The UCLA football team had a tremendous offensive effort against USC. How did it happen? Go Joe Bruin looks at it in the weekly “Play calling/outcome analysis” segment.
The UCLA football team took out the USC Trojans, 34-27. It was a methodical performance backed by Joshua Kelley‘s 289 rushing yards.
RELATED: Joshua Kelley’s 289 Yards Leads UCLA Past USC
So how did it all happen? Let us find out…
A deliberate yet unpredictable running game
The run-pass balance for UCLA against USC this week was 64% in favor of the run. This was the highest run-favored balance of the season, other than the Cal game (that was 71% run). Chip Kelly used a dizzying array of variety in the number of ways he attacked the ground game and attacked the edges far more than has been typical for UCLA this year.
This week the Outside Zone Read was a huge feature after being all but ignored the entire season. OZRs were used 13 times in the three nonconference games, 12 times before USC in Pac-12 play (that’s across 7 games), and now 12 times specifically against USC. The USC results tell us exactly why Chip Kelly may be hesitant to consistently rely upon the current incarnation of the OZR play; it’s a real feast or famine call. When compared to the other basic run constructions (IZR and Non-Mesh), OZR had by far the biggest upside on production (10.0 YPP), but was the lowest performing in terms of success rate (42%) against USC. It succeeds less often (because it’s easy for defenses to diagnose), but when it works, it’s a blur style demoralizing chunk play.
Outside Zone Read was better going left than to the right by a factor of four both in terms of production (15.8 YPP left) and success rate (67% success left). I don’t think this means you can easily conclude that LT Andrew James and LG Michael Alves are outperforming RG Christaphany Murray and RT Jake Burton. I counted at least three flavors of O-line blocking assignments on OZRs vs USC: an unbalanced sweep (more on that later), a balanced sweep (C and a G pulling), and a traditional fishduck-defined blur approach (no pulling, just the line moving together). There are a ton of moving parts in Chip’s OZR scheme, and right now it’s working when the ball moves left.
We can defend the right side of the line with their success rate on non-mesh runs to the right side at 67%. It’s worth noting the non-mesh right run production (6.7 YPP) was a fair bit worse than the non-mesh left run production (14.6 YPP), but I’ll take success rate every time. Non-mesh runs up the middle (2.9 YPP / 33% success) and IZRs (3.7 YPP / 54% success) (also generally up the middle) are a necessary threat with lesser production.
There was also a lot of “arc” motion by Theo Howard, Christian Pabico, and Devin Asiasi that kept the defense moving and having to communicate as the ball was being snapped.
Unbalanced Line Formations
This week Chip resurrected the unbalanced line formation that was a source of excitement in Week 1 vs Cincinnati (it got Kaz Allen his first long TD run), but has been rarely (if at all?) used since. Above you’ll see a really poorly constructed cartoon of what we’re talking about (can anyone recommend football specific software for this kind of thing)? When they run an OZR sweep out of this alignment, the two Guards and the ‘extra’ Tackle block inside, while the C TE, and conventionally lined up Tackle pull to the play side. The Guards have it so easy here, while the Tackles have to think and cover a lot of ground. This unbalanced OZR sweep only worked the first time it was sprung on USC (for 61 yards), and was 0/5 the rest of the game. Below is a list of all plays vs USC that used the unbalanced line formation.
Unbalanced line was used always to run, mostly on first down in the first half, and was constructed to use 12 personnel out of Shotgun or with 13 personnel from under center. The big drawback to this approach is the screaming announcement as to which way the play is headed. UCLA had to drop it from the play call sheet at halftime because USC adjusted after the first surprise (only a 25% success rate in the game).
The good news is that Chip was able to pull off OZRs with both unbalanced (17% success rate unbalanced OZR) and balanced (67% success rate balanced OZR) lines while also using the unbalanced line for something other than OZRs. This kind of variable simply wasn’t possible earlier in the year. I’m sure Chip wants to keep using this unbalanced line and is looking for ways to diversify (misdirected runs to the weak side or play action passing).
YPP Chart
There were rolling average spikes at the start of 1st and 4th Quarters. This is a lot like last week’s ASU chart (consistency wise), just better overall production. The 2nd quarter play count kept defenses fresh for a physical 2nd half. I often note the offense stumbles in YPP production after halftime, but that didn’t happen against USC. If you focus in on possessions 5 and 6, the YPP average took a hit and bottomed out for the game, but those were long sustained drives which ate up clock and resulted in a single TD. This shows how YPP isn’t a reliable indicator in a vacuum. The two outlier positive large chunk plays were runs, which is unusual for 2018 UCLA (only 5 of the 20 longest plays >=30 yards are runs so far this season).
Game Summary
Usage of Check-With-Me saw a big jump up this week (used 20 times) after almost disappearing last week (used only 6 times vs ASU), and continues to be a reliable favorable outlier (65% success rate against USC). Check-With-Me had two in-season firsts that are worth noting: they checked from shotgun into pro ace (under center), and also sideline checked from under center. Chip is trusting the players with more and more complexity now that they have built a solid foundation.
Dear Chip: full time Check-With-Me for Stanford? Is that even feasible?
