My non analytical brain is tickling and said it wasn't an awful decision to ride the hot hand with Lamar Jackson. But going for it again was kind of wild when it didn't work the first time.
However, I also don't think we know whether or not Harbaugh was going based off of analytics or based off of instincts. He is a head coach with decades of experience and has had a SB ring under his realm, and maybe based off of all the film-study and patterns he saw during practice, there was a reason behind why he chose to go for it twice.
Reading the piece, I saw what you were doing on the "statistically independent" bit, since this is analogous to the much debated "hot hand". Presenting a subtly fallacious argument then refuting it was fair play.
On the other hand, giving a wrong number on the odds, then correcting it, is cheating the reader, who can't be assumed to have the data. Trivially, the observed frequency of a binomial outcome is the maximum likelihood estimate, so you should have started by saying "some people think it's a coin flip, but the observed frequency is less than 0.5".
Well, it's unfortunately more complicated than that. There are years (e.g., 2023) where the observed frequency is more than 0.5, there are years (2024) where it's less. The success rate in 2024 was significantly lower than it had been over the last half-decade.
I can send you a link to more statistics if you'd like to review them.
I have to say, I'm not sure I understand your animosity towards statistical analysis here, or in general. Sometimes a given real world event can be usefully modeled as a given statistical game, and sometimes it can't, but the only way to know is to try. And statistics can help you figure out whether your statistical analysis is working!
And sometimes an "appropriate" statistical game for a given situation just tells you that there's not actually any way to "win" the game. These kinds of analyses don't get a ton of hype, but they are perfectly good statistical analyses!
oh so we're not talking about midwest emo? K I'm out
My non analytical brain is tickling and said it wasn't an awful decision to ride the hot hand with Lamar Jackson. But going for it again was kind of wild when it didn't work the first time.
However, I also don't think we know whether or not Harbaugh was going based off of analytics or based off of instincts. He is a head coach with decades of experience and has had a SB ring under his realm, and maybe based off of all the film-study and patterns he saw during practice, there was a reason behind why he chose to go for it twice.
That's also fair. Who knows what he was thinking! But he's the king of blowing 2nd half leads...
Well the special teams meta is shifting too, as the Eagles demonstrated with those two blocked fgs….go birds
Three blocked field goals in ten minutes across the league. Statistical aberration or sign of the apocalypse?
Reading the piece, I saw what you were doing on the "statistically independent" bit, since this is analogous to the much debated "hot hand". Presenting a subtly fallacious argument then refuting it was fair play.
On the other hand, giving a wrong number on the odds, then correcting it, is cheating the reader, who can't be assumed to have the data. Trivially, the observed frequency of a binomial outcome is the maximum likelihood estimate, so you should have started by saying "some people think it's a coin flip, but the observed frequency is less than 0.5".
Well, it's unfortunately more complicated than that. There are years (e.g., 2023) where the observed frequency is more than 0.5, there are years (2024) where it's less. The success rate in 2024 was significantly lower than it had been over the last half-decade.
I can send you a link to more statistics if you'd like to review them.
🛌
I have to say, I'm not sure I understand your animosity towards statistical analysis here, or in general. Sometimes a given real world event can be usefully modeled as a given statistical game, and sometimes it can't, but the only way to know is to try. And statistics can help you figure out whether your statistical analysis is working!
And sometimes an "appropriate" statistical game for a given situation just tells you that there's not actually any way to "win" the game. These kinds of analyses don't get a ton of hype, but they are perfectly good statistical analyses!