Can you fly, Bobby?
The startling change in Oregon’s offensive efficiency between the first and second halves of the Alamo Bowl can be explained by four factors:
For the first two factors, I can’t detect any real pattern or strategic change Oregon or Oklahoma made at halftime – the Ducks were effectively 6 for 13 on these plays, about what both teams’ regular seasons would predict. I think it’s just a coincidence that the bad perimeter blocks and breaking the pocket at inopportune times stacked up so neatly in the first half, and good blocks and wise scrambles stacked up in the second.
The second two factors, however, were definitely strategic. Both the rushing offense and attacking the sideline (Oklahoma’s secondary plays with bizarre leverage outside the numbers) were effective at the same rates throughout the game; the halftime adjustments were to lean into the former to sustain drives and to take the latter deeper for big gains. Oklahoma didn’t change its scheme or defensive personnel at any point in the game, and I couldn’t spot any real difference in effort or fatigue problems as the Sooners maintained a solid rotation in their defensive front despite their opt-outs.
Given that all four of those changes involve the passing offense, it’s no surprise that there are such stark differences in the efficiency and explosiveness numbers between halves. Excluding garbage time (the final drives of both halves), Oregon was successful on 18 designed passing plays vs 22 failed ones given the down & distance, or 45.0%. In the first half that was 6 vs 16, or 27.3%, and in the second it flipped to 12 vs 6, or 66.7%. Overall, Oregon averaged 8.4 yards per passing attempt and 15% of dropbacks resulted in a gain of 15+ yards. In the first half it was 4.8 YPA and 5% explosive, in the second it was 12.7 YPA and 28% explosive. If it were sustainable, the second-half version of the passing offense was hitting championship-caliber numbers.
Since they were such a dramatic departure for #13 QB A. Brown, who hadn’t been able to hit the deep ball since he transferred from Boston College where he was doing so regularly, let’s look at all six of those attempts in this game:
(Reminder – after pressing play, you can use the left button to slow any video to 1⁄4 or 1⁄2 speed)
I have no explanation for why Brown was suddenly able to connect on these deep strikes. He was playing with almost an entirely replaced wide receiver corps, with only #14 WR Hutson as a season-long starter. True freshmen #11 WR Franklin and #10 WR Thornton got most of the action, with one catch for the third true freshman #15 WR Brevard and five targets for 2018 4-star recruit #88 WR Crocker (who essentially hadn’t seen the field in a Ducks uniform until December).
While I think those five wideouts are very talented, I don’t have any good data to suggest that they were outplaying the original group who were unavailable in this game – compared to the season-long charts, they were getting open and making catches at the same rates. Instead, I think the two major effects that show up in the data are Oklahoma’s peculiar secondary and Brown simply placing the ball where it needed to be. The former I documented pretty extensively in my preview, the latter is a complete mystery which will likely never be solved.
The offensive line performed at about the same effectiveness level throughout the game. There were no rotations in this game, instead playing #56 LT Bass, #77 LG Moore, #78 C Forsyth, #74 RG S. Jones, and #71 RT Aumavae-Laulu on every snap. As usual for this season, they were more effective run blocking than pass blocking. Oklahoma’s opt-outs affected them most significantly on the edges, and both Oregon tackles graded out pretty well on my tally sheet in pass-pro. But as expected, the deep and talented interior of the Sooners’ d-line replaced their opt-outs with no real dip in performance, and the issues I’ve been seeing all season long with the Ducks’ interior o-line in pass-pro showed up again. I continue to think it’s been a very odd choice to have Moore and Jones play guard when both are clearly built like tackles, and to be concerned that Forsyth isn’t able to hold his ground against powerful DTs.
Here’s a representative sample of failed passing plays:
While I think strategic choices were most responsible for the upswing in passing effectiveness in the second half, those outside passes were available all game long, and Brown was making better choices about throws over the middle and when to scramble after halftime. Some examples:
By contrast to the passing game, Oregon’s rushing attack was uniformly effective throughout the game. They had 18 successful designed runs vs 10 failed ones, or 64.3% efficiency, with identical 9 vs 5 splits between halves. The Ducks averaged 6.3 yards per carry on such runs and 25% of them went for 10+ yards. Those numbers all line up with their season long averages and exceed them for explosive rush count, and constituted a significant dip in performance for the Sooners.
