Summer School 2020: A glossary of analytics terms

“Stats,” “data,” and “analytics” have become the hot new buzzwords around the NFL, and in doing so have also become a hotly debated subject with regards to football. For some, “analytics” is a way to make sense of the chaos and complexity that make up the game of football. For others it is an invasion of pencil-necked geekery into a world in which nerds have no place.

The coverage and fandom of the New York Giants is one of those places where the debate over analytics rages — and all too often devolves into flame wars. The Giants have struggled with implementing a large-scale analytics operation, while other teams like the the New England Patriots, Baltimore Ravens, and Philadelphia have embraced analytics whole-heartedly.

Between the success of teams who make great use of analytics and the digital revolution, analytics have become much more popular and accessible to the public at large. All of that together makes it almost impossible to talk about the NFL in 2020 without mentioning stats or analytics. Analytics itself is also fit for a number of Summer School topics in and of itself.

But with the NFL, ESPN, Pro Football Focus, Football Reference, Football Outsiders, and many other outlets all using their own metrics, we’ll have to define some terms before we can dig into what Analytics is and isn’t.

Analytics - The systematic study of large data sets to discover trends and patterns that might not be immediately apparent. Those trends and patterns can then be analyzed for potential competitive advantages

Grade (Pro Football Focus) - PFF’s proprietary grading system. Individual plays are graded on a scale of -2 to +2, with “0” being the “expected outcome.” PFF grades are not based on stats, but rather analysts’ determination from watching the play. The raw grades are then normalized to a 0-100 scale which is the same across positions.

(Further Reading - PFF Player Grades)

Expected Points Added (EPA) - The number of points expected to be scored on the next play based on down, distance, field position, and the contrast between a team’s situation at the start of the play and the end of the play. For instance, a short pass on first down would have a small EPA, a run for a loss (or a sack) would have a negative EPA, while a longer pass that results in another first down would have a larger EPA. The framework for EPA was created when Cincinnati Bengals quarterback Virgil Carter and Robert Machol (a professor at Northwestern) published a paper (linked below) in 1971 examining 8,373 plays from the 1969 season and calculating the point values of various field positions.

(Further Reading - Operations Research In Football)

ANY/A - An abbreviation for “Adjusted Net Yards per Attempt.” ANY/A is a metric which shows yards per total attempts (including sacks and interceptions) while also taking into account sacks (as negative yardage), interceptions (as a negative value), and touchdowns (as a positive value). ANY/A is intended to give a more comprehensive view of quarterback play than yards per attempt alone.

ANY/A formula: (pass yards + 20* (pass TD) - 45* (interceptions thrown) - sack yards)/(passing attempts + sacks)

Total QBR - ESPN’s proprietary grading metric designed to measure quarterback efficiency. Total QBR combines EPA with a variety of quarterback stats, including attempts, completions, yards, touchdowns, interceptions, rushes, rushing yardage, and sacks taken. That is then normalized to a 0-100 scale, with a “50” being average quarterback play.

Total Pressures - Takes into account sacks (plays in which the quarterback is tackled or a fumble is forced behind the line of scrimmage), quarterback hurries (plays in which a quarterback was forced to throw earlier than intended), and quarterback knockdowns (plays in which the quarterback is knocked to the ground shortly after throwing the ball).

Pass Block Win Rate - The number of passing plays in which an offensive line — or individual lineman — gives up a quarterback pressure within 2.5 seconds of a snap compared to the total . A play in which pressure is not generated within 2.5 seconds is considered a “win” for the blockers. A team with a good pass block win rate but a high number of sacks could be running a more slow-developing offense or have a quarterback who does not process or play quickly enough.

Pass Rush Win Rate - The inverse of Pass Block Win Rate, this stat compares number of passing plays in which a pass rusher is able to create a quarterback pressure in 2.5 seconds or less to the total number of passing plays. A defense that has a high number of sacks but a relatively low “win rate” likely has a talented secondary which gives the pass rush more time to get to the quarterback.

