Josh Bynes’ return to the 2019 Ravens with LJ Fort (1 week later) precipitated a 12-game win streak when a set of flailing ILBs (Onwuasor, Young, and Board) received reduced roles. Bynes went on to play more than 50% of the team’s defensive snaps from week 5 on and the Ravens were one of the most successful defensive teams for the rest of the season.
Josh’s impact was apparent in first game back with the Ravens (10/6/19 at Pittsburgh) when he picked off an overthrown pass and made 5 tackles. Did history repeat itself? Let’s review:
- (Q1, 14:17): He assisted Campbell to tackle RB Austin Ekeler RR-1
- (Q1, 13:43): He flashed between Herbert and RB Ekeler to force an overthrow on a swing pass right despite ATS
- (Q2, 12:52): He stood up TE Jared Cook to blow up RB Ekeler RM2
- (Q2, 12:28): On 3rd/3, he dislodged the ball from juggling WR Mike Williams for a PD to end the drive
- (Q2, 6:19): He diagnosed the pitch left and evaded pulling LT Rashawn Slater to tackle RB Joshua Kelley RL-2 with Averett
- (Q3, 6:44): He had a diving PD/near INT in coverage of WR Williams 17 yards  negated by 12-men on defense as Queen was slow to get off the field
- (Q3, 5:02): He was underneath in coverage of TE Cook 6 yards  as Herbert was pressured and thew high
- (Q4, 4:08): He tackled FB Gabe Nabers PM6 (5 + 1) 
Bynes played every snap that was not an obvious passing situation (see packages below).
Neither Queen’s thigh injury, nor the fact he was slow to get off the field when the Ravens were flagged for 12 of defense (Q3, 6:44), helped his cause. Humphrey, Board, and Young all converged to celebrate what they thought was a drive-ending PD by Bynes. They clearly appreciated the awareness, physicality, and effort he brings to the job. The shot then went to Harbaugh, clearly seething on the sideline as he talked through the headset with clenched teeth.
What’s Next at ILB?
I’ll offer my sense as to the short-term future of the ILBs:
- Bynes will play all downs which are not obvious passing situations until he proves himself incapable
- Since the bar has been set low by the other ILBs, I don’t expect that to happen from a single bad game
- Board will continue to play in obvious passing situations as dictated by either down-and-distance or score
- Most of Board’s snaps will come in dime packages with only 1 ILB
- Harrison and Queen will play rotationally with Bynes as the 2nd ILB, allowing each to master a smaller set of responsibilities
- Kristian Welch remains an excellent special-teams player, which creates a question about the number of ILB activations the Ravens can afford for any game
- Because Queen is the only ILB who does not play special teams, his game day roster spot is in jeopardy
- I do not believe it is reasonable, prudent, or likely given Harbaugh’s background that Queen will take a ST role
Versus the Chargers, the Ravens dressed just 4 DL, 4 CBs (excluding Stephens), and 4 OLBs, all positions where the Ravens have frequently activated 5. I don’t believe they’ll continue to take in-game risks for other position groups to activate all 5 ILBs. I think this means Patrick Queen is likely to be inactive vs the Bengals.
The Chargers entered the game 10/11 on 4th down conversions. What happened on the closest decisions on Sunday?
- (Q2, 12:22) Staley sent on the punt team on 4th and 3 from the LAC 32 with the score 14-0.
- (Q2, 9:13) Staley went for it on 4th and 3 from the LAC 39 with the score still 14-0. This was a little strange with the field position and game situation so similar. Humphrey dislodged the pass intended for Mike Williams and the Ravens took over. Tucker kicked a 52-yard FG 4 plays later.
- (Q3, 5:58) Staley went for it on 4th and 1 at the LAC 19 down 24-6. Chris Board delivered a QH and Herbert overthrew Josh Palmer on the left sideline. Tucker kicked a 39-yard FG 4 plays later.
- (Q4, 5:28) With the score 34-6, Herbert ran a 2-yard QB sneak on 4th and 1 at the Ravens 37.
- (Q4, 2:12) 5 plays later, and on the same drive, Staley went for it on 4th and 13 at the Ravens 28. Herbert overthrew Guyton on the right sideline as Tavon Young rushed unblocked for pressure.
It’s fascinating that this set of aggressive, yet mathematically viable decisions came from a coach who the previous week had failed to have his QB kneel 3 times then kick for a near-certain win (Austin Ekeler was dragged into the end zone and the Browns got a chance to score a winning TD in the last 90 seconds).
I see at least 2 additional career risks for coaches who lean on mathematical models for in-game strategy:
- Such models may require more complexity than the data allow. While a coaching decision may be adding expected wins under league-average conditions, those rarely exist, so the marginal decisions will come under greater scrutiny of all Monday Morning QBs from fans to media to owners.
- If you’re selling yourself as an expert in game-strategy decisions, any hole in your theory (such as the unwanted Ekeler TD vs the Browns) will create additional scrutiny. While perhaps unfair, I think a coach who had simply ordered a power run for the TD in the Ekeler case would take less heat than Staley did for failing to execute a kneel-kneel-FG sequence.
