Syracuse Sabermetrics: When Cal Can Get Back in This

This was born because I was bored and I find this stuff fascinating. When can we actually say "Cal can get back in this," and not have it be annoying or a meme? Especially when very early guessing work far less accurate than predicting the weather says that the Orange and Golden Bears may run into each other one more time. So for all of the announcers who may really want Cal to get back into a game with the Orange, I'm going to figure out just how down in the odds Cal can be before they get back into this thing.

First and foremost, I'm basing my win probability model of KenPom's old win probability model that hasn't really changed all that much. It was the most thorough thing I could find online and I found it easiest to account for favorites. (Back to that in a sec.) If you don't want to click the link, it divides a basketball game into five minute blocks with different leads as the Y-axis with percentages giving win probability to the team in the lead.

But Andy, very rarely are basketball games featuring two teams of an equal talent level! Very true! This means the handy dandy chart needs to be adjusted. The KenPom article talks about using a log-5 function to help compensate for favorites, but I found that unnecessarily tedious. A team comes in as a specific point favorite set by the spread and that's how the commonly accepted favorite/underdog narrative is set. Thus, instead of not treating time as linear (as Pomeroy does) I simply did as the spread suggests: add points to the favorite's line and thus allow the win probability to work it's magic from there.

Finally, I decided that "getting back into this" is an incredibly subjective term that I need to define at some point. Here's my definition: at any point in the second half of a game where Cal is within one possession (4 points) after their win probability dips below 40%.

Sound good? Kay because here comes the fun part: Here are all of the times Cal was "out of this thing" this season, arranged by their win probability:

11/27/14 vs Dayton: 19:19, down 10, 22%
1/11/14 vs Oregon State: 19:32 down 10, 22%
1/29/14 vs AZ State: 18:15 down 16, 20%
1/26/14 vs UCLA: 15:25 down 19, 2%

Ok, so in actuality, there were not as many games when Cal's win probability dropped below 40% and then found a way to get back into the game. In fact, here's exactly when in each of these games Cal did get back into it with the end result of the game.

vs Dayton: 16:16, cut deficit to 4, lost 82-64
vs Oregon State: 16:28, cut deficit to 3, won 88-83
vs AZ State: 4:09 left, cut deficit to 3, lost in overtime 89-78
vs UCLA: 9:17 left, cut deficit to 3, lost 76-64

So here's the gist of it: in a huge rivalry game, Cal got back into it after having an only 2% chance of winning. There's their magic for the year. The only game Cal won after getting back into this was against Oregon State in which the Golden Bears had just a 22% chance to win at their largest deficit.

You have your objective announcers: once Cal's win probability drops below 22%, please PLEASE know Cal is not getting back into this.