dbmain.bsky.social profile picture

DB Main @dbmain.bsky.social

There are two basic problems with these statistical analyses:

  1. Statistically, they are treating each state/county that shifted as "independent" events. They're not. There are correlations between the events. Swing voters in NV and OH are both motivated by the same issues.

1
dbmain.bsky.social profile picture

DB Main @dbmain.bsky.social

When correlation is factored in, the "impossible" becomes "likely". One guy, when confronted with this said, "Yeah, but the math is easier to understand if you treat them independently".

Replies

  1. dbmain.bsky.social profile picture

    DB Main @dbmain.bsky.social

    1. They don't understand a basic statistical law. That election results were noisy at first and then just converged at certain levels isn't evidence of an algorithm: It's evidence of "Law of Large Numbers".
      This guy did a beautiful job of dissecting it.
      bsky.app/profile/trev...

    1