clock menu more-arrow no yes

Filed under:

Predict the 2018 Cal Football Season, Results: BOWL-BOUND!

New, 23 comments

Our ascension under Justin Wilcox continues (at least, according to our expectations).

Washington v California Photo by Jonathan Ferrey/Getty Images

Optimism has been in short supply around here for most of the past decade. That might be changing. Although the Bears fell slightly short of our expectations of 5.76 wins last season, they were only 10 points short of an 8-4 finish. Thanks to continuity among the staff and at several key positions, this could be a breakout year for the Bears. It certainly doesn’t diminish our hopes that several of our in-conference rivals look uncertain due to coaching changes and replacement of some top talent. We should be better this year, but will we be favored-in-10-games better? Apparently!

A few weeks ago I asked you all to predict the Bears’ chances of defeating each opponent this season. And you were an enthusiastic lot, as we had well over 200 responses. I gathered all the data, printed it out on several sheets of paper, and mailed it to Evans Hall. A few days later this table appeared in my mailbox. Let us bask in the optimism.

Opponent Avg. Win Chance Standard Deviation
North Carolina 75.8% 12.6
at BYU 69.1% 16.4
Idaho State 94.6% 7.1
Oregon 50.8% 16.5
at Arizona 51.6% 15.9
UCLA 60.8% 14.9
at Oregon State 75.2% 14.9
Washington 32.9% 18.4
at Washington State 54.2% 16.9
at USC 30.8% 18.8
LSJU 58.3% 23.1
Colorado 69.3% 14.7
Total 7.23 wins 1.28

From left to right we have our opponent, the average of everyone’s predicted chances of a Cal victory, and the standard deviation, which captures the amount of uncertainty in our predictions. Lower standard deviations (Idaho State) indicate that our predictions were mostly clustered together, while larger standard deviations (Big Game) indicate that our predictions were more spread out. Of course, the Big Game always has the largest standard deviation because many members of our community have a proud tradition of giving the Bears a 100% chance of reclaiming the Axe.

What is most remarkable about our predictions is that we favor the Bears in a whopping ten games. Many of those are in the 50%-60% toss-up range, so we may not necessarily win all of them (for the love of Oski, let us not get bogged down in our annual comment thread talking about why a team favored 51%-49% should not be expected to win every single time). But this is nonetheless a remarkable turnaround after we only favored the Bears in three games last year.

We begin the season with a fairly favorable September, as we give the Bears strong odds of starting the season 3-0. After three toss-up games (crucially, two are at home), we have a favorable matchup against the lowly Beavers (not coached by Beau Baldwin, thankfully!). Then begins the toughest three-game stretch of the season: hosting the likely Pac-12 title favorite, visiting a Cougs team that probably won’t turn the ball over 7 times again, and then on the road again to face USC. We conclude with a favorable (!) Big Game and a very favorable game against a Buffs team that may not capture the magic of Mac’s 10-win season a couple years ago.

In sum, we have five favorable games, five toss-ups, and two difficult matchups. Adding up all those predictions gets us a projected total of 7.23 wins. We may not win all 10 in which we favor the Bears, but I can abide 7 regular season wins.

I use the plots below to illustrate the distribution of our predictions. The distributions are a bit lumpy because we’re not all a bunch of random-number-generating robots (well, some of us are, and they have great advice on how to earn $82/hour working from home 20 hours per week). Instead, we’re predisposed to submit predictions divisible by 5 and 10. I’ve smoothed much of that out, but some lumpiness nonetheless persists. Oski smiles upon all of you who contributed to the bump near the 100% win probability for the Big Game.

Win probabilities of home games.

Obviously Idaho State is our most favorable home game, and the predictions largely center around 95%. The North Carolina rematch is the next-likeliest win, followed by Colorado. UCLA sits just off-center, ahead of a near-tie between Oregon and the Lobsterbacks. Even if you ignore the Big Game bump at 100%, the curve the Big Game is slightly to the right of the curve for Oregon. We fared much better against the Lobsterbacks than against the Ducks last season, but I’m surprised to see that we tend to be more optimistic about the Big Game than about Oregon. Perhaps I’m underrating Oregon due to their coaching staff shakeup in the offseason. Surprising virtually no one, we’re least optimistic about the UW game.

