Yeah, that was a particularly good analysis, and confirmed what I've felt all along. I particularly like the very precise formulations he suggests when making predictions based on poll movements.
However, I would not get too despondent over Nate's lack of enthusiasm. Nate made his name by being accurate and cautious and never getting ahead of the facts. He takes quite a while before upgrading or downgrading a race.
Consider PA-Sen. He still gives Toomey an 83% change of winning. I forget the precise details, but something like 3 recent polls have put Sestak ahead and 2 or 3 others have shown a dead heat. I think Toomey still has to be considered the favorite, but since Nate relies so much on ALL of the polls, including less recent polls, their weight in his model causes it to lag.
All in all, Nate will have the outcomes correct at the end, to the decimal pont, I suspect. But for now his models poorly incorporate the recent positive gains the the dems and paint a bleaker picture than may be the case for the election.
My personal forecast has been upgraded from "apocalyptic" to "catastrophic with a 20 % chance of shitty."
Nate Silver's stuff is so data filled. He knows his audience as well. His article today about "comebacks," which is not unlike the article I liked here, started off with the Boxer race, pointing out that Fiorni (or whatever) didn't really have much of a chance, despite a little spike in polls here and there.
He is getting a little nervous about counting Sestak out.
3 comments:
Wow, great article.
Yeah, that was a particularly good analysis, and confirmed what I've felt all along. I particularly like the very precise formulations he suggests when making predictions based on poll movements.
However, I would not get too despondent over Nate's lack of enthusiasm. Nate made his name by being accurate and cautious and never getting ahead of the facts. He takes quite a while before upgrading or downgrading a race.
Consider PA-Sen. He still gives Toomey an 83% change of winning. I forget the precise details, but something like 3 recent polls have put Sestak ahead and 2 or 3 others have shown a dead heat. I think Toomey still has to be considered the favorite, but since Nate relies so much on ALL of the polls, including less recent polls, their weight in his model causes it to lag.
All in all, Nate will have the outcomes correct at the end, to the decimal pont, I suspect. But for now his models poorly incorporate the recent positive gains the the dems and paint a bleaker picture than may be the case for the election.
My personal forecast has been upgraded from "apocalyptic" to "catastrophic with a 20 % chance of shitty."
Nate Silver's stuff is so data filled. He knows his audience as well. His article today about "comebacks," which is not unlike the article I liked here, started off with the Boxer race, pointing out that Fiorni (or whatever) didn't really have much of a chance, despite a little spike in polls here and there.
He is getting a little nervous about counting Sestak out.
Post a Comment