David Salazar Profile picture
Dilettant'n Data Science. I enjoy multiplying random numbers over and over again. #rstats and Python.
3 subscribers
Sep 7, 2020 6 tweets 3 min read
Smell BS in election models?

Well, most of them are way too confident and too volatile.

@nntaleb shows that one cannot invoke probabilistic thinking in one's forecasts of a highly uncertain election unless the forecasts are stable and around 50/50. Why?

Explanation⬇️ Image With a single event, frequency calibration is meaningless. Yet there are other ways to ensure coherence in one's belief.

De Finetti showed that coherent subjective probabilistic statements must imply a bet that precludes the possibility of arbitrage. Excerpt from De Finetti⬇️ Image
Jul 20, 2020 33 tweets 20 min read
I find stats talk around causality wanting: most of the time, the concept is simply side-stepped. To remedy this, I'll read through @yudapearl's seminal work 'Causality' using his 'Book of Why' as an interpretative key. In this thread, I'll share what I learn. Bayesian Networks and Probability Distributions. How can we model a joint probability distribution in a computer? What is the probabilistic connection? d-separation and testable implications from graphical methods, Ladder of Causation..

Blogpost:
david-salazar.github.io/2020/07/18/cau…
#rstats
Apr 23, 2020 6 tweets 7 min read
Going over @rlmcelreath's latest book in #rstats. However, I am using #tidyverse tools. So far, I have watched the lectures, completed homework for week 1 and the exercises in chapter 2 of the book. Hope to make it till the end!
Btw, the book is amazing.
david-salazar.github.io/2020/04/19/sta… ImageImage Week 2 of Statistical Rethinking! Lectures and homework completed. @rlmcelreath is equal parts entertaining and knowledgeable. Homework made crystal clear how difficult it is to set priors for polynomial models #tidyverse #rstats

david-salazar.github.io/2020/04/28/sta… ImageImage
Apr 23, 2020 55 tweets 33 min read
Trying to get my head around fat-tails by studying @nntaleb's latest technical book. Replicated some plots in #rstats #tidyverse
david-salazar.github.io/2020/04/17/fat… #rstats and @nntaleb's work. By fattening the tails, one learns that the tail events are convex to the scale of the distribution. Thus, the problem compunds: tail events have an increasingly large role, but we cannot estimate their probabilities reliably

david-salazar.github.io/2020/05/09/wha…