DVOA - A metric created by Football Outsiders which stands for “Defense-adjusted Value Over Average”. DVOA is used to break down games on a play-by-play basis, taking stats such as down, distance, field position, score, and game clock into account. DVOA also takes opponent quality into account, and can be applied to determine a team’s efficiency in number of aspects to every other team in the league.

DYAR - Similar to DVOA, DYAR was created by Football Outsiders and stands for Defense-adjusted Yards Above Replacement. DYAR is a total stat designed to compare individual skill position players to an average “replacement level” player.

Adjusted Sack Rate - A stat devised by Football Outsiders to give a better measure of pass protection than raw sack rate. Adjusted Sack Rate divides total sacks by total passing plays, as well as adjusting for down and distance (sacks are more common on third and long than a 1st and 5), and opponent quality (a sack facing a team with a great pass rush is more forgivable than a against a team with a poor pass rush).

Adjusted Line Yards - The first of three metrics used by Football Outsiders to measure a team’s ability running the ball. Adjusted Line Yards attempts to separate the yards blocked for by the offensive line from the yardage gained by the running back. A team ranked highly by Adjusted Line Yards has a good run-blocking offensive line. Adjusted line yardage accounts for runs for a loss to 4-yard gains.

Second Level Yards - Second level yards is used to measure rushing gains between 5-10 yards and weights the contributions of running backs and the offensive line equally.

Open Field Yards - Open Field Yardage is used to measure runs of 11 yards or more. This yardage is attributed solely to the running back.

(Further Reading - Methods To Our Madness)

Completion Percentage Over Expectation - Also abbreviated “CPOE,” this is a stat created by NFL Next Gen Stats using GPS tracking based on accuracy and the separation between the intended receiver and all nearby defenders. It compares a quarterback’s actual completion percentage to what is expected based on players’ positions on the field. A CPOE of greater than 0 suggests a particularly accurate quarterback, able to deliver catchable passes into tight windows, while a negative CPOE suggests a passer who should be completing more passes than he is.

Air Yards - The number of yards a pass travels downfield in the air before being caught. This metric separates passing yardage by a quarterback from yards after catch created by a receiver. Completed Air Yards shows the average distance traveled by the ball on completed passes while Intended Air Yards measures the distance traveled by the ball on all passes.

Aggressiveness - This stat from NextGenStats tracks the percentage of passes a quarterback attempts into tight coverage. “Tight coverage” is defined as a passes targeting a receiver who is within one yard of a defender. A quarterback with a high aggressiveness percentage (AGG%) attempts a high number of passes into tight windows.

Average Separation - The the average space between a receiver and the nearest defensive back at the catch point, as measured by NextGenStats GPS tracking.

Expected Yards After the Catch - A measure of yards after catch based on the distance between the receiver and the nearest defenders, the direction and speed of the players invovled at the time of the catch, and the number of defenders in the area at the time of the catch. This is compared to a receiver’s actual yards after catch to determine Yards After Catch Above Expectation, which measures how much better (or worse) a receiver is with the ball in his hands compared to his peers.

Efficiency (running back) - A stat from NextGenStats measures the total distance a running back travels behind the line of scrimmage. Lower efficiency (EFF) numbers show a runinngback who is a more “North/South” runner.

8+D% - An abbreviation for “Eight-plus Defenders In The Box”. This stat from Next Gen Stats shows how often a running back faces “loaded” tackle boxes of eight, or more, defenders.

Speed Score - A stat which combines a player’s speed (as timed in the 40-yard dash) and size to get an idea of their overall athleticism. A player who is 250 pounds and runs a 4.5-second 40 yard dash will have a higher speed score than a player who is 210 pounds and runs a 4.5-second 40 yard dash.

SPARQ - An acronym which stands for Speed, Power, Agility, Reaction, and Quickness. Nike holds the proprietary formula, but it has been reverse engineered elsewhere. SPARQ ratings are sport-specific and can be used give a comprehensive view of a player’s athleticism compared to other athletes in his sport and at his position.