Baseball has always been the leader in sabermetrics due to the sheer volume of repeatable plays. It's "easy" to put a lens on an individual player's performance and value when you can point to hundreds of plays over the course of a season. Football's different. Instead of a pitcher facing a batter, or a fielder facing a ball in play, there are eleven players on both sides of the ball each running their own unique role in the down.
Over the course of the year, offensive linemen and quarterbacks can be expected to play every snap. Defensive back and inside linebackers can be counted on as well. Everyone else? It comes down to coaching and positional need, which can greatly reduce the amount of opportunities those players have to make an impact play.
A baseball player can see a minimum of three plate appearances if they play a full game- and they play 162 of them- and the player with the fewest average pitches per appearance is seeing 3.34 P/PA, good for a projected 10 data points a game. There's no rule that a receiver has to see 10 targets- heck, that's a busy game- and there's only 16 regular season games. The amount of data points in baseball allows their projections to be much more concise than their football counterparts.
As stated before, there are a few players who perform the same actions that could rival baseball production. I'm sure someone could develop a value for defenders in the same vein as fielding statistics, but even that won't have the same precision due to different defensive schemes preventing season-long consistency. There are offensive linemen, who are already graded out on their blocking (bonus for pancakes!), but their value is less statistical and more observational- something that Pro Football Focus already has under wraps.
The last player who sees the same high volume of plays is the quarterback.
Now, receivers will run their different routes. Defenders will play their unique sets. Quarterbacks will have their target. But how does it all come together? It's easy to point to "Completion Percentage" and say that tells the story of a quarterback's accuracy- and it does to an extent. That number shows the final portrait of a quarterback and the quarterback deserves to be valued on it. However, it's important to note each individual brush stroke to see if maybe, just maybe, the artist could have improved his painting.
Much like a batter has a strike zone for the pitcher to focus on, so too do wide receivers for quarterbacks.Scouts will talk about a receivers catch radius, and Patriots fans were witnesses to Brandon Lloyd's absurd efforts to corral in some inhuman receptions. The strike zone for a receiver is wherever the ball is catchable for them. Does the ball hit the player right in the mitts? Then that's a strike for the quarterback. Can the player get his hands on the ball with a normal range of motion (*subject to my imperfect perception and All-22 footage)? Then that's a strike. Does the player have to lay out to make the play? Then that's not a strike- that's added value by the receiver.
Of course, this strike zone is subjective and the value attributed to the quarterback and to the receiver is very fluid. Some passes are designed for the receiver to lay out- and some passes just have to be thrown that way to avoid an interception. Still, it's fairly simple to develop a three tier strike zone:
Zone 1: Quarterback throws the ball right into the player's chest. If the receiver doesn't have to move his arms outside of his body, then that's a strike in the middle zone.
Zone 2: Quarterback throws the ball in the outreached arms of the receiver. This adds the receiver's arm lengths to the catch radius. If the receiver has to adjust their arms to make the catch (above their head, by their legs, to either side), then that's still a strike, but the value is weighted towards the receiver.
Zone 3: Quarterback throws it where the receiver cannot reach out for it. This includes passes that are too high, too far, into the dirt, out the sideline, and, yes, those that Brandon Lloyd smiled into the end zone with.
As a result, we're able to develop a strike zone of which to evaluate the quarterback that looks a little like this:
The middle number (bold) represents the strikes from the quarterback. The middle ring (italics) represents strikes that are assisted by the receiver. The outside band (plain) are non-strikes. Fairly simple and it tells a story- not the whole story, though.
Unlike baseball, with set distances between the pitchers mound and home plate, the quarterback will always be at varying distances from the receiver. Add in the fact that the receiver is a moving target and you'll find the strike zone much more difficult to pin down. As a result, we can't just create one strike zone for the quarterback- we have to create multiple.
With football, there are three main fields- left, middle, right- and there are four primary depths- behind the line of scrimmage (LoS), 0-10 yards past the LoS, 11-20 yards, and 21+. Now there's no official "left" or "right", as NFL.com essentially adds a direction if the receiver is outside of the hashmarks, while PFF chooses the numbers as their marker. For this exercise, Left/Middle/Right is determined by the receiver's first football move. If the receiver is running left and they catch the ball prior to the numbers, but their momentum and first step brings them towards the sideline, then that registers as a "pass left".
Combining these fields and depths can generate a set of 12 strike zones to better record the quarterback's accuracy. Of course, this generates 300 squares for data and, when a quarterback will only throw 550-650 times a year, there will be plenty of blank space. This is okay- sometimes empty space is just as important as populated data.
The Test Subject
The Patriots' quarterback Tom Brady is a prime subject for observation as his 637 attempts provide a gold mine of potential information. I was able to map out his passes in every single game, broken down by receiver, in order to glean any type information. Who was his favorite target? How did his perform without Hernandez? Without Gronk? How did he grow as the season wore on? Which players stepped up? Which passes did Brady flourish with? Struggle on?
There was plenty to learn from watching and mapping the 2012 season- and we'll be breaking down the data further as I continue to evaluate the numbers. For now, let's look at the big details:
Tom Brady's Heat Zone
From the bottom to the top we have the four different depths and from left to right we have the three different zones. In each box is the strike percentage for the zone.
