Lightnews — Scholar-powered news

Will Lowe

@conjugateprior.org

3.3K followers 730 following 9.4K 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

Videos

Computer science 32%

Political science 28%

Posts Media Videos Starter Packs

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).

6 19 89

Will Lowe @conjugateprior.org · 3h

Introducing... Starter Patches

Will Lowe @conjugateprior.org · 3h

And you'll have a nice concrete example of the consequences of differential item functioning, too.

1 3

Will Lowe @conjugateprior.org · 4h

If you're unhappy with structured assumptions about what could have been observed but wasn't, then there's not much left of statistics except counting things. Any confidence interval is populated almost entirely by data that could have been observed but wasn't and is in it (or not) by assumption.

Will Lowe @conjugateprior.org · 15h

Finally, some descriptive representation.

1 3

Will Lowe @conjugateprior.org · 16h

if you gift it to him, perhaps

Will Lowe @conjugateprior.org · 17h

Turns out the 26.1 beta is quite a bit less silly-looking, should you wish to fix the most egregious fails.

Will Lowe @conjugateprior.org · 18h

That's certainly something the current Charlemagne knows how to write.

Will Lowe @conjugateprior.org · 22h

fwiw my beef with the "writing is thinking" crowd, banging on about LLMs, is not that they're wrong but that they don't take the idea seriously enough.

Will Lowe @conjugateprior.org · 22h

which is why this as an identity not a prerequisite relationship: it's the code that realizes your thoughts about a problem's structure and solution. Sometimes those are creative, novel, even inspiring thoughts (sometimes not), but when your brain turns to lunch, those thoughts are still there.

2 2 3

Will Lowe @conjugateprior.org · 22h

writing is thinking (code edition)

1 3 11

Will Lowe @conjugateprior.org · 1d

Wait, ICD-11 is also my mother in a log flume?

Will Lowe @conjugateprior.org · 1d

I would also echo the bandcamp suggestion in the other replies. I find great stuff there quite regularly.

And if you want to pay people more serious money for their music, buying stuff there is much more effective than selecting among low (and very marginally different) stream payment rates.

Will Lowe @conjugateprior.org · 1d

Almost every other service had better quality streams when I jumped from Spotify to Qobuz. It was nice and I would have stayed with it had they not incomprehensibly locked the language of sleeve notes, commentary & other text content to that of the billing address. I ended up on Apple.

1 1

Will Lowe @conjugateprior.org · 1d

I think this speaks to more to the mesmerising weirdness of Belgian junk food. Brussels in the 90s is my own mental exemplar of an uncanny valley food landscape. Strange, strange things in convenience stores...

Will Lowe @conjugateprior.org · 1d

In a similar vein, I tend to tell students that using Bayes Theorem doesn't make you a Bayesian. Dealing with uncertainty by marginalization makes you a Bayesian.

1 6

Will Lowe @conjugateprior.org · 1d

Just tried to make this come out as true and I think I pulled a muscle.

Will Lowe @conjugateprior.org · 1d

Happy to be having a little trouble remembering Nick Clegg.

Will Lowe @conjugateprior.org · 1d

omnomnom

bart simpson and lisa simpson are standing in front of a house made of gingerbread

Alt: The bars are to stop them getting out, obvs.

media.tenor.com

1 1

Will Lowe @conjugateprior.org · 1d

Perhaps we could persuade the Greens to put together complementary legislation that would rename Gummibären and suchlike?

1 1

Will Lowe @conjugateprior.org · 1d

Maybe grab a Fruchtzwerg or a Flutschfinger for after then? That should cover the fat and gristle. Unless you're telling me that...

1 1

Will Lowe @conjugateprior.org · 1d

Aha, das muss das traditionelle amerikanische Spiel "Hide the Sausage" sein, von dem ich so viel gehört habe.

Will Lowe @conjugateprior.org · 1d

und laut Wittgenstein "Wovon man nicht sprechen kann, da es ist ja über ekel digga, muss man schweigen".

1 1

Will Lowe @conjugateprior.org · 2d

Gegen Kant, es ist nicht dass man das Ding-an-sich der Wurst nicht sehen kann, sondern dass man es nicht sehen will.

1 1

Will Lowe @conjugateprior.org · 2d

Kostenlose schlechte Werbeidee: Eine vegetarische Wurst namens (natürlich) "Vurst" und beworben von einer Cartoon-Wurstfigur namens "Fürst von Vurst". Wenn ein Gegner gebraucht wird, regiert der "Worst von Wurst" im nächsten Fürstentum entlang.

Ganz ehrlich, hätte ich im Marketing arbeiten sollen.

1 1

Will Lowe @conjugateprior.org · 2d

Let me guess, the underneath says "und wir mögen es so"

Will Lowe @conjugateprior.org · 2d

I mean, I do live here 😉

I suspect we keep Dr Marten's in business too.

1 1

Will Lowe @conjugateprior.org · 2d

Berlin not beating the '90s teen bedroom decor' rap here 😁

www.slowtravelberlin.com @slowberlin.bsky.social · 2d

Mucha in Pankow ❤️ #jungendstil #berlin

1 3

Will Lowe @conjugateprior.org · 2d

I may be missing something, but there's nothing special about 15 years ago either. Poland just... grows.

Polish GDP yearly growth rate 1991-2024. Looks like 2-6% a year since 1991

Will Lowe @conjugateprior.org · 2d

An uncharitable interpretation is that that is a device for making sure you don't accidentally kill the person you're torturing.