanks for pointing that out.

Login or Register to post comments

5 (+91)-9560052561 High Profile Cha

by RenukajainJain // Aug 31, 2018 - 4:18am

Chandigarh escort agency offers you wide choice of Russian and Indian mobies for sensual pleasure. We are one of the finest escorts service provider in Chandigarh. Our escorts are students, models, busty, air hostess and housewives, charming and very friendly. You will find more than your expectation which gives you memorable pleasure with our sexy escorts.

Login or Register to post comments