In the previous post we saw that despite the upheaval on the QB position and the consequent lack of consistency in play and execution, the game-time decisions to run or pass remained consistent with the norm and game script. I was bemoaning the lack of passing on 1st down due to the generally higher returns on a pass than run, however, as we will here see: there are good reasons why the run/pass selection was mundane.
Part II: It is the work of an enemy defense.
It is the core tenet of sports analytics to make the most optimal decision from an outcome perspective:
- In football, on average, an offense should keep passing the ball until the expected yards gained per pass. (Per last article: Pac-12 and Cal left a lot of yards on the field by running the ball too much).
- In basketball, most of the time a 3 point shot is more efficient than a mid-range 2 point jumper since the difference in the expected point value justifies it.
However, like the 3 point shot (take a more sure 2 point shot when down by 1 late in the game even if the 3 point’s expected value is higher), decisions to run or pass are made with a team’s performance and how the enemy worked against the pass and the run.
In the case of Cal, the distribution of yards it has had in the air and on the ground indicates that we did not have this yardage advantage through the air.
Here we can see the distribution of yards gained per each type of play: run and pass. In Cal’s case we can see a pretty predictable pattern: a lot of the passes are clustered around the 0 mark due to blown screens/incompletions, while the run generally gained some yards. Which reinforces the notion that Justin Wilcox wants to run the ball to gain some ground on 1st down than risk a bigger play.
Zooming in on the -10 to 20 yard range we can still see that Cal was more likely to churn out 2 to 7 yard gains on the ground than it did through the air. It is the same across the 4 downs, the decision to run more on any down was dictated by the desire to matriculate the ball down the field.
Getting the 2nd and 8/5 is a beneficial outcome since it brings the Cal offense into a zone where it can get 5 yards, it just can’t get 10 after an incomplete on 1st down.
One sign of a good offensive coordinator, staff, and team is the ability to adjust quarter by quarter. The outcome variable here is the yardage gained, and as we can see there are is a significant dip in incompletions in the 3rd quarter and increase in effectiveness on passing plays, however, this reverts in the 4th leaving the run being more effective between the 0 and 10 yard range.
The run game comparatively was more reliable and consistent throughout the game. With the slight bump in the 4th quarter it maintained a consistent production.
By Opponent and Down
Now that we looked at the “opponent”, or how the Cal opponents defended the pass/run. For passing we will exclude incompetents to see how well each completion did (we already know that there were struggles with accuracy by Garbers/McIlwain but the large sample of in completions would make the charts much harder to process).
Even after factoring out incompletions, Cal passing offense failed to yield anything worth noting besides the very noise 4th down (low sample causing high variance), and a poor showing on 2nd down, failing to produce above average past the 10 yard mark).
Here we can see that Cal didn’t produce anything worth writing home about on besides mediocre consistency across all downs. It was simply average against opponents.
One blimp is the 3rd down run where the share of very short runs (probably stuffed) was lower than average showing that on 3rd and short the offense was not very effective.
Here we look at only opponents, showing the distribution of actual runs/passes over the average opponent’s defensive performance.
This is ugly, even accounting for the fact that deep plays are hard to perform and excluding inompletions, each Cal completion was very short, clustered around the 10, and fell short of the general distribution of each opponent’s average defensive performances. One can look at the Colorado game to see the plodding nature of the passing game.
I think a lot of the inefficiency on second down can be seen on a game to game basis is clustered around the 10 yard completions, and the dearth of completions past the 15+ yard mark.
The run game acquitted itself down the middle of the road. With a handful of rushes of the +20 yard nature, and a worrying amount of 2-5 yard plays and nothing besides that. This could speak to the lack of ability of the QBs/RBs to generate yards past the LOS. Across the opponents there wasn’t a game where Cal broke its own distribution of runs by yards.
Taking in both of the distributions, even on the down by down or opponent specific distributions we can see that the pass did not yield the added yards passes an average offense would yield. On the extreme: in the USC game there was only 1 10+ yard completion... Overall, this can explain the lack of a desire to favor the pass: something we knew but now have proof.
Conclusion: The World Cal Offense Lived in 2018 Was a Football Run Era One.
“There are three things that can happen when you throw a pass, and two of them are bad.” - HC Darrell Howell
Cal offense lived by this decades old adage, especially with the lack of explosive passing plays or consistent completions past the 5+ yard range makes a strong point for OC Baldwin and HC Wilcox favoring the run. Without factoring in turnovers, and only focusing on completions, the running offense was a safer call on most downs. Which explains the previous posts’ run/pass tendency.
When an offense is stymied by its own failures and the work of an enemy defense it can only shift its run-pass tendency to maximize its outcomes.
This is why I call it a bizarre season: Cal going against all analytical knowledge had good reason, besides scheme (ex. triple option), to run the ball more often than the national average.
Hopefully, with the coming of an improved passing game with a QB who was given 1st team snaps for nearly a year throughout bowl game prep, spring and fall camps, a more seasoned OL, and a renewed focus on being decent we can expect more production across the 10+ yard range of completions. Hopefully this will let Cal enter the realm of 21st century offensive run/pass ratios.
All in all we don’t need the offense to be Jared Goff/Davis Webb level great, we don’t even need them to be good, we just need them to be mediocre and the defense can carry the team to 9 wins.