Peters Wins/Peters Loses
Marcus Peters had 2 juxtaposed plays which underscore not only his own ability to gamble effectively, but how NFL defenses should approach risk. Let’s review:
- (Q2, 3:55): Peters guessed on the out route to Erickson and undercut it for a room-service interception by the right sideline and ran 89 yards for the pick-6. Gannon mentioned on air that Peters baited Finley on the throw, but this was far different from the interception versus Russell Wilson where he backed off the receiver at the snap to induce the throw before breaking.
- (Q2, 1:56): On 1st and 10 at the Bal 25, Peters was beaten by Auden Tate on a double move along the right sideline. Effectively, Peters undercut, expecting the same out route, but Tate continued down the right sideline and was wide open at the 5 for a walk-in TD, but Finley overthrew him under pressure from Brandon Williams.
There are multiple ways to look at the risk/reward of these plays and I am a proponent of expected win models over expected points. However, each of these occurred during the first half when expected points are a reasonable proxy for expected wins.
There are lots of smaller, debatable assumptions that are folded into the following analysis, but we’re trying to take broad strokes to estimate how often Peters would need to succeed in order to make gambling on this undercut worthwhile.
Using the pro-football-reference.com expected points model, the first play gained the Ravens 11.58 points (the Bengals had a +4.58 expectation which immediately changed to -7). To be fair, that’s what actually happened, not what might have happened with Peters anticipating correctly, so I’m going to arbitrarily cut the positive result by 20% to account for times where Peters drops the INT or somehow fails to run it back for a TD. So we’ll call that positive outcome +9.26 EP.
Ideally, we’d evaluate this same play for what would have happened on a double move, but since we don’t have good EP numbers for that, we’re going to use the 2nd play as a proxy.
Evaluating the 2nd play is a little trickier, because it resulted in a forced error that cost the Bengals .55 points (3.91 EP before, 3.36 EP after). However, let’s say this play should have resulted in a TD 80% of the time and some sort of forced or unforced incompletion 20% of the time. The cost of Peters’ lost gamble can then be estimated as;
.8 X (7.00 – 3.91) + .2 X (3.36 – 3.91) = 2.47 – 0.11 = 2.36 EP which is expressed from the Bengals point of view or -2.36 EP for the Ravens.
So, how often does Peters need to be correct to make the gamble worthwhile?
For that, we need to solve for p (probability Peters wins the gamble) in the equation:
p X 9.26 + (1 – p) X (-2.36) = 0, which reduces to
11.62 X p = 2.36
p = .203
In words, Peters needed to have only a 20% chance to guess correctly in order make that gamble worthwhile.
When all players with 15+ INTs are adjusted to the NFL’s 1945-2016 average, Peters has the highest interception rate in NFL history (8.64 per 16 games played), nearly 1 full interception more than Ed Reed (2nd, 7.67). If you’ve been reading here recently, you know those 2 tower above all others as the league’s greatest interceptors with Asante Samuel (6.48) and Darren Sharper (5.64) a distant 3rdand 4th respectively. Thanks to @Yoshi2052 for extending the study down to 15 INTs.
Pierce’s Injury Hampers Defense
Michael Pierce hobbled off after 3 snaps with an ankle injury and did not return. Without him the Ravens run defense allowed 12 runs of 5+ yards the rest of the day and 157 total rushing yards (3.9 YPC).
On one hand, the Bengals’ rushing was of small consolation in a game they lost 49-13. This was reminiscent of the game the Bengals finally broke the Raves 50-game streak without a 100-yard rusher. That day (12/23/01), Corey Dillon rushed for 127 yards, but the Ravens shut the Bengals out for the 3rd straight year in Baltimore, 16-0.
More significantly, a long-term injury to Pierce could hamstring the Ravens defense in upcoming games against the Texans, Rams, 49ers, Bills, and Browns, all of whom have formidable rushing attacks.
Note: all snap totals exclude penalties resulting in no play, kneels, spikes, and specials team plays resulting in a run or pass. As such, they will be lower than other published totals.
Base (3): The Bengals lost 2 yards total on these snaps, which were all defensive wins.
Standard Nickel (29): The Ravens most commonly called package was a standard 4-2-5 nickel with 3 CBs. The Ravens allowed 165 yards on these plays (5.7 YPP), including Peters’ pick-6.
3-3-5 Nickel (12): Martindale used this package more than any other game this season in what appears to have been an attempt to slow down the Bengals run game. The Ravens held the Bengals to 38 total yards on these plays (3.2 YPP), including just 13 yards on 8 runs (1.6 YPC).
Dime (26): The Ravens continued to run dime primarily with Carr inserted as deep, cover-2 safety and Clark in Levine’s typical dime spot. Effectively, Levine has lost his spot defensively to Carr with the return of Jimmy Smith. They allowed 116 yards in the dime (4.5 YPP).
Quarter (2): Martindale called the quarter package for just 2 plays. The first was the sack/fumble by Ricard (Q3, 2:24). The second was an incomplete pass to deny 3rd and 14 (Q4, 2:40).
Racecar: The Ravens continued to use a modified version of the racecar package with all 4 active 4 OLBs (Bowser, Ferguson, Judon, Ward), 1 DL, and no ILBs as a common dime look. On 16 such non-penalty snaps vs the Bengals, they allowed 81 yards (5.1 YPP). They had a 17th such snap negated by roughing the passer on Brandon Williams.