What we can tell:
Brady Leans Left - Brady throws an astonishing 13% better to his left than to his right. Of course, Brady threw the ball 70 times more to his left than to his right (258 vs 182), so if Brady were throwing high volume passes to Wes Welker, you would expect the left's number to be inflated. Well, Welker only counted for 13 additional to the left. The real star was Brandon Lloyd whose sideline feats accounted for 44 additional passes to the left sideline, and his reliability as a target shone through. Of course, Lloyd provided nothing with YAC, but the fact that he and Brady developed that rapport on the left sideline helped move the chain on more than just a few occasions.
Screen Plays - The Patriots offense loves to take advantage of open space and will use quick bubble or slip screens to get the shifty receivers (Welker, Julian Edelman, Aaron Hernandez, Danny Woodhead) in open space. Well, the Patriots favored the left to the right 43-to-25. Brady threw 33 quick strikes into Zone 1 to the left, while only 12 to the right. In fact, Brady threw 8 passes into Zone 2 on both sides, despite the volume difference. And when looking at Zone 3? Brady had twice as many hit the dirt going right than left.
Dead Zone - The biggest discrepancy of completion from left-to-right is between the 10-20 yard depth, with a shocking 15.4% difference. Brady had twice as many sail out of bounds, over the head of the receiver, when throwing right compared to left. His touch was much better going left, although he did managed to throw 7 passes just wide of the mark. The difference? Lloyd again. His back shoulder prowess, as well as his deeper hitch routes were extremely reliable to the left, but the back shoulder connection never happened to the right.
Up the Gut - Want to know how much Brady missed Rob Gronkowski? Brady's passes up the seam are some of the more reliable passes in the play book. The Brady/Gronk connection had only one Zone 3 pass on attempt 10+ yards of depth in the middle. That's a 95% strike rating for passes up the middle and that's huge for the offense. When you see that passing 10-20 yards up the middle is just as reliable as throwing short to the left or middle, or, yeah, more efficient than throwing right? Well, you have to use that to your advantage.
Strong Right - Speaking of tight ends, the discrepancy in strikes from the left to the right can possible be explained by the fact that Gronk (R24-L23), Hernandez (27-25), Daniel Fells (4-1), Kellen Winslow (2-0), and Donte Stallworth (2-0) were the only targets with more to the right and to the left. The need for a standout 3rd receiver has been around for a while, but it was partially resolved with Hernandez on the field. Without him, the Patriots have no one to fill the spot on the right side (if you view Gronk as the in-line tight end). Someone is going to have to step up.
Obvious Note - Brady behind the LoS: 86.8%; 0-10: 80.6%; 11-20: 70.5%; 21+: 52.6%. Narrative, yeah, Brady's deep ball isn't spectacular. But throwing up the seam is just as good as throwing mid-range to the sidelines (which makes sense when you think of distances). Brady threw the ball well over the head of his intended receiver 36 times on 95 attempts.
Depth Perception - An obvious curve ball is when the Patriots ran Welker up the field instead of having him pull up short, just to catch the defense off guard. Brady targeted him 21+ yards 21 times for 10 strikes. Welker accounted for 22% of the deep passes. 9 of those passes went over his head and he had a few in Zone 2 that he couldn't pull down. Adding Danny Amendola's extra height should make this wrinkle even more dangerous.
Ol' Reliable - Kind of understated this whole off-season is the loss of Danny Woodhead. I believe that Shane Vereen will prove to be a more-than-capable replacement, but Woodhead's impact needs to be highlighted. I mapped 54 targets and 42 strikes, primarily within 10 yards of the LoS (51 targets). Two of those targets went over his head (which Vereen should be able to help with), but his chemistry with Brady led to a 77.8% strike rate, good for third best behind Gronkowski and Hernandez. His value shone in the dump off passes in the middle of the field traffic- can Vereen show that same wiggle to create yardage?
The Loss - Another narrative that has been much overdone is the loss of Welker, Lloyd, Hernandez and, to a lesser extent, Woodhead. Where will this impact the Patriots the most? Below is a table of the differential between Brady and the four departed players.
The Patriots are going to be hurt trying to replace Woodhead behind the line of scrimmage as the four players provide a 13.8% premium on passes to the right, behind the line of scrimmage. Oddly enough, these players have actually reduced Brady's accuracy on passes L10-20, M10-20, and M20+, and I feel like that has to do with lack of overthrown passes to Welker/Woodhead, as well as the sideline passes to Lloyd.
The Patriots will absolutely miss having the depth of so many top tier targets, but to point out any one player as "the biggest loss" might be more difficult that originally imagined.
We'll delve deeper into the game-by-game numbers, but straight upfront, it's clear that the Patriots offense is slightly more predictable than previously imagined (although whether that makes it easier to defend is another question). Brady clearly favored the left side of the field and struggled when asked to throw to the right. His deep ball up the middle was better than expected, although it could explain why he struggles against defenses that feature a ball hawking single deep safety.
The loss of Hernandez clearly weakens the right side of the field, the loss of Lloyd hurts the mid range, and the loss of Woodhead hurts the short yardage, as expected. The loss of Welker, though, could bear improvement on the mid-range passes as Amendola is more able to catch those passes down the field.
Going forward, what interests you? What do you want me to look at? What do you want to see?