I wasn’t able to detect significantly different performance quality for Oklahoma up front or with their inside linebackers defending the run, and I don’t think their opt-outs really affected them in that regard because their rotation was extensive throughout the year and everyone who played in this game had plenty of reps going into the bowl. I don’t have any individual amongst the Sooners’ replacement defenders who graded out significantly differently than his teammates in this game, and while there are two players they were missing who did provide extra value vs their backups during the regular season, both are edge players – OLB Nik Bonitto and DE Isaiah Thomas – and those are far more important in the structure of Oklahoma’s pass defense than its rush defense.
There was one very interesting change, which is that against inside zone runs during the regular season Oklahoma’s unblocked end always crashed inside on the back, and in the bowl game he always crashed outside on the QB. In terms of rush performance that didn’t change things much, since the whole point of a zone read play is that it doesn’t matter which potential ballcarrier the unblocked end attacks – the other guy is going to get the ball as long as it’s an obvious crash, and the Sooners provide that no matter what. What it did effectively stop was the RPO – Brown was being attacked every time, and had few opportunities to extract any advantage from those offensive play structures.
Simply put, I think Oregon has a very effective ground game and they should have been leaning into it, and away from RPO plays and inconsistent passing, much earlier in the contest. Some examples:
Oregon’s unsuccessful rushing plays are fairly unremarkable – Oklahoma simply defeated Oregon’s blocks a certain number of times. I don’t have any real difference in performance between the various backs on my tally sheet, only that designed QB runs weren’t working because of the Sooners’ surprise switch in whom they attacked. Some examples:
The biggest discrepancy in performance compared to these two teams’ numbers during their regular seasons is that Oklahoma’s rushing was significantly more explosive in the bowl game, and arguably that’s where the game was won for the Sooners.
In terms of per-play rushing efficiency, the numbers weren’t too far off of what the regular season figures would have predicted: Oregon successfully defended 17 designed runs vs 18 failures, or 48.6%. That’s almost exactly what Oklahoma normally posts, and only about 2 percentage points worse than Oregon does – a single play going one yard shorter would have made it perfectly match.
But Oklahoma averaged 7.8 yards per carry on those runs, with 26% of them going for 10+ yards. That’s 1.6 YPC longer and 11 percentage points more explosive than the Sooners normally gain, and 3.8 YPC longer and 15 percentage points more explosive than the Ducks normally allow.
I didn’t note any real differences in Oklahoma’s rushing scheme and their formational tendencies lined up almost exactly with their regular season counts. They used the same blocking personnel in terms of offensive linemen and tight ends as expected, with the minor difference that they didn’t have their third most used TE. Towards the end of the game they put in their third-string running back whom I hadn’t seen much of during the season, though I don’t think that affected things at all as his numbers lined up with the second-string back’s. All their situational tendencies were the same with one exception – they ran the ball about 70% of the time in 2nd & medium or long, way up from their 40% tendency during the regular season – clearly pressing their advantage in explosive rushing.
The major difference was available personnel – Oregon’s defensive front was significantly depleted with injuries and opt-outs. The Ducks both had more guys out than the Sooners’ defense did and their backups were mostly playing out of their natural positions (including #58 OG Powers-Johnson switching squads to play nose tackle). I’ve heard some, including the broadcast commentators, dismiss those absences by saying “both sides had guys out,” which strikes me as quite a false equivalence.
