Tim Morris

Ooh interesting! Relevant to what I was thinking about, so thanks. I’ll see if I can find it.

Yes! Thanks very much

That's the one! Thanks Cameron. At least it's very explicit now.

It was logistic regression in scikit learn (scikit-learn.org/stable/modul...). Apparently, you could not fully turn off regularisation there for some time.

Aaah it was the scikit learn part I forgot - thanks Björn!
(Can I acknowledge you in some slides where I mention this?)

A few years ago there was a hoo-ha when someone discovered a well-used implementation of logistic regression in python used some regularization by default. Can anyone recall more details (or point me to a link because I'm clearly a bad googler)?

I'm just using typst in-browser where I think we have an interactive overleaf sitch. I'll DM both of you a link to a test doc now so we can see.

Introducing the Simulation Studies Preregistration Template: Q&A with Björn S. Siepe, František Bartoš, and Samuel Pawel Initially submitted through COS’s open call for community-designed preregistration templates, the Simulation Studies Template is now part of the expanding collection of preregistration resources available on the OSF.

Simulation studies are widely used, but have historically lacked clear guidance around preregistration.

In response to this gap, a team of researchers have developed a new template for preregistering simulation studies, now available on OSF.

🚀 Read our Q&A: www.cos.io/blog/intr...

Just started playing with typst a few weeks back. I've written a few short documents and it's been good for that. It looks like it's under active development and some quite basic things don't appear to be documented. So I'm not yet confident to write a paper or slides, say. Still, worth a look!

With typst it’s instantly clear what’s exciting… I just can’t work out how make a typst template to follow Novartis brand guidelines.

Doing slides for ISCB with quarto (since they accept it I thought why not) and don’t really see what the fuss is about – with two slides it takes ~10s to compile. Looks ok.

Genuine question: what is exciting about it? Is it code-related stuff?

Oh nice, thanks for looking into this Sean! That’s reassuring.

Good question… not having a good answer, I’ll go with ‘I’m a busy man who doesn’t have time to type “1.96*” twice!’

Same. This blog gives is a really good walkthrough of what tmle doing (doesn’t tell you why it works but that’s not what I needed).
Helped me finally understand what a “clever covariate” is!
www.khstats.com/blog/tmle/tu...

😂

Things like their dad can’t speak it? 😜

Oh not you too Matt

Thanks, and same! I realised I was saying this a lot and should probably check.

In draf, I had written “randomized” and “randomization” but changed them all to “randomised” and “randomisation” just to spite @faustobustos.bsky.social, @solomonkurz.bsky.social and @pausalz.bsky.social, who didn’t know how to compromise.
bsky.app/profile/timp...

Cover picture with blog title & subtitle, and results graph in the background

Post just up: Is multiple imputation making up information?

tldr: no.

Includes a cheeky simulation study to demonstrate the point.
open.substack.com/pub/tpmorris...

You’re saying you currently pronounce it “zee-eye-vich”?? Didn’t see that one coming.

I was going to make one exception but your escalation means we will be pronouncing it “Jay-zed” henceforth

Once did a simulation study where I evaluated coverage of 94% confidence intervals. People kept interpreting 94% as undercoverage. Next time I went for 89% to make it clearer. Was asked to stop the weird distractions… Fair point: 0.05 is so ingrained that this is distracting!

Alright Americans, I know you’re having a hard time with stuff at the moment, so here’s an olive branch:
I’ll start spelling it randomization if you start pronouncing the letter Z as “zed” instead of “zee”.

LMK

I’m glad you brought up sons. All the good photos I get are with my son, which you can’t use as a headshot. Also true of sunglasses.

I was thinking you have to give yourself more slack as you get older, like change it every age/10 years or something

When was your current profile pic taken? Just wondering if I’ve ever seen you with a different one…

What is the right update frequency I’m losing my mind here

The same-headshot-for-quarter-of-a-century strategy also appears vain because you’re saying you’re unwilling to use a photo with more wrinkles & grey hairs

It’s a fine line to tread between using the same headshot for quarter of a century and updating it so often you appear vain

Interesting! Yes, it’s possible to get that with FPs but I never saw it in real data. My recollection is that Royston and Sauerbrei’s book has some practical advice on this sort of thing, though I forget exactly what it is (it’s been a while)!