forums

MEGA Week 6 Games Thread

MiamisMoperiestMopes said:
View attachment 375526


Cracking Up Lol GIF by MOODMAN
Click to expand...
ChatGPT’s analysis

That is a remarkable collapse, particularly because of how sustained the downward slope is. What stands out isn't simply that Florida State lost its advantage. It's that once the probability started falling, it almost never recovered.

Three things I find particularly interesting


1. The asymmetry is striking.


FSU's probability rises from roughly 50% to 90% through a series of relatively gradual increases. But the descent isn't a mirror image.


The first major downward adjustment is almost vertical, from approximately 85–90% to 65%. After that, the decline continues with remarkably little sustained upward movement.


2. There is almost no resistance on the way down.


Look at the second half of the graph. You see brief upward movements, but almost every one is followed by another decline.

In financial-chart terminology, this resembles a sequence of lower highs and lower lows.


That suggests Louisville wasn't simply benefiting from one enormous play. The win-probability model was repeatedly receiving new information that made an FSU victory less likely.


It looks like an extended loss of control rather than a single turning point.


3. The most damaging stretch is not necessarily the initial plunge.


FSU goes from approximately 90% to 50% during the second quarter. That's dramatic.


But going from 50% to 10% through the second and third quarters is arguably more revealing.


At 50%, FSU still has a competitive game. Yet the graph shows virtually no meaningful recovery after that point.


That's what makes the shape unusual: the team doesn't just surrender its advantage; it never establishes another sustained period of positive momentum.

The mathematical perspective


One interesting distinction: a drop from 90% to 50% isn't mathematically equivalent to a drop from 50% to 10%.


Expressed as odds:


90% probability = 9-to-1 odds of winning.


50% probability = 1-to-1.


10% probability = 1-to-9.


Each transition represents a ninefold deterioration in FSU's odds of winning.


So the two sections are equally substantial on a log-odds scale, despite looking different on the probability chart.


My biggest takeaway


What makes this graph fascinating is the persistence of the negative trajectory.


A normal back-and-forth football game produces something resembling a jagged sawtooth, even when one team eventually pulls away.


Here, after FSU peaks, the probability curve resembles a descending staircase with very few upward steps.


One qualification: win probability is influenced by score, possession, field position, and remaining time. A trailing team's probability can steadily decline simply because the clock is running out. So we cannot attribute every downward movement to poor execution without examining the play-by-play.


But the sustained descent beginning while FSU still had a substantial advantage is particularly notable.


I'd be interested in measuring the longest sustained decline in win probability without a meaningful recovery. Not just total probability lost, but how long a team goes without recovering, say, 10 percentage points from a local low.


That would distinguish a dramatic comeback from a genuine, prolonged collapse.
 
MiamisMoperiestMopes said:
ChatGPT’s analysis

That is a remarkable collapse, particularly because of how sustained the downward slope is. What stands out isn't simply that Florida State lost its advantage. It's that once the probability started falling, it almost never recovered.

Three things I find particularly interesting


1. The asymmetry is striking.


FSU's probability rises from roughly 50% to 90% through a series of relatively gradual increases. But the descent isn't a mirror image.


The first major downward adjustment is almost vertical, from approximately 85–90% to 65%. After that, the decline continues with remarkably little sustained upward movement.


2. There is almost no resistance on the way down.


Look at the second half of the graph. You see brief upward movements, but almost every one is followed by another decline.

In financial-chart terminology, this resembles a sequence of lower highs and lower lows.


That suggests Louisville wasn't simply benefiting from one enormous play. The win-probability model was repeatedly receiving new information that made an FSU victory less likely.


It looks like an extended loss of control rather than a single turning point.


3. The most damaging stretch is not necessarily the initial plunge.


FSU goes from approximately 90% to 50% during the second quarter. That's dramatic.


But going from 50% to 10% through the second and third quarters is arguably more revealing.


At 50%, FSU still has a competitive game. Yet the graph shows virtually no meaningful recovery after that point.


That's what makes the shape unusual: the team doesn't just surrender its advantage; it never establishes another sustained period of positive momentum.

The mathematical perspective


One interesting distinction: a drop from 90% to 50% isn't mathematically equivalent to a drop from 50% to 10%.


Expressed as odds:


90% probability = 9-to-1 odds of winning.


50% probability = 1-to-1.


10% probability = 1-to-9.


Each transition represents a ninefold deterioration in FSU's odds of winning.


So the two sections are equally substantial on a log-odds scale, despite looking different on the probability chart.