Pass Rush Gambles Pay Off
Martindale harassed Finley with numbers and scheme and the Bengals averaged just 4.7 yards per pass play.
Finley had ATS on just 8 of 32 drop backs (25%). In addition to 6 QHs, they often flushed Finley left, the much worse direction for a right-handed QB.
By number of rushers:
3 or 4: 13 plays, 93 yards (7.2 YPP)
5: 12 plays, 20 yards (1.7 YPP), including the sack/fumble TD
6+: 7 plays, 36 yards (5.1 YPP), including Peters pick-6 on a 7-man rush
In terms of elements of deception, the Ravens used 13 individual blitzes, ran 6 stunts, and had 3 drops of 2+ men. That deception is at a similar per-drop-back rate to that used against Tom Brady the previous week. However, they used fewer combinations of deceptive elements and had only 5 deceptive pass rushes as I define them (2+ elements).
Finley did not complete a pass of more than 13 air yards and the Bengals had just 2 offensive plays over 16 yards. The Bengals 5 longest pass plays were to Boyd (24, 13 + 11 YAC), Mixon (23, -1 + 24 YAC), Tate (15, 6 + 9 YAC), Mixon (14, 0 + 14 YAC), and Eifert (14, 12+ 2 YAC).
- Patrick Onwuasor (21 snaps) played poorly. He was juked by Mixon by the right sideline on a play with 24 YAC (Q2, 6:07). He is also was either blocked effectively or unable to quickly bring down the ball carrier on 4 runs I’ll leave for Game Pass review (Q1, 3:08; Q2, 8:39; Q2, 3:42; Q2, 3:06). After having difficulty matching up with the size the Bengals present with their run game, he played just 1 snap in the 3rd quarter and 4 more in the 4th.
- The Ravens had 12 on the field after they called a timeout (Q2, 2:03). Based on other packages played during the game, it appears either Sieler or Carr was on the field when he should not have been.
- Patrick Ricard had 3 significant contributions among 12 defensive snaps. He beat a double team from LG Price and C Hopkins to tackle Bernard for no gain (Q3, 10:24). He again beat a double team by RG Miller and C Hopkins when both politely disengaged so he could deliver his first career sack (Q3, 2:24). Ricard dislodged the football which Bowser returned 33-yards for his first career TD (there may be a fight over that football!). Patrick had what was effectively another sack when he tackled the scrambling Finley for a 1-yard gain to deny 3rd and 7 (Q4, 13:59).
Brandon Williams (56 snaps) was both overworked and outstanding in Pierce’s absence. He finished with contributions to 7 tackles for gains of 4, 0, 0, 3, 1, 3, and 1. Of those, 5 were defensive wins by the Football Outsiders definition. Further detailing:
- (Q2, 7:25): On 3rd/1, he penetrated past RG Miller to tackle Mixon RM0.
- (Q2, 1:56): He bulled C Hopkins for pressure as Finley overthrew the wide-open Tate for a touchdown 18 yards .
- (Q3, 8:08): He cleaned up Sieler’s redirection of Mixon to tackle RL0.
- (Q4, 8:11): He bulled C Hopkins for pressure.
- (Q4, 6:18): He received an absurd roughing the passer flag on Finley that negated a PD by Humphrey. We’ve seen some awful RTP flags his season, but I am confident Williams will not be fined by the league for this one.
Marlon Humphrey (72 snaps) shadowed Tyler Boyd wherever he lined up. As I scored it, he was targeted 8 times with 4 receptions for 31 yards, 2 PDs and no TDs allowed. Reviewing my notes on him:
- (Q1, 1:34): On 3rd/11, with 8-man pass rush Finley threw to Boyd PM13 (10 + 3 YAC)  with Marlon covering.
- (Q2, 14:24): Finley threw for Tate 6 yards . Humphrey came off Boyd as throw was low and incomplete with Fort underneath.
- (Q2, 14:19): Finley threw to Boyd PR3 (3 + 0)  and Marlon tackled quickly.
- (Q3, 11:11): He knocked away Finley’s throw to Boyd 2 yards .
- (Q3, 9:43): He blitzed off the slot to pressure Finley on 4th/8, pass was complete for 0 yards, then fumbled.
- (Q3, 6:51): On 3rd/2, Finley threw for Boyd 25 yards  incomplete with tight coverage from Humphrey.
- (Q3, 6:47): On 4th/2, Finley again threw for Boyd 6 yards  incomplete, as Marlon used the right sideline well.
- (Q3, 1:59): He was blocked by RT Hart as part of Mixon’s RR9.
- (Q3, 0:50): Finley threw for Morgan 11 yards  and Humphrey punched it free for a PD.
- (Q4, 8:11): He was soft as Finley threw to Boyd PR6 (2 + 4) .
- (Q4, 6:18): He stripped Boyd for a PD 20 yards , negated by RTP flag on Williams.
- (Q4, 0:29): He was soft as Finley threw to Morgan PR9 (7 + 2) .
- Marcus Peters
- Brandon Williams
- Marlon Humphrey
Honorable mentions were earned by Patrick Ricard, Matthew Judon, and Chuck Clark.