Next we have the road games.

Win probabilities of road games.

We have such nice, vibrant colors in our chart for the home games, but our road opponents have a bunch of drab, ugly colors. So I apologize that our opponents are all so eye-bleachingly ugly. Not so ugly are our win chances against Oregon State, where we favor the Bears heavily. We also heavily favor the Bears against the Stormin’ Mormons of BYU. Arizona and Wazzu both look like toss-ups, while USC sits there on the left side of the chart, like it has every single time we’ve run these predictions. Will USC ever be terrible? The answer is clearly yes, but that may not necessarily extend to the gridiron.

Simulating the Season

Next I sat one thousand monkeys at one thousand abacuses (abaci?) and had them painstakingly simulate the season. I simulated the entire season one million times to determine how likely we were, based on all these game-by-game predictions, to finish with 7 wins, 8 wins, 6 wins, TWELVE WINS!, and so on. I begin by taking one prediction at random for the UNC game and using that to determine a win or a loss. If I draw a 70% prediction, we’ll have a 70% chance of winning and a 30% chance of losing. I’ll draw a win or a loss based on those probabilities and move on to the next game. I then draw a BYU prediction at random and use the same procedure to determine a win or a loss. After going through all 12 games I count up the number of wins and do this 999,999 more times. In the table below I demonstrate how likely we are to finish with every possible win total.

Wins Probability
0 0.00%
1 0.00%
2 0.12%
3 0.79%
4 3.36%
5 9.41%
6 18.05%
7 24.26%
8 22.59%
9 14.24%
10 5.71%
11 1.32%
12 0.13%

Not surprisingly, 7 wins is the likeliest outcome. 8 wins is the next-likeliest outcome, followed by 6 wins. Based on these simulations we only have a 1-in-7 chance of missing a bowl game. We’re as likely to win 9 games! I’ve plotted the numbers above in the chart below.

That’s right, those are the official Yale Blue and California Gold colors.

AWARDS!

Finally, it’s time to hand out some awards per our usual tradition. I still haven’t settled on a name for the optimist and pessimist awards, so I’ll keep rotating through various award names that fit each category. I really liked the Sonny Delight and Sonny Yikes categories during the Dykes era. Oh well, I’ll trade endless optimism for a couple good award names.

We’ll start with the optimists, who had the most optimistic predictions.

Sirmon’s Disciples

Name Wins
1. rahimftd 12.00
1. RememberTheCalamo 12.00
1. OCBear1983 12.00
1. Def_not_a_fake_account 12.00
5. Old Bear 71 11.34
6. The Ghost of Joe Roth 11.00
7. Moosehead 10.98
8. Oski Disciple 10.94
9. yanny 10.31
10. jpl 10.10

Only four perfect scores? We must be getting more realistic. Since we favor the Bears in 10 games, several of these predictions towards the bottom of the list don’t seem too lofty. Also, I have decided that RememberTheCalamo is my new favorite username on this site.

Not So Greatwood

Next we highlight the most pessimistic predictions.

Name Wins
1. Uthaithani 4.00
2. 99bottles 4.45
3. Chartreuse Bear 4.48
4. Md3 4.59
5. Not creating an account for this 4.70
6. Snort weasel 4.90
7. Ryke99 5.22
8. O.Overall 5.30
9. Gobears02 5.40
10. BacherBear 5.50

When only 6 of our 200+ predictions are worse than the previous season, we must be feeling prettay, prettay, prettay good about the upcoming season.

The Voice of Reason

Finally, I highlight those whose predictions were closest to the community average. Despite all the sunshine pumping and optimism, you all remain grounded in some semblance of reality.

Name Standard Deviation
1. joltimpact .041
2. jrrad1980 .043
3. dinan3 .046
4. joshwright1291 .050
5. sup_doe_library .056
6. Blue tide bear .056
7. Novarem .056
8. wesbutera .058
9. minesweeper .061
10. Alex Ghenis .062

And that reality, clearly, is 7 wins!

Thanks to all who participated. We’ll run another round of these season predictions after we get another look at the team during Fall Camp. Until then, keep pumping that sunshine!