Stephen Wild
stephenjwild.bsky.social
Stephen Wild
@stephenjwild.bsky.social
I try to put straight lines through things but usually fail. Try to be Bayesian when I can. Views my own. RT/like != endorsement.
If I were an ambitious Scottish general encouraged to violence by my wife, I would simply distinguish myself by killing the king and ascending to the throne.
November 11, 2025 at 8:53 AM
{ggdag} is also a nice package and has some nice vignettes.

r-causal.github.io/ggdag/articl...
An Introduction to Directed Acyclic Graphs
r-causal.github.io
November 11, 2025 at 1:53 AM
BUY IT, JOEL
November 11, 2025 at 12:09 AM
*thought about it a bit
November 10, 2025 at 5:53 PM
Having thought about it about it able bit, I think what bothers we is how early the curves separate. This to me suggests that it's probably caffeine withdrawal triggering it.

Lesson: Don't stop drinking coffee
November 10, 2025 at 5:53 PM
Whoa, whoa, whoa. I am clearly implying the effect should be larger.
November 10, 2025 at 4:40 PM
I dunno. Something about it makes me question it. And I say this as a dude who likes his coffee!
November 10, 2025 at 4:29 PM
That KM curve seems off. Something about it is setting off an alarm bell in my head.
November 10, 2025 at 3:16 PM
Exactly.
November 10, 2025 at 2:54 PM
Feels like most days, tbh
November 10, 2025 at 1:09 PM
Depends on your ideological view 🤪

Most of the research in this area is terrible. But also, no one agrees on what "masks work" means.
November 10, 2025 at 1:08 PM
But yeah, I agree. It usually stands out
November 10, 2025 at 2:10 AM
A bright flashing light you say?

In the paper they use they asymmetric Laplace:

bsky.app/profile/step...
That's why you need things lkle this (this comes from a paper that reported positive results. Some of them are clearly driven by the prior)
November 10, 2025 at 2:06 AM
This seems like something I would have learned from you, @guidobiele.bsky.social. Any ideas?
November 10, 2025 at 12:18 AM
I don't think so. I think your prior would need to be symmetrical. But I need to look into it more
November 10, 2025 at 12:17 AM
Depends on the prior. If you set a gamma prior on something that should/can be negative, for example, you may have a problem.
November 8, 2025 at 11:10 PM
It's always an interesting assumption that you're going to buy more than one home
November 8, 2025 at 10:34 PM
I think it was @rmkubinec.bsky.social who directed me to this
November 8, 2025 at 5:51 PM
Tangential to your point, but nonetheless
November 8, 2025 at 5:49 PM
As a former retail employee, the "smile/greet them if they're within ten feet" rule is annoying. As a customer, I find it offputting and creepy because it comes across as fake.
November 8, 2025 at 5:47 PM