UCLA Football: An introduction to UCLA play call/outcome analysis

PASADENA, CA - SEPTEMBER 01: Head coach Chip Kelly of the UCLA Bruins during a timeout against the Cincinnati Bearcats during the fourth quarter at Rose Bowl on September 1, 2018 in Pasadena, California. (Photo by Harry How/Getty Images)
PASADENA, CA - SEPTEMBER 01: Head coach Chip Kelly of the UCLA Bruins during a timeout against the Cincinnati Bearcats during the fourth quarter at Rose Bowl on September 1, 2018 in Pasadena, California. (Photo by Harry How/Getty Images) /
facebooktwitterreddit

This new GJB Play Call/Outcome series will take a deep dive into Chip Kelley’s UCLA football offense. We will look at how game situation and pre-snap details inform post-snap play calls (such as run/pass balance) and results (primarily success rate and Yards-Per-Play / YPP).

Throughout the season, I’ll be keeping a detailed log of every UCLA football offensive play. This log will be updated every week, and publicly available via a Google doc at this location (bookmark it!).

RELATED: UCLA vs. Cal – Gameday Info – TV, Radio, Live Stream, Stats, Odds, and More

My hope is that our readers and fellow fans will help us dig into the data to find tendencies what’s working, and what’s not working for the “not-so-Blurry” offense. You can catch up on the week to week analysis so far this season on my twitter account @orz_bftb. The first follow-up to this introduction will be a breakdown of the upcoming game at Cal.

Here’s a sample slice of data from the UCLA vs Washington game; you can see that one TE (11 personnel) in the game was a huge passing tendency, while zero (10) or two TE (12) in the game was a similarly massive run tell.  Running from 12 is something Chip has been doing all season, but that much running from 10 was a significant break from prior weeks.  They got good YPP production running the ball with 10 personnel (probably because the defense was expecting pass), but were most successful overall with that mainstay 11 personnel passing game.

Related Story. GJB talks to California Golden Blogs. light

Personnel

We use a simple and fairly standard two digit annotation for the offensive personnel on the field for any given play. After the quarterback and offensive line, there are 5 skill players left to fill out the huddle. Those skill players come from running backs, tight ends, and wide receivers. We identify the personnel with the first digit for the number of running backs and the second digit for the number of tight ends, where the unwritten number of wide receivers is the balance of the 5 skill spots.

So far this year, 11 personnel (1 RB, 1 TE, 3 WR) has been the foundation of Chip’s offense, having been used on 65% of plays over the first 5 games. UCLA has used 10, 11, 12, 13, 21, 22, and 02 through the first five games.  There aren’t many reasonable personnel groupings that might still show the first time (maybe 01 or 23?) the way this offense is organized.  The reason personnel is important comes from Bill Connelly (here he’s arguing for less variance in personnel usage):

"Defensive coordinators have to make their decisions based on your personnel, not your formation. Don’t give them anything to work with."

The more versatile your individual players are, the more you can use the same personnel to be less predictable as an offense. The UCLA game vs Washington (where they minimized substitutions) was the best example of Chip Kelly keeping it simple with personnel so far this season. If you go to a live college (or other level) football game, a lot of times you’ll see an assistant on the current defense’s sidelines holding up cards showing the offense’s personnel. This is helping the defensive coordinator make his calls and gives the defensive players on the field some cues to follow the game plan.

Related Story. Know Your Opponent: UCLA Football vs. California 2018. light

Formation

Formation terminology can vary widely; we’re going to break up formation tracking into what the backfield (QB and RB) looks like, and what the players out wide (generally WR and TE) are doing. We will also pick a somewhat arbitrary set of terms and try to just stick with it.

While personnel tracking pays close attention the difference between tight ends and wide receivers; formation tracking will pay attention to how the players are lining up (we don’t care if it’s Michael Ezeike or Caleb Wilson in the slot). I’m heavily inspired by Fishduck in how I’m keeping track of the backfield. While seven years have passed since his infamous Oregon blur tutorial videos, most of the concepts are still fairly valid (but maybe less prominent) today.  They have only used a single “ace” backfield formation when operating from under center so far.

Backfield FormationShotgun:
Behind – (AKA Pistol) RB directly behind QB (traditionally sets up IZR)
Behind (Offset I) – RB directly behind QB with TE next to QB
Next – RB next to QB, no horizontal space (traditionally sets up OZR)
Hybrid – RB next to and behind QB (blends Behind and Next)
Empty – QB alone in shotgun
Max Protect – 2 skill players next to / behind QB in backfield)
Under Center:
Pro Ace – A single RB lines up deep behind QB
Receiver FormationSingle = 1 WR in game
Base = 1 WR on each side
Double = 2 WR on one side, no WR on the other side
Spread = 2 WR on one side, 1 WR on the other side
Trips = 3 WR on one side (includes no or 1 WR on the other side)
2×2 = 2 WR on each side
3×2 = 3 WR on one side, 2 other

Success Rate

I believe success rate originated from Football Outsiders, but these two articles by Bill Connelly were my introduction to the concept. Successful plays are where the offense gets 50% of to-go yards on 1st, 70% of to-go yards on 2nd, and 100% of to-go yards on 3rd/4th. Consistently successful plays will keep you out of obvious passing situations (see below). Sometimes we will look at YPP on successful plays to see how explosive the successes are.