My biggest takeaway


What makes this graph fascinating is the persistence of the negative trajectory.


A normal back-and-forth football game produces something resembling a jagged sawtooth, even when one team eventually pulls away.


Here, after FSU peaks, the probability curve resembles a descending staircase with very few upward steps.


One qualification: win probability is influenced by score, possession, field position, and remaining time. A trailing team's probability can steadily decline simply because the clock is running out. So we cannot attribute every downward movement to poor execution without examining the play-by-play.


But the sustained descent beginning while FSU still had a substantial advantage is particularly notable.


I'd be interested in measuring the longest sustained decline in win probability without a meaningful recovery. Not just total probability lost, but how long a team goes without recovering, say, 10 percentage points from a local low.


That would distinguish a dramatic comeback from a genuine, prolonged collapse.
Click to expand...
ChatGPT doing it's thing...wasting a lot of words just to say "FSU sucks"
 
cookies said:
I ask again—is Jedd fisch on the hot seat?
Click to expand...

Fisch did a pretty good job and recruited well the first two years, but this season has been a train wreck and a big step backwards.

They're down 20-0 in the second quarter against Iowa. At home. About to be 0-3 in the B1G after being a trendy Top 25 pick.

So yeah, somewhat on the hot seat, to answer your question.

But UW's administration is terrible, kind of one-foot-in, one-foot-out with football as far as giving the program the kind of resources you need to win at this level.

They got away with it in the Pac-12, but that ain't gonna happen in the B1G obviously.

Not sure they'll get it together anytime soon. Administration probably resents the football players making more money than the women's basketball and volleyball teams.
 
MiamisMoperiestMopes said:
MiamisMoperiestMopes said:
View attachment 375526


Cracking Up Lol GIF by MOODMAN
Click to expand...
ChatGPT’s analysis

That is a remarkable collapse, particularly because of how sustained the downward slope is. What stands out isn't simply that Florida State lost its advantage. It's that once the probability started falling, it almost never recovered.

Three things I find particularly interesting


1. The asymmetry is striking.


FSU's probability rises from roughly 50% to 90% through a series of relatively gradual increases. But the descent isn't a mirror image.


The first major downward adjustment is almost vertical, from approximately 85–90% to 65%. After that, the decline continues with remarkably little sustained upward movement.


2. There is almost no resistance on the way down.


Look at the second half of the graph. You see brief upward movements, but almost every one is followed by another decline.

In financial-chart terminology, this resembles a sequence of lower highs and lower lows.


That suggests Louisville wasn't simply benefiting from one enormous play. The win-probability model was repeatedly receiving new information that made an FSU victory less likely.


It looks like an extended loss of control rather than a single turning point.


3. The most damaging stretch is not necessarily the initial plunge.


FSU goes from approximately 90% to 50% during the second quarter. That's dramatic.


But going from 50% to 10% through the second and third quarters is arguably more revealing.


At 50%, FSU still has a competitive game. Yet the graph shows virtually no meaningful recovery after that point.


That's what makes the shape unusual: the team doesn't just surrender its advantage; it never establishes another sustained period of positive momentum.

The mathematical perspective


One interesting distinction: a drop from 90% to 50% isn't mathematically equivalent to a drop from 50% to 10%.


Expressed as odds:


90% probability = 9-to-1 odds of winning.


50% probability = 1-to-1.


10% probability = 1-to-9.


Each transition represents a ninefold deterioration in FSU's odds of winning.


So the two sections are equally substantial on a log-odds scale, despite looking different on the probability chart.


My biggest takeaway


What makes this graph fascinating is the persistence of the negative trajectory.


A normal back-and-forth football game produces something resembling a jagged sawtooth, even when one team eventually pulls away.


Here, after FSU peaks, the probability curve resembles a descending staircase with very few upward steps.


One qualification: win probability is influenced by score, possession, field position, and remaining time. A trailing team's probability can steadily decline simply because the clock is running out. So we cannot attribute every downward movement to poor execution without examining the play-by-play.


But the sustained descent beginning while FSU still had a substantial advantage is particularly notable.


I'd be interested in measuring the longest sustained decline in win probability without a meaningful recovery. Not just total probability lost, but how long a team goes without recovering, say, 10 percentage points from a local low.


That would distinguish a dramatic comeback from a genuine, prolonged collapse.
Click to expand...
20261009_220738_C5E0668B-1B49-4F0E-9A89-7DDA2BA6CCCA.png
 
Back
Top