Like other posters, I felt a lot of frustration and disappointment around yesterday's collapse. I will confess that as the lead ebbed away, I turned off the television and went for a walk to avoid kicking in the set as the Cats let go of a 27 point lead.
Today, after reading the NY Times and SF Chronicle's stories about the comeback, I realized that my emotional swing was both sparked and tempered by (in truly nerdy NU fashion) a belief in the world governed by Beyes Rule/Theorem.
Before the game, with Mac and Ash both out, I had extremely low expectations. I thought the chances of the Cats winning were at best maybe .05. If you had asked me if I thought MSU would have a 15 minutes stretch where they outscored the Cats by 30, I would have guessed that the probability of such an event was about one in ten.
When the Cats played so well in the first 15 minutes, it clearly played on my emotions as I recalculated the probability of a win. ESPN moved that to 97% at one point, and though I was fearful of a collapse, I did not really think it would happen. I fell victim of a fallacy: I assumed that the growth of the lead indicated a fundamental change in my perception of the disparity of skill between the two teams. Clearly, the Cats played well the first 15 minutes, but what I missed was how badly MSU played, especially at the defensive end.
As a Beyesian, I should have focused on the original estimate that an MSU +30 run originally had about a 10% probability. Once the Cats opened the lead, that 10% probability was influenced by two key factors: the first was that Izzo would get MSU really fired up and focused due to the deficit, and secondly, the Cats had expended a huge amount of energy in extending the lead, and suffered from a short bench, which would lead to rubber legs in the second half. The probability of a big MSU run probably increased significantly as a result of these two factors, and while the Cats still had a very good chance to win, the probability of an MSU run that could level the score probably increased to about 1 in 4., which was disappointing but not apocalyptic. MSU's run could have been predicted by the NU run and was not contradictory to it.
The whole set of emotions, that is, going from feeling there was no chance to almost a sure thing to disappointment should have been tempered by an understanding of how building the lead created the condition for it dissipating. I don't teach undergrads statistics anymore, but this game would be a great model to help at least the sports fans in an intro class understand what Beyes Rule means.
Today, after reading the NY Times and SF Chronicle's stories about the comeback, I realized that my emotional swing was both sparked and tempered by (in truly nerdy NU fashion) a belief in the world governed by Beyes Rule/Theorem.
Before the game, with Mac and Ash both out, I had extremely low expectations. I thought the chances of the Cats winning were at best maybe .05. If you had asked me if I thought MSU would have a 15 minutes stretch where they outscored the Cats by 30, I would have guessed that the probability of such an event was about one in ten.
When the Cats played so well in the first 15 minutes, it clearly played on my emotions as I recalculated the probability of a win. ESPN moved that to 97% at one point, and though I was fearful of a collapse, I did not really think it would happen. I fell victim of a fallacy: I assumed that the growth of the lead indicated a fundamental change in my perception of the disparity of skill between the two teams. Clearly, the Cats played well the first 15 minutes, but what I missed was how badly MSU played, especially at the defensive end.
As a Beyesian, I should have focused on the original estimate that an MSU +30 run originally had about a 10% probability. Once the Cats opened the lead, that 10% probability was influenced by two key factors: the first was that Izzo would get MSU really fired up and focused due to the deficit, and secondly, the Cats had expended a huge amount of energy in extending the lead, and suffered from a short bench, which would lead to rubber legs in the second half. The probability of a big MSU run probably increased significantly as a result of these two factors, and while the Cats still had a very good chance to win, the probability of an MSU run that could level the score probably increased to about 1 in 4., which was disappointing but not apocalyptic. MSU's run could have been predicted by the NU run and was not contradictory to it.
The whole set of emotions, that is, going from feeling there was no chance to almost a sure thing to disappointment should have been tempered by an understanding of how building the lead created the condition for it dissipating. I don't teach undergrads statistics anymore, but this game would be a great model to help at least the sports fans in an intro class understand what Beyes Rule means.
Last edited:
