Probability & Statistics with a focus on Monte Carlo simulations & epistemology: https://xitter.cf/t.co/fkfSEBF0f9, podcast https://xitter.cf/t.co/fDorXuJxXq. #Stats #RStats #Statistics #Probability
Toronto
Joined April 2009
- Tweets2.2K
- Following1K
- Followers7.8K
- Likes144
Ockham’s razor is aspirational metaphysics, as in, “Wouldn’t it be nice if all of reality could be modeled with tidy, tractable equations?”
Roundup of links to my recent writings and podcast episodes related to probability, stats, and epistemology:
statisticsblog.com/2022/12/2…
A podcast episode about Pascal's Wager and it's implications for today:
mattasher.com/2021/10/04/ep-…
This is looking sadly more and more accurate as time goes by. Here's everyone's model island for excellent response 15 months on, in data and liberty:
xitter.cf/DrEliDavid/status/1434…
theatlantic.com/ideas/archiv…
To be as clear as possible: The fight against #Corona is like a 15-round prize fight. Lockdowns are socially extreme, costly, and hard to enforce. It's like recommending your boxer throw 50 punches in round 1. They may be up at the end of that round, but what happens after that?
Statistics Blog retweeted
The average person's mechanism for determining a truth claim is simply whether it gives them a positive emotional response.
A gentle reminder that all scientific progress rests on recognizing that things that work... work.
This may seem obvious or tautological, but I keep seeing people talk about supposedly useful interventions in the absence of real-world correlations. No correlation → no causation
For the past year, Science has meant bad predictions by zealots and partisans, used to justify extreme measures which needlessly destroyed millions of lives, while benefiting the already rich and powerful.
Also, you are not allowed to question The Science.
#truth #science
What are the most important statistical ideas of the past 50 years? @StatModeling
fermatslibrary.com/s/what-ar…
Statistics Blog retweeted
Everybody should know about Berkson’s paradox.
👉 When two positive traits are (spuriously) negatively correlated, in a population *selected on these traits*.
Examples: 1/
“Handsome men are jerks.”
via @anecdatally
Note: Something can’t be a factor in a decision making process unless it is sometimes The Decisive Factor. This is a deductively true as a matter of logic, not opinion.
Kea parrots perform domain-general statistical inference
nature.com/articles/s41467-0…
Three-dimensional model of electricity consumption in Manchester
collection.sciencemuseumgrou…
#dataviz
A reminder that while correlation doesn't prove causation, you can't prove causation *without* correlation.
Or, even worse, with inverse correlation.
Past thefilter.org guest Andrew Gelman and @avehtari on the most important statistical ideas of the past year:
stat.columbia.edu/~gelman/re…