Joe Blitzstein
stat110.bsky.social
Joe Blitzstein
@stat110.bsky.social
Statistician at Harvard. Probability and paradoxes. Statistics and data science. Books and chess.
I'm glad it was an impactful experience. I remember having you in the class and hearing about your music research!
November 15, 2025 at 5:05 PM
Awwww thank you, I am very proud to have had you and @mayasen.bsky.social as students too! As for this quote, I just said what should be obvious to anyone with a decent sense of morality who read the emails, but sometimes it's important to say obvious things on the record.
November 15, 2025 at 5:02 PM
Yep, only possible for fall semester courses
December 13, 2024 at 3:32 PM
@hadleywickham.bsky.social says at adv-r.hadley.nz/oo.html "Generally in R, functional programming is much more important than object-oriented programming... Nevertheless, there are important reasons to learn each of the three systems". Definitely both are valuable.
Introduction | Advanced R
adv-r.hadley.nz
December 10, 2024 at 10:40 PM
Thanks! :) Yes I did coin that (as far as I know), by analogy with object-oriented programming. OOP seems to have gone out of fashion but certainly still has some useful concepts such as encapsulation and inheritance, which are useful both for coding and for probability and statistics.
December 10, 2024 at 10:07 PM
It's amazing that a free chess app running on an iPhone can play at a 3600 ELO rating level. No human has ever gotten a rating of 2900+. The app is called Smallfish, which runs stockfishchess.org
Stockfish
Strong open-source chess engine
stockfishchess.org
November 19, 2024 at 1:18 AM
The leading chess engines are now so much stronger than any human that this isn't really the case anymore. For a while there were some interesting "advanced chess" or "centaur chess" tournaments where a human worked with a computer, but I haven't seen that recently: en.wikipedia.org/wiki/Advance...
Advanced chess - Wikipedia
en.wikipedia.org
November 19, 2024 at 12:51 AM
It would also be helpful for probability and statistics if the Riemann-Stieltjes integral were taught more often. For example, then we can write E(X) as the integral of x dF(x), where F is the CDF of X, rather than having to write separate formulas for discrete X and continuous X.
November 18, 2024 at 10:26 PM
so do I, for over 15 years!
November 18, 2024 at 10:22 PM