13 Comments

The link to the incorrect margin of error is dead, and I'm enough of a nerd that I wanted to read it!

Expand full comment
Oct 18Liked by Carl Allen

Me too!

Expand full comment

Me threeeeeeeeeee!!!!

Expand full comment
author

Fixed fixed fixed sorry about that and thanks for the heads up

Expand full comment
author

Thoughts?

Expand full comment

If you haven't already read it, I think you might dig the book Fluke by Brian Klaas. In it, he discuses a little of the history of probability science.

Expand full comment

Yeah, but who’s gonna win?

Expand full comment

Believe Carls model has Kamala substantially higher than most pollsters right now. Around 60% I think

Expand full comment

You've attached only screenshots of Silver's statements for 2012 and 2023. Can you provide links instead? I'm asking because while it's obvious that he's talking about predictions at the *state* level in the 2012 screenshot, it's not obvious if the 2023 one refers to state or national level. If it refers to the *national* level, then it makes perfect sense, as the results of the US presidential elections are not that closely correlated with national polls due to the Electoral College.

Expand full comment
author

Thoughts?

Expand full comment
author

"If it refers to the *national* level, then it makes perfect sense"

And what if they refer to poll averages at the state level?

Expand full comment
author

It's state level

Expand full comment
author
Oct 18·edited Oct 18Author

Updated article with links, also given below

2012: https://fivethirtyeight.com/features/oct-26-state-poll-averages-usually-call-election-right/

2023 chart I referenced: "presidential, Senate, House and gubernatorial general elections since 1998, by how close the poll showed the race"

https://fivethirtyeight.com/features/2022-election-polling-accuracy/

Expand full comment