I can confirm that every player mentioned in the above link as being absent in practices did not in fact play in the bowl game, which includes 17 defensive players of whom 12 play in the defensive front. While I think two of the absent defensive backs would have been behind the DBs who played, I believe each one of the 12 missing defensive front players would have started ahead of those who played. There was one type of Sooners’ rushing playcall the Ducks didn’t defend properly that looked to me like a bad job of film study by the coaches, but the rest of Oregon’s failures in rush defense pretty clearly came down to being short-handed:
When the key personnel on a play were starters for Oregon they held up much better, and there were several encouraging plays from the backups as well:
Oklahoma’s passing offense, and Oregon’s defense against it, was much more in line with both teams’ regular season numbers. In terms of per-play efficiency it was perfectly even, 15 successes vs 15 failures. The Sooners averaged 8.9 yards per passing attempt, and 16% of dropbacks went for 15+ yards. Oklahoma hit one very long pass when #15 DB B. Williams, in his first game back after missing the last nine with a leg injury, got beat off the break; if that 55-yarder is excluded from the dataset then every efficiency and explosive passing number for both teams falls perfectly in line with what their regular seasons would have predicted.
With the absence of transformational player #5 OLB Thibodeaux, Oregon’s pass rush wasn’t as dramatic and Oklahoma didn’t devote any additional blocking resources to stop it as other opposing offenses had with Thibodeaux in. Outside of that, however, the pass rush did a fairly good job and deep ball coverage was better than I was expecting with both starting corners out:
The Sooners only hit two deep-shot passes during the bowl, the aforementioned stutter-go and a post route against freshman #28 CB Dickerson (he pulled the receiver’s helmet off after the catch resulting in the touchdown being called back but a 15-yard penalty assessed, however it had already gone 20 yards through the air). Dickerson didn’t have any other notable plays one way or the other so it’s tough to assess him since cornerbacks get the worst camera angles on the broadcast.
Otherwise, while Oregon gave up a few more longer passing plays, they all represented screen passes or short routes in which linebacker or safety was out of position and allowed it to go bigger than it should have, rather than a DB getting beat on a long pass. While Williams returned, #7 DB Stephens was still out and so #32 DB Happle played most snaps, and the issues I’ve been noting for a while now regarding his athleticism and the club on his hand interfering with tackling persisted in this game. The next most experienced backup in the secondary was #11 CB Bridges; I thought he did a decent job in coverage but my tally sheet has seven dings on it for him on poor tackling or leverage. I have very few notes on #8 CB Manning; I’d been critical of the freshman in previous games and there’s one play in the bowl where I think he wasn’t quick enough recovering on a comeback route, but generally when a CB is never on camera that’s a good sign.
Here’s a representative sample of Oregon’s failed pass defenses:
In my preview of Oklahoma before the bowl game, I expected the Sooners to keep the basic scheme and tendencies on offense that they did during the regular season despite the departure of that offense’s architect, and I think it played out as described. I’m a little surprised that so few passes went to their outside receivers (I thought more highly of them than apparently many Oklahoma fans did), but just like rushing more often on 2nd & long, I suspect that was simply pressing their massive advantage running the ball against Oregon’s depleted defensive front. Their backs performed as advertised, but I think their tight ends and slot receiver were more productive than I had described them. I thought the Sooners’ backup center and starting right guard would be real vulnerabilities, and they did grade out the lowest on my tally sheet in the bowl again, contributing to Oregon’s relatively effective pass rush up the middle.
The expectation that Oklahoma would stick with the same scheme and tendencies on defense despite their DC’s departure also held up. The personnel predictions in my preview about who would replace their opt-outs in the front, and where they’d handle them well vs where they wouldn’t were also borne out. I was pretty shocked that they reversed their longstanding practice of crashing the back on zone read runs, though in hindsight the fact that this takes away two of the three options in Oregon’s staple RPO makes a lot of sense and now I’m wondering why more defenses hadn’t tried that. I think the biggest coup in my preview was describing the Sooners’ DBs, both their problems tackling and the way they defend the sideline, and if anything I’m surprised the Ducks took advantage of it so well in the second half. I hedged my bets in this regard in the preview, thinking that perhaps Oklahoma’s corners had gotten it sorted by this point after so much rotation and personnel experimentation during the first three-quarters of the regular season, but that didn’t show up at all so perhaps I should have been bolder.