Boom !
-
@mastodonmigration @KamalaHarrisWin
I don't because it's useful to highlight when he's wrong.
Other people can do whatever they want, I think blinders are less useful.
-
@mastodonmigration @KamalaHarrisWin
I think people who tell you to ignore different stats methods when things don't go their way aren't very good interpreting stats and they only use stats as propaganda material.
I use stats to try to understand what's actually going on.
Other people can do whatever they want, they can use blinders and lie to themselves all they want.
-
It's all good. Just letting you know Rosenberg's take on Silver.
-
@mastodonmigration @KamalaHarrisWin
By the way, whoever says "Don't trust him, trust me and my work instead" is very transparently incompetent in my view and automatically untrustworthy.
But that's just me and I don't tell people what to do.
-
Lots of questions about Silver and his relationship with Peter Thiel.
-
@mastodonmigration @KamalaHarrisWin
Silver's model is being used by his former mates of FiveThirtyEight under the umbrella of ABC News.
I think it's not very smart -I actually think it's extremely stupid- for a statistician to ignore polls because somebody said he may be pals with Thiel.
I am aware of the fact that polls are sometimes used to lie. It's a core feature of pollsters.
But knowing how much they're lying and their biases and interpreting that is extremely useful.
-
Not disagreeing with you, but here is some background on the Thiel/Silver connection.
https://www.alternet.org/nate-silver-stuart-stevens-polymarket/
-
Jim E-H πΊπ¦replied to Mastodon Migration last edited by
@mastodonmigration @KamalaHarrisWin We especially need to run up the score this time, because it makes it less likely their intended cheating will succeed.
-
@mastodonmigration @KamalaHarrisWin
The data is the data. Ignoring Silver is just as bad as ignoring YouGov, or Ipsos, or any other serious analyst. You learn a lot from the biases.
I even take a look at the Real Clear Politics average, even when they include Rasmussen and Fox data (really terrible polls, more wrong than right most of the time), because that helps you estimate how skewed are somebody's numbers, knowing that RCP always give 1 or 2 points more to the Republican candidate.
-
Agreed.