I went back and looked at the different win probabilities ESPN has given us this season during games.

Last week, ESPN gave Utah a 99.7% chance of winning on the last 4th and 10 with 1:35 left in the fourth quarter, leaving BYU with only a 0.3% chance of pulling off the comeback.

Against Oklahoma State, ESPN had Ok State at a 99.2% chance of victory with 52 seconds remaining when we had to convert 4th and 7, giving BYU a mere 0.8% chance.

When you combine the probabilities of these two miraculous wins happening in the same season, you get:
0.003 × 0.008 = 0.000024


That’s a joint probability of 0.0024%, or about 1 in 41,667 for just those two games. 

The other two times this season where I really felt like we were on the ropes was against SMU and KState.

After a really bad interception from Jake deep in their own territory, BYU was down 9-7. SMU had a 2nd and 5 on the BYU's 18-yard line. ESPN gave SMU an 81.1% chance of victory with 11 minutes left in the third quarter. Defense held to a field goal.

Against Kansas State, they were up 3-0 and inside our rezone, also with a 2nd and 5 with 11:50 left in the second quarter, ESPN gave them an 81.4% chance of victory. Again, the defense held to a field goal.

Assuming these events are independent, the combined probability of BYU winning all four games is:
0.003 × 0.008 × 0.186 × 0.189 = 0.0000002118

That translates to a probability of 0.00002118%, or roughly 1 in 4.723 million. Let that sink in for a second.  

Joel Klatt and other national commentators have said BYU reminds them of TCU 2022. I went to to see if in their entire season they had comparable probabilities:
- Oct 15th TCU v OK State, with 13:36 left in the fourth quarter, OK state had a 96.1% chance to win. - This game went into double overtime, TCU won.
- Nov 19th: TCU v Baylor, Baylor had a 92% chance to win with 8:52 left in the fourth. TCU won on a last second field goal.
- Calculating this joint we get: 0.039 × 0.08 = 0.00312
- So, the joint probability of TCU winning both games is 0.312%, or about 1 in 320.

If we just look at the first two games for BYU and compare them to these two games for 2022 TCU, we get
0.00312 /0.000024 = 130

Meaning that in just those two games, BYU’s results are 130 times more unlikely..

Whatever happens for the rest of the season, just enjoy it - 1 in 4.7 million odds don’t come around often!

PS: I got laid off from PwC last month in a structural pivot from the firm. If you know of anyone or are hiring in the data science space, would appreciate any warm leads. 



Cheers,