Published on [Permalink]
Reading time: 2 minutes
Posted in:

A few mildly related pre-election observations

  1. It won’t be close. Most pollsters are hacks who commit even greater statistics crimes than physicians so their 50/50 is most likely to mean a landslide either way.
  2. That link above is to Nate Silver’s Substack post, but please remember that he is also a hack who builds prediction models from the polling garbage he describes above while knowing it is garbage. That is even worse than what the pollsters are doing because shouldn’t he know better?
  3. Worse yet are economists who excuse the pollster behavior: they see crimes being committed and think yep, that’s how it should be. This is a University of Michigan professor of economics and a senior fellow of some pretty serious Think Tanks who doesn’t realize that fiddling with your results after you’ve collected them in order to better align with the aggregate of other people’s results is scientifically unsound. I’d send all of his papers to Retraction Watch for a close inspection.
  4. From 538 to the NYT to Nate, every poll aggregator has for months been fed back its own bullshit. Little wonder then that they all converged to a 50/50: complete ignorance.
  5. Prediction “markets” are no better than equity markets in reflecting reality. Which is to say, they reflect the reality of vibes and wishful thinking, not the ground truth. They are best ignored.
  6. Sátántangó (2019) by the Hungarian director Béla Tarr is a 7-hour masterpiece shot in black and white; perfect for watching on a crisp autumn evening like tonight’s, no other screens allowed.

✍️ Reply by email

✴️ Also on Micro.blog