Passing Downs

Passing downs are a pre-snap situation defined as 2nd and greater than or equal to 8 yards to-go or 3rd/4th and greater than or equal to 5 yards to-go. I’ll give credit to these two articles for my understanding of the idea, but it seems like a fairly open source concept. Success Rate and Passing situations are somewhat correlated.  Success rates are much lower in obvious passing situations.

Related Story. UCLA's Time of possession, offensive balance, and tempo. light

UCLA Offense in 2018

This chart shows the game to game variations of a handful of summary indicators for the UCLA offense so far this season. While almost every UCLA fan correctly came away from the Washington game feeling upbeat about the offensive progress, there’s no overwhelming game-over-game constantly improving trend.

The item closest to a linear trend is the steady decline of run/pass balance (towards passing) since Oklahoma. Passing Success rate took a huge dump vs Colorado and a huge leap forward vs Washington. I wouldn’t call that the evolution of DTR until we get out of small sample sizes. YPP was flat (well below the national average of 5.73) all season until this most recent Washington game. Overall Success Rate and Run success rate are all over the place. It’s interesting that they stayed away from obvious passing situations the most vs Fresno St; the offense was mostly sunk by turnovers that game.  The way all the indicators started out clustered in the upper 30’s in week one (Cincinnati) is a little creepy. We will need to see steady improvement in these indicators to declare Chip’s offense as being on it’s way to consistently better productivity.

Related Story. Sure there is progress, but UCLA still has to be cautious. light

Google Doc Glossary

Here’s a detailed reference for the spreadsheet we’re going to use.  Hopefully, readers can help me crack that code for what’s going on every time Chip uses 12 personnel in the 2nd quarter while trailing on the 50-yard line with Kazmeir Allen in the game lined up in Trips pre-snap snap motion!

WeekGame number in a given season
W/LWhether this game was a Win or a Loss
GTGY for Goal to Go, 2 for Two Minute Drill
LeadWhether the team was (T)ied (Leading) or (B)ehind at the time of the snap
Quarter Which quarter of each game the play took place in.
PossessionPossesion # within a single game
Down1st/2nd/3rd/4th down
DistanceDistance to go
Ball OnI count own territory as 1-50 and opposing territory as 51-99
Passing DownPassing downs are a pre-snap situation defined as 2nd and greater than or equal to 8 yards to-go or 3rd/4th and greater than or equal to 5 yards to-go.
RunningBackThe tailback in the game (if using 2RB, the one in the more conventional RB alignment)
QBThe QB in the game (the DTR show so far this year)
Personnel1st digit # RBs, 2nd digit # of TEs.
Rbs + TEs + WRs adds up to five; so 5 minus both digits tells you the # of WRs.
Backfield FormationThis is the QB/RB alignment, VERY heavily informed by fishduck oregon blur tutorials.
Shotgun:
Behind – (AKA Pistol) RB directly behind QB (traditionally sets up IZR)
Behind (Offset I) – RB directly behind QB with TE next to QB
Next – RB next to QB, no horizontal space (traditionally sets up OZR)
Hybrid – RB next to and behind QB (blends Behind and Next)
Empty – QB alone in shotgun
Max Protect – 2 skill players next to / behind QB in backfield)
Under Center:
Pro Ace – A single RB lines up deep beind QB
Receiver FormationSingle = 1 WR in game
Base = 1 WR on each side
Double = 2 WR on one side, no WR on other side
Spread = 2 WR on one side, 1 WR on other side
Trips = 3 WR on one side (includes no or 1 WR on other side)
2×2 = 2 WR on each side
3×2 = 3 WR on one side, 2 other
WR Heavy SideWhether receiver heavy side is (L)eft, (R)ight, or (B)alanced.
TE Heavy SideIf TEs are attached to Oline, whether receiver heavy side is (L)eft, (R)ight, or (B)alanced. Blank means no attached TEs.
MotionWhether there is pre-snap motion. I don’t count formation shifts here.
Play CallIRZ for a run between tackles after a mesh*
OZR for a run outside the tackles after a mesh*
Run (L)eft/(M)iddle/(R)ight for conventional handoffs
Option for conventional QB option run play
Pass for called pass plays (includes sacks)Sometimes will annotate RPO when I suspect there was a  run/pass option, but this is impossible to determine accurately.
Play Fake?If there is a run fake on a pass play; Y for obvious fake action, ZR for meshes that could have been read keep or designed play action.
Carrier/RecWhich skill played wound up with the ball
ResultYards gained or lost
1st?Whether the play resulted in a first down
EventIn-completions, Sacks, Turnovers noted here.
TD?Whether the play resulted in a touchdown
Successful?Successful plays are where they got 50% of to-go yards on 1st, 70% of to-go yards on 2nd, and 100% of to-go yards on 3rd/4th.

Next. The 30 Greatest UCLA Football Players of All-Time. dark

*I suspect (and Chip has outright stated) there are mesh looking handoffs that have no read, meaning they are designed runs with a zone read veneer.  I’m calling these IZR and OZR as a simplification, because it’s impossible to really know without being inside the coaching staff.