John Harbaugh is an aggressive 4th-down coach, but he’s no longer the most extreme animal of the group.
By the way, I looked up names of groups of animals. Those closely related to our favorite football team…a murder of crows, a conspiracy of ravens, and the alternate term, an unkindness of ravens. If you knew all 3, feel free to award yourself 10 Scooby-Doo bonus points.
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.
The Chargers ran 53 such snaps.
Base (7): The Ravens used their base 3-4 defense with 2 ILB and 2 OLB on first-down plays and a single 4th and 1 (Q4, 5:28). After (Q2, 10:07), Martindale did not deploy it again until the final Chargers drive. Josh Bynes was on the field for every base snap. 7 plays, 44 yards, 6.3 YPP.
Standard Nickel (32): Martindale again used the standard nickel (including 2 down linemen, 2 OLB, and 2 ILB) as the most common response to 11 and 12 personnel. 32 plays, 122 yards, 3.8 YPP. Josh Bynes was on the field for every nickel snap vs the Chargers, paired with one of Board, Harrison, or Queen.
Jumbo NIckel (1): The Ravens again turned to jumbo nickel for 1 play versus the Chargers (4th and 3, Q2, 9:13). The version used on Sunday included 3 down linemen, 1 OLB, and 2 ILB. Jimmy Smith was the 5th DB. 1 plays, 0 yards.
Rush Dime (10): The Ravens lined up with 1 DL, 3 OLB, 1 ILB, and 6 DB on 10 occasions. Chris Board was the sole ILB on each of these plays. These were all obvious passing situations by down-and distance (3/11, 3/10, 3/5, 3/10, 3/7, 2/21, 3/26, 3/6, 4/13) or game situation (1/10, Q4, 11:55). 10 plays, 17 yards, 1.7 YPP.
Racecar Dime (1): Martindale inserted his 4 OLB package with 1 DL, 0 ILB, and 6 DBs (Q2, 5:16). 1 play, 8 yards, 8.0 YPP.
Quarter (1): The Ravens played their first snap of the season with 7 DBs on the last play of the first half (Q2, 0:06) and allowed a stat-padding, 16-yard screen pass to Ekeler. 1 play, 16 yards, 16.0 YPP.
Goal Line (1): The Ravens played a single snap of jumbo on 1st and 1 when the Chargers scored their only TD (Q2, 4:08). This included 4 down linemen, 2 OLB, and 3 ILB, and just 2 DBs (Humphrey and Clark). 1 play, 1 yard, 1.0 YPP.
Martindale employed another conservative pass rush versus Herbert, who had been one of the fastest QBs in terms of time to throw. The Ravens were effective in almost every split in route to a 4.4 YPP effort by the Chargers.
For the game, Herbert had ATS on 15 of 41 drop backs (37%) resulting in a pass or sack. He also had 9 ball out quick (BOQ) instances and the Ravens generated 17 pressure events within 3 seconds. By pass-rush situation:
ATS: 11/15 for 109 yards (7.3 YPP)
BOQ: 7/9 for 48 yards (5.3 YPP), 1 INT
Pressured: 4/15 for 25 net yards (1.5 YPP), 2 sacks for 13
From the perspective of the defense, each of these results is excellent for the circumstances.
Summarizing by number of pass rushers:
3: 4 plays, 20 yards, 5.0 Ypp, 1 sack
4: 25 plays, 115 yards, 4.6 YPP, 1 TO
5: 8 plays, 25 yards, 3.1 YPP, 1 sack
6+: 4 plays, 22 yards, 5.5 YPP
Total: 41 plays, 182 yards, 4.4 YPP, 2 sacks, 1 TO
The Ravens rushed 5+ on 12 of 41 drop backs (29%) on which the Chargers averaged 3.9 YPP with 1 sack.
Martindale used just 3 individual blitzes the entire game (.07 per pass play) from off the LoS. Herbert completed just 1 of the 3 plays for 1 yard. This is the lowest total I can recall in the Martindale era, which I hope to be able to fact check this week.
The Ravens used 0 stunts as they did against the Chiefs. This is indicative of a moratorium on stunting from Martindale because some stunts are called by the players.
The Ravens showed simulated pressure just 4 times as well where 2 or 3 dropped from the LoS.
I define a deceptive pass rush as incorporating 2 or more of the above elements (off ball blitzes, stunts, and simulated pressures). The Ravens ran 0 deceptive rushes for the game for just the second time in the Martindale era (week 2 vs KC).
Justin Houston was by far the Ravens most effective pass rusher with contributions to 4 pressures plus a sack and a holding call drawn on Rashawn Slater.
Other Individual Notes
- Calais Campbell had another fine game and played 41 of 53 snaps. He had contributions to 4 run tackles, each of which was a defensive win (RR-1, RM2, RM1, RM-2) and had 1 pressure.
