Welcome back to The Good, the Bad, and the Rockfights, our weekly exercise in condensing the previous game’s data into one of four types of Cal football games. Under our new data overlords at BCF Toys, all our measures capture some aspect of how well the offense moves the ball or how well the defense stops the ball from moving. These measures tend to capture: productivity of moving the ball (i.e. total distance covered), efficiency in moving the ball, and ability to turn drives into scoring attempts. In contrast to the subjective player ratings that we used with PFF’s data, there is a comforting level of objectivity in our new data. Drives are generally easily quantifiable across multiple dimensions, and assessing any single drive is fairly straightforward…or so I thought until Syracuse’s first drive of the fourth quarter.
That fourth-quarter drive was a masterpiece in utter chaos. 16 plays, 1 interception called back by penalty, 1 lost fumble called back by penalty, 3 near-interceptions, and a drive-ending field goal that bounced off the upright. It was spectacular. Glorious. A master-class in absurdism that I have not seen since the Cheez It Bowl. And as much as I appreciate the objectivity of our new drive-based data, no set of measures could adequately capture what we witnessed on that drive. That drive was neither math nor numbers; that drive was art.
Fortunately, there were 22 other drives that could be more adequately represented with BCF Toys’ measures, and our goal is to make sense of the rest of the game (although absurdity reigns because the outcome is something we have never seen in this series, so maybe sensemaking is futile here).
Ratings Comparison
We begin with the offense, whose ratings we illustrate below with boxplots. Recall that boxplots capture the spread of the data, with the box capturing the spread between the 25th and 75th percentiles while the horizontal line represents the midpoint of the data.
Offense
Across the board it was a bad day for the offense. Game Rating and Efficiency capture very similar measures: a starting-position-adjusted measure of expected scoring. The primary difference is that Game Rating is opponent-adjusted. These both tell us that Cal had a woeful scoring rate when accounting for where their drives started. Points per drive is somehow even worse, a reflection of Cal’s 10 points on 11 drives. Scoring rate captures the proportion of drives ending in TDs or field goal attempts (whether a miss or a make), so Cal’s missed field goal gives that measure a slightly higher rating than points per drive. Yards per play was the highest rating of the game, thanks to a handful of explosive plays. But Cal’s inability to accumulate the yards needed to score the ball led to a woeful Available Yards rating. That is six measures and too many words to say that Saturday’s offense was a disaster.
Defense
Defensively, things were better, although results were mixed. The Game Rating is perfectly middling, while Defensive Efficiency was pretty good (although that score is somewhat inflated by the poor quality of Syracuse’s offense). Points per drive was quite good, aided by the defense’s ability to force Syracuse into field goal attempts that they happily missed. Success rate suffered a bit, as Cal allowed Syracuse many scoring opportunities. And Yards Per Play and Available Yards were roughly middling, indicating that Cal allowed Syracuse to move the ball moderately well. In sum, that is a strong profile of a bend-but-don’t-break defense: give up yards, but limit scoring.
Special Teams
This was generally a favorable set of ratings for the Bears, although this is likely a reflection of the ratings punishing Syracuse rather than rewarding Cal.
Ratings Clusters
I took our seventeen ratings and fed them into a clustering algorithm. Based on the combination of woeful offense, decent defense, and reasonable special teams, I was expecting this one to fall into the realm of the Rockfights. I was wrong. I was so very wrong.
The Bad! The Bad?! I don’t disagree with the classification, because this was not a good performance. But I am absolutely astonished because this is the first time we have ever seen a win in The Bad. Cal has 56 games classified in The Bad—all losses. It makes sense that the vast majority of games in The Bad will be losses, but it took 56 Bad games to eke out a win? Either Cal is horrifically unlucky, or utterly unable to overcome poor play. Or Both.
One caveat to note is that these clusters are ever-evolving groups. As more and more data enter the plot, the boundaries of the clusters tend to change slightly over time. Since this game is on the edge of The Bad, it may flip into a Rockfight at some point in the future. But for now, take a moment to bask in the highly improbable.
Deja Vu
I’m adding a new section to these pieces: after each game I will go through every single game in our database and rank them from most similar to least similar to the game we just watched (I calculate similarity based on Euclidean Distance between this game and all previous games in the data). And then we’ll highlight the recent game’s doppelganger.
Despite 20 seasons of data, we do not have to go very far back to find the game most similar to this one. It’s another midday ACC matchup featuring a blue and orange team: Cal vs. Virginia in 2025. As you might expect, that one was very similar to this: Cal was outgained by about 200 yards, Cal fell behind by double-digits early in the game before repeatedly pulling within three points, Cal also went 3-for-12 on third down, and the Bears lost the time-of-possession battle by about 10 minutes. That game also featured many, many punts and some excruciating stretches of football by both teams. One fifteen-minute stretch featured this woeful stretch of possessions: punt, punt, punt, missed FG, punt, turnover-on-downs, punt, punt. And, critically, this game was decided by turnovers as Cal tossed a red zone pick in the final minute of the 3rd and the game-sealing pick-six in the final minute of the game. Those made the difference as Cal lost a frustrating, we-could-have-won-that kind of game that Syracuse fans have been lamenting for the past several days. Better them than us.
Finally, a quick programming note: FEI, whose data underlies BCF Toys’ ratings, does not evalute FBS-FCS games. So we will not have any data from the Wagner game. Not that the game would be very interesting, because this team is bad. I’m talking, bad bad. See you in two weeks after Clemson. Go Bears.
BK97 for all the outstanding data analytics you've done over the years for the CGB / W4C community, those first two paragraphs are some of the best prose ever written here. Bravo!
Fantastic analysis. In the article you mentioned BCF TOYS doesn't track FCS data. One could argue the Cal/'Cuse game was an FCS game, not a Power 4 game, with'Cuse wanting to lose more than the GBs.