Predict the 2020 Cal Football Season: Results!
A short schedule means fewer opportunities to regress to the mean, for better or for worse.
After months of uncertainty we are finally on the verge of the 2020 Cal Football season. The uncertainty doesn’t end with the start of the season, however, as the Pac-12 is in a particular state of disarray this season. Many teams across the conference are replacing head coaches, several impact players, or both. Nevertheless, I asked you all to predict the Bears’ chances of defeating each confirmed opponent on the abbreviated schedule. Today, we’ll review the results.
2020 Preseason Predictions
In the table below I provide the Bears’ chances of winning each game, along with the standard deviation indicating how much variation or uncertainty is in those predictions. Fortunately, we only have the Bears as underdogs in one game, a clash with Oregon with potential division title implications. The rest of the games are toss-ups (UW, @ASU) or favorable matchups (@OSU, LSJU, @WSU). Adding up all those chances gives us a projected 3.9 wins, or nearly a 4-2 record.
Below I have taken the predictions for each game and plotted their distributions. This helps to provide a better visualization of how our predictions span the range from 0 to 100.
That 4-2 record hinges on winning the 3 favorable games (@OSU, @WSU, LJSU) and going 1-2 in the toss-ups (UW, @ASU, UO), or winning 2 favorable games and going 2-1 in toss-ups. An interesting feature about a shortened season is that there’s an increased chance of having some extreme outcomes across those toss-up games. Consider flipping a coin three times; getting three heads or three tails isn’t all that unlikely (1 in 8 chance). If this were a typical season, we’d have around 6 toss-up games. Flipping 6 heads or 6 tails is much, much less likely than flipping 3 (chances drop from 1 in 8 to 1 in 64). So there’s a better chance that we sweep those games this year than in a typical year. On the other side of the coin, there’s a better chance that we lose all those toss-up games. With such a short season, there’s less time to regress to the mean. So if we overperform, it’s less likely we’ll come crashing back down to earth. Or if we crater, we’re less likely to dig ourselves out of a hole. So now we have some statistical madness that may happen this season in addition to the madness induced by covid, extremely limited spring practices, and the general wackiness that is college football.
Here’s a look at how likely we are to be at each win threshold after each game. The number in each cell represents the probability to have that number of wins after each game. So after UW we have a 63% chance of being 1-0 and a 37% chance of being 0-1. After ASU it’s a 16% chance of being 0-2, a 48% chance of being 1-1, and a 36% chance of being 2-0. And so on.
We’re quite likely to be somewhere in the 3-5 win range (which is a huge spread with only 6 games), but some of those extreme outcomes I talked about earlier aren’t too unlikely. There’s a 9% chance of finishing 2-4 and a 7% chance of being 6-0. Those are obviously unlikely events, but they are still well within the realm of possibility. Fortunately, the only impossible outcome is a winless season.
In these uncertain times, we can rely on our ol’ triumvirate of awards that we carry over from CGB. We highlight the most optimistic predictions, the most pessimistic, and those closest to the community average.
I’m surprised that only 4 predictions have us above that 5-win mark.
It feels like we’re not far from the days when 3-win projections were pretty common in the Old Blues category. 3 wins in a 12-win season is pretty miserable, but 3 wins in a 6-win season isn’t too bad.
Voice of Reason
Finally, we have the predictions that were closest to the community’s score for each game. Spots 2-4 are pretty close together, but Sailor Berry was miles ahead of us. I don’t know how anyone can be that reasonable given all the insanity that we’ve seen in 2020 (and I’m writing this before whatever insanity unfolds on Election Day!)
Thanks to all of you for participating and for sticking with us through this long, long, looooong offseason. We certainly picked a sub-optimal time to start a new site, didn’t we?