It’s always great to see the success rate climb as the offense works deeper in their sets of downs (1st 46%, 2nd 54%, 3rd 57%). Run YPP production peaked early in the downs (8.2 YPP on 1st) while passing YPP was better after first down (2nd down 9.0 YPP, 3rd diwb 8.3 YPP passing). Anytime you can confidently run the ball 81% of the time on 2nd down, your offense is not falling behind schedule.
Chip was able to also confidently run on 56% of passing downs and achieve a 44% overall success rate on passing downs (national average PD success rate is around 31%). This is indicative of a team that is asserting a run identity. They were also in passing downs a season-low 24% of plays.
The offense peaked in the 1st quarter in almost every category as compared to the other quarters, and most explosive as compared to any slice of data in the game summary table (17.2 YPP on successes). Chip brought in a solid game plan for USC, took a step back after initial adjustments in the 2nd quarter, then put things back together slowly over the 2nd half. The 2nd quarter saw chip veer toward too much passing and a low run success rate. That 2nd quarter in this way felt more like the rest of the UCLA 2018 season.
Personnel and RB Alignment
Chip broadened the personnel lens this week, featuring more than 2 personnel for the first time since the Cal game. The last four games (Arizona, Utah, Oregon, Az St) relied on only 2 personnel groups.
Passing from 12 and 13 was especially bad this week, but it wasn’t for lack of trying . That lone (unsuccessful) pass play from 12 personnel was the 3rd quarter play action Speight arm punt (interception) into the end zone. Even though 12 personnel had been a 100% run tell, the USC secondary was disciplined enough to hang in there on pass coverage. The same was true on the cute statue of liberty-esque ball hiding play fake to Asiasi.
Maybe that opposing defense discipline in the Pac-12 is why Chip is able to be so run dominant from 12 (95% run) and 13 (79%) without negative repercussions of the super obvious tell. Just out of morbid curiosity, I want to see an opposing defense put 9 or 10 in the box and go cover 0 to see what would happen, and how Chip would adjust.
The fun thing to note about RB alignment this week was the 3 straight Pistol plays in a goal to go situation in the 2nd quarter. Using the pistol there makes sense, as you get the RB better momentum going straight ahead when you only need a yard or two. We hadn’t seen UCLA use the Pistol at all since the Arizona game (it was gone for 3 whole games). Like most of these kinds of observations, this could be situational (the situation they planned to use it didn’t come up) or schematic (planned to not use it based on scouting).
2 QB Personnel
I made a 100% throwaway joke on twitter after the Arizona game suggesting chip use 2 QB personnel with Speight and DTR together on the same play. Now I get to be smug about it as if it was serious commentary! Chip reads Go Joe Bruin: CONFIRMED. Here’s a snip of it.
In terms of the play log, this is just 11 personnel, Speight was technically playing running back. DTR correctly kept the ball, as USC #56 crashed Speight. If Speight got the ball, he likely would have had a run pass option with 3 WRs running routes together on the left side. The DTR keep was a power/sweep run outside to the right, which was all the offensive line was worried about. The LT, C, and RG pulled while the LG, RT, and right TE blocked inside.
One of the TV commentators said they saw more variations in practice, and I want them all. If you put DTR at RB you’d have to design the QB keep to be primarily pass oriented; this is how they wound up at the initially counter intuitive choice to play Speight at RB.
Season Long Trend Chart
We’ve already covered how well they stayed out of passing downs (season-low passing downs on 24% of plays) and how much they ran the ball (64% run plays). The run emphasis empowered passing game to have the best pass success rate of the season (56%).
More from Go Joe Bruin
- UCLA Football: It’s time for the nation to meet Dante Moore
- UCLA Football: Where are they ranked heading into week 4
- UCLA Football: Position battle breakdown for Utah showdown
- UCLA vs. Utah: Location, time, prediction, and more
- UCLA Football: Highlights from Chip Kelly’s appearance on the Jim Rome Show
The 6.8 overall YPP was also exceptional (2nd best of the season to last week, well above national average of 5.7). It’s been true for a few weeks that if you look at the chart starting at Colorado and throw out Utah (and maybe squint a little), the most important indicators are trending favorably in a sustained way. Chip is getting it all put together. The UCLA staff definitely has a much better visualization, but this is the kind of view they will take into upcoming recruiting visits
This week was a blast to log after the last few UCLA games featured nothing but 11 personnel running vanilla IZRs and a misdirection free passing attack. This week we saw Pistol, OZRs, more personnel variety, a 2QB gimmick, Arc blocking, unbalanced line with multiple callable plays, and increased Check-With-Me and pre-snap motion. It’s almost as if Chip knew this one was important for the narrative and was work-shopping/building the pieces through the middle of the PAC-12. Almost.
Dear Chip: the Sanford streak is important for the narrative!
Footnote
The google doc linked here has every UCLA offense play logged with the situation, pre-snap details, and post snap results. I like to look for how the situation and pre-snap looks inform play calling and results. In the past, I’ve done this (on staff) for the SBN NFL Chargers blog, and got a lot of great reader participation.
My hope is that readers look at it and find stuff along with me (something pre-snap is a tell, something else very successful). This post introduces the column which has context for everything in the log.