- Chuck Clark delivered one of the most satisfying plays in the form of a leaping PD in the backfield (Q1, 12:00) to end the first Chargers drive. He also made contributions to 2 defensive wins vs the run.
- Brandon Stephens made 3 tackles, 2 of which were defensive wins, including a tackle for loss (Q4, 11:55).
- (Q1, 4:09): On 3rd/5, he rushed unblocked off the ORS for a drive-ending S-6
- (Q2, 9:18): On 3rd/3, he rushed unblocked off the OLS for a QH as Herbert threw behind RB Ekeler and Humphrey nearly intercepted
- (Q2, 4:39): Herbert threw for TE Cook 15 yards  but Elliott undercut for the interception
- (Q3, 4:54): On 3rd/10, he dislodged the ball from the bobbling Cook 10 yards  to force a punt
- (Q4, 13:43): With ATS, Herbert threw to WR Keenan Allen PL21 (21 +0)  in front of Elliott
Deshon returned in style with 3 drive-ending plays and another 3rd-down QH which led to the 4th-and-3 failure.
- (Q1, 12:04): He bulled TE Stephen Anderson for pressure as Herbert threw high for WR Williams
- (Q2, 10:07): He rushed unblocked off the ORS to flush Herbert left where he threw incomplete for WR Williams
- (Q2, 6:59): He beat RT Storm Norton outside combined with Oweh’s pressure to flush Herbert for RR10. Norton was not flagged for an obvious hold.
- (Q2, 5:16): On 3rd/17, he bulled RT Norton for pressure as Herbert threw PL8 to WR Joshua Palmer
- (Q2, 0:06): He bulled RT Norton for pressure as Herbert threw to RB Ekeler PR16 (-5 + 21)  to end the half
- (Q3, 6:39): On 3rd/2, he stood up pulling RG Michael Schofield to blow up RB Ekeler RR1
- (Q4, 10:41): On 3rd/16, he beat LT Slater inside and drew holding to negate PM13
- (Q4, 2:53): On 3rd/6, he beat RT Norton outside for a fast S-7 on a 3-man pass rush
Justin accumulated his pass rush success on just 25 plays that resulted in a pass or sack and he dropped to cover on 3 of those plays.
- (Q2, 10:44): With ATS, Herbert threw to RB Ekeler PM18 (3 + 15)  with Humphrey trailing
- (Q2, 9:18): On 3rd/3, he tipped and nearly intercepted RB Ekeler’s deflection
- (Q2, 9:13): On 4th/3, he dislodged the ball from WR Williams 8 yards  to end the drive
- (Q3, 7:04): He made a quick tackle on WR Williams PL1 (-2 + 3) 
- (Q3, 5:58): On 4th/1, he had close coverage of WR Palmer 6 yards  as Herbert threw incomplete under pressure from Board
- (Q4, 12:37): He had tight coverage of WR Palmer 5 yards  as WR Williams PD at LoS fell at his feet
- (Q4, 6:02): He was close as Herbert missed a back shoulder throw for WR Palmer 10 yards  under pressure from Oweh
Marlon was in coverage on 2 drive-ending plays.
- Deshon Elliott
- Justin Houston
- Marlon Humphrey
Honorable mention to Josh Bynes, Calais Campbell, Chuck Clark, and Brandon Stephens.
Always appreciate reading your work and listening to your pod. Thanks Ken.
Thanks for reading and taking the time to comment, Bruce.
I’ve never been a fan of Patrick Queen’s play, but do you think that a potential decision to deactivate Queen over Welch represents an overvaluation of Special Teams on the Ravens’ part. It seems like they have enough core STers to sacrifice one for the sake of keeping Queen on the field in a limited role so he can gain confidence and hopefully develop with experience.
Very fair question. I do agree that having the decision based on special teams is letting the tail wag the dog. I also agree that the only way Queen is going to develop is with live fire. It’s strange that having an on-field handler like Bynes seems to have done more for him in 1 week than Rob Ryan has since training camp.
“If you’re selling yourself as an expert in game-strategy decisions, any hole in your theory (such as the unwanted Ekeler TD vs the Browns) will create additional scrutiny.” this reminds me of something from Nassim Nicholas Taleb in this essay (https://www.edge.org/conversation/nassim_nicholas_taleb-the-fourth-quadrant-a-map-of-the-limits-of-statistics) and developed in his other books like The Black Swan. Black Swans are events for which there is little or no precedent, which makes them highly unpredictable. Pulling an opposing running back into your endzone to score a TD on yourself is a Black Swan event. basically what happened (according to Taleb’s framework) is that there were some people who were very good at using “analytics” based on all that is known from the history of the game, but because the models are based on what is known, they don’t properly account for what is unknowable (basically Black Swans or football plays with little or no precedent).
another way to say it is that a model based on every known play in the history of the game might tell you to run there, but if you improve that model by including risk assessment for the possibility of unknowable events, your model might have told you to kneel kneel kick instead.