Will Lowe | Lightnews

Light up
your news

About Privacy Terms Help

Will Lowe

Will Lowe

@conjugateprior.org

3.4K followers 740 following 10K posts

Señor Research Scientist, NPC at the Hertie School in Berlin 🇩🇪 via Princeton, Mannheim, Edinburgh and a bunch of other ivory towers that will probably be billiard balls and decorative boxes by the end of the decade.

Rome Statute appreciator. .. more

Computer science 32%

Political science 28%

more

Posts Replies Media Videos

Pinned

Will Lowe @conjugateprior.org · Apr 15

For Monty Hall problem aficionados: a #causal DAG, with explanation in the alt text.

Who knew that the M in M-bias stood for Monty?

A DAG for the Monty Hall problem.

G, the door Monty opens, is a function of your first door choice D_F and of where the car is C.

Conditioning on G, which happens because it determines your door choices when you have the opportunity to switch, induces collider bias between C and D_F, such that they are now less likely to be equal. (Why? because G cannot be C and G cannot be D_F. These are in the same 'direction' so, conditional on G, D_F is less likely to be C.)

You decide to switch conditional on G, which you just observed, and D_F, because that defines what you can switch to.

Conditional on G you now know D_F is less likely to be equal to C. So you switch.

Note: In the original problem, W is a deterministic function of C and D_S so it's slightly cheeky to put it in a DAG, but I'm going to do so anyway. The argument doesn't require it, but if you want to think informally about opening paths to W from D_F via C, it's helpful.

That makes this problem a variation on M-bias (or a clipped wing butterfly bias, if you prefer insects to goats).

Will Lowe @conjugateprior.org · 4h

Eh, we should be happy they don't vary by day of the week. And don't ask if the thing you're capitalizing contains parts that are not English because then language simulation glitches, the jinns enter, and there's non-stop screaming until someone reboots it from Latin.

Will Lowe @conjugateprior.org · 4h

English is feral (Part 128474343)

Will Lowe @conjugateprior.org · 4h

There's also the 'Dark Dennett' view: folk psych works great (because we socialize each other to ensure it does) but the microfoundations are arbitrary or there aren't distinct sets of psych/neurological mechanisms worth making fields out of. Most social sciences survive that view fine but not psych

Reposted by Will Lowe

econmaett @econmaett.github.io · 21h

Enough with astrology. Tell me which Kalman filter you are! 🚀💥
#stats

www.buzzfeed.com/danielmlacha...

Which Kalman Filter Are You?

Which Kalman Filter best represents you? Take the quiz to find out.

www.buzzfeed.com

Will Lowe @conjugateprior.org · 1d

Extremely niche typesetting observation: typst is mostly just context for people who like rust more than lua.

Will Lowe @conjugateprior.org · 1d

The curriculum office once tried to list my usual Fall course as "casual interference". That's now just what it's called in my brain.

Will Lowe @conjugateprior.org · 1d

Diese "Saw"-Filme haben wirklich nachgelassen.

Will Lowe @conjugateprior.org · 1d

The complete Ladybird 'What to look for' collection.

Helen Day @lbflyawayhome.bsky.social · 1d

Two lesser known books today, both
illustrated by Ronald Lampitt (1972)

The front covers of the two books which are guides to better understanding the design and features of British churches

Reposted by Will Lowe

Helen Day @lbflyawayhome.bsky.social · 1d

Two lesser known books today, both
illustrated by Ronald Lampitt (1972)

The front covers of the two books which are guides to better understanding the design and features of British churches

Will Lowe @conjugateprior.org · 2d

No, that seems brutally correct

Will Lowe @conjugateprior.org · 2d

"ERSTAUNLICH"

Will Lowe @conjugateprior.org · 2d

Wow

Will Lowe @conjugateprior.org · 2d

Kartoffelcore ist im Trend. I don't make the rules.

Will Lowe @conjugateprior.org · 2d

🥔🥔🍟🥔🥔🍟🍠🍟😈

Will Lowe @conjugateprior.org · 2d

Tomaten sind so was von letzer Woche. Die Zutaten der Zukunft findet man nur wenige Kilometer südlich von Leipzig

www.morgenpost.de/berlin/artic...

Vier Millionen Kilo Kartoffeln suchen ein Zuhause

Zum 100. Mal jährt sich die Internationale Grüne Woche. Zu diesem Anlass verschenken Berliner Morgenpost und Ecosia tonnenweise Kartoffeln. So können Sie mitmachen.

www.morgenpost.de

Will Lowe @conjugateprior.org · 2d

Eh, I've been rumbled. Can I borrow him as cover?

Will Lowe @conjugateprior.org · 2d

That's a excellent Saturday's worth of stopwatch and notebook escalator tourism right there.

Will Lowe @conjugateprior.org · 2d

I think we have to assume that if a url is in the suggested citation then they're planning on keeping it. (my optimism knows no bounds)

Will Lowe @conjugateprior.org · 2d

I haven't seen a clinical practice guideline without either an ISBN or a URL you're supposed to use in the citation (though mine is a rather small sample). Government documents seem more heterogenous.

Will Lowe @conjugateprior.org · 2d

and here's something a little tighter*

*the narrowest street in Shrewsbury, or so I'm told by people who you can't get _not_ to tell you things like this in rural pubs.

Grope Lane, Shrewsbury. Lots of wood in these buildings, obvs.

Will Lowe @conjugateprior.org · 2d

Perfect. If they name another county, I'll be disappointed if they don't choose that.

Will Lowe @conjugateprior.org · 2d

That's fair. Wessex has been hard to find since about 900 😉

Will Lowe @conjugateprior.org · 3d

Ha! The perfect antidote to all the "omg these big assumptions are so fierce and strict and I'll never know if they hold for little old me and my smol data analysis"

Will Lowe @conjugateprior.org · 3d

If some models are useful, just how wrong can they be?

#BoxCutters

Will Lowe @conjugateprior.org · 3d

For models, wrongness comes in degrees - like usefulness, really.

#BoxCutters

Will Lowe @conjugateprior.org · 3d

Some models are right - but probably not yours.

#BoxCutters

Will Lowe @conjugateprior.org · 3d

All models are wrong, if by wrong you mean 'not identical to the thing you're modeling', which you apparently do (and also don't.)

#BoxCutters

Reposted by Will Lowe, Daniel G. Saunders

Thomas Lumley @tslumley.bsky.social · 3d

Which is why assumptions you can check aren't very helpful: if an assumption doesn't cost anything it won't buy much

Reposted by Will Lowe

Will Lowe @conjugateprior.org · 3d

😵‍💫