Predicting the 2021 Cal Football Season: Results!
Expectations climb towards 8 wins...can the program keep up?
Seven wins is a safe number of wins to predict in a given football season. Slightly better than bowl eligibility and a few wins below a standout season, a seven-win season sits at a comfortable level of achievement for any middle-of-the-pack program. For those reasons we have predicted seven regular-season wins in four of the last five seasons (2019, 2018, 2016, and 2015). Both the 2018 and 2019 seasons achieved those seven wins, so we’re on a seemingly upward trajectory (last year’s bizarro season notwithstanding). As a result, Cal fans have begun to set their sights slightly higher this season. That’s right, we’re looking at 8 wins! Well, 7.67 wins rounded up, but that’s still a sizeable leap over recent predictions.
A couple weeks ago I asked you all to rate the Bears’ chances of defeating each opponent on the 2021 schedule. Thanks to your contributions, we have a wealth of data to give us a sense of our expectations for the 2021 season. I have the game-by-game results in the table below. We keep track of the average win chance and the standard deviation, which captures how much variation is in our predictions (a small SD indicates convergence of our predictions while a large SD suggests much greater uncertainty in our predictions).
Cal Football Season Predictions
What stands out most to me in the above table is that we favor the Bears in nine games. Of course, being favored to win does not necessarily mean the Bears will win. Likewise, being underdogs doesn’t guarantee a loss in the three games in which we favor the opponent (at UW, at Oregon, vs. USC). So the Bears will probably lose a game or two (or three) where we favored them, but they have a strong chance of winning one of the games where they are underdogs.
Another detail that stands out in the table is that we only have four games in the toss-up range (40%-60%). Thanks to the team’s offensive struggles in recent years, we have tended to put many games per season in that toss-up range. Now we have bumped several games into a much more comfortable 70+ win chance (including four conference games!). And the toughest games are now in the 30s-40s rather than the 20s. Unfortunately, the uncertainty on offense seems to put a ceiling on our win chances in several games. Other than an FCS matchup, it’s hard to give Cal an 80+% chance of winning when it’s not clear if the offense will be able to muster 30 points.
To give us a sense of the spread of our predictions in each game, I’ve plotted the distribution for each game below. The x-axis represents the win chance and the y-axis shows the number of predictions at that given chance.
Interestingly, Oregon is the game with the lowest win chance but the predictions are pretty well spread out (in fact, it has the highest SD in the schedule). Perhaps last year’s win is still fresh in the minds of some of us while others fear that their ever-increasing talent will pose quite a challenge this year.
In keeping with our proud tradition, we observe a nice little bump at the 100% mark for the Big Game.
Simulating the 2021 Season
Obviously the team can’t win 7.67 games; that total has to fall to 7 or 8 (or 9 or 10 or 11 or 12—I’d gladly accept any of those). So to see how likely it is that the team wins 7 or 8 or 12 games, I used our predictions to simulate the season.
The simulation process is simple. I take the first game of the season, draw a prediction at random and use that to determine whether Cal wins or loses. If I draw a 75% for the first game, I then draw a 1 (win) or 0 (loss) based on that percentage. Once the outcome is decided, I move to the next game and repeat the process. And so on, through the rest of the season. I then repeat the process 999,999 more times until we have 1 million simulated seasons.
I’ve plotted the results below using the site’s new team-appropriate colors. Each bar refers to a specific number of wins and the y-axis shows us the probability of achieving that number of wins.
Although the simulations help provide more insight into how likely the Bears are to finish with a 7- or 8-win season, their true value lies in showing the team’s likely records after each game.
In the chart below I list the probability of finishing with each possible record after each game. The number in the square reflects the probability, and darker squares indicate likelier outcomes. We start the season with a 100% chance of having a 0-0 record before the Nevada game. After Nevada we have a 75% chance of being 1-0 and a 25% chance of being 0-1. After the TCU game we have three possibilities: 12% chance of being 0-2, 48% chance of being 1-1, and a 41% chance of being 2-0.
The highest number indicates the likeliest outcome. So after Sac St. we’re likeliest to be 2-1; 3-1 after UW, 4-1 after Wazzu, 4-2 after Oregon, 5-2 after Colorado, 5-3 after Oregon State, and 6-3 with bowl eligibility after Arizona. After that, it’s 6-4 after USC, 7-4 after the Big Game, and 8-4 after the season finale in Pasadena. While likeliest, an 8-4 finish is still only a 25% chance. The Bears have over a 30% chance of finishing with 9+ wins and, mercifully, only an 8% chance of missing bowl eligibility. Anywhere in the range of 6 to 9 wins seems likely, as the team has an 80% chance of falling in that window.
Ever since the CGB days we’ve been handing out awards along with these season predictions. We highlight the most optimistic predictions, most pessimistic predictions, and those most closely aligned with the community’s predictions.
First, the eternal optimists.
100% win chance for all 12 games. I like it!
Next, the lowest predictions of the bunch.
It’s always a good sign when nearly all of the worst predictions are still on the precipice of bowl eligibility.
The Voice of Reason
Finally, those whose predictions were closest to the average predictions for each game.
Thanks again to everyone who participated. Go Bears! And may we finally end the painful losing streak to Nevada...