Dan Larremore
@danlarremore.bsky.social
Prof | CU Boulder BioFrontiers Institute + Computer Science | Santa Fe Institute
LarremoreLab.github.io
LarremoreLab.github.io
I don't know who made this but 10/10.
![In this meme, a young John Connor is talking to the Terminator.
Wait a second, you re saying you're from the future?
Great question! Thanks so much for asking. Let's think about the answer...hmm! The current year is 1983. Working backwards logically. I'm from the year 1965, which is 300 years in the future. Anything else you'd like to know?
[Blank stare]
I'm sorry, I can see that was not the right answer. I actually came from the year 829392 AD, which is 10 years in the past from this year, which is](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:qhc5hk27hzotcxv7cehkqvpd/bafkreidzwql5qqyhv26bfqgc2mh46lbkopr5mwf6k3vkbggrijqz3xd4am@jpeg)
July 17, 2025 at 8:11 PM
I don't know who made this but 10/10.
A nice behind the scenes thread from Johan Ugander for the 2012 "4 Degrees of Separation" paper, computed on the Facebook graph, which just won the WebSci "Test of Time" award. 👏
This Fig showing convergence of avg distances to 4.74 (i.e. 3.74 deg of sep) as FB graphs assembled together is 🧑🍳🤌😘
This Fig showing convergence of avg distances to 4.74 (i.e. 3.74 deg of sep) as FB graphs assembled together is 🧑🍳🤌😘
May 23, 2025 at 7:17 PM
A nice behind the scenes thread from Johan Ugander for the 2012 "4 Degrees of Separation" paper, computed on the Facebook graph, which just won the WebSci "Test of Time" award. 👏
This Fig showing convergence of avg distances to 4.74 (i.e. 3.74 deg of sep) as FB graphs assembled together is 🧑🍳🤌😘
This Fig showing convergence of avg distances to 4.74 (i.e. 3.74 deg of sep) as FB graphs assembled together is 🧑🍳🤌😘
This table, compiled by Carol J Auster for the @aaup.bsky.social caught my eye. It's remarkable how social norms in academic research vary so dramatically by area of study...
Table from www.aaup.org/article/unde... with more details available in additional tables www.aaup.org/sites/defaul...
Table from www.aaup.org/article/unde... with more details available in additional tables www.aaup.org/sites/defaul...
May 15, 2025 at 5:17 PM
This table, compiled by Carol J Auster for the @aaup.bsky.social caught my eye. It's remarkable how social norms in academic research vary so dramatically by area of study...
Table from www.aaup.org/article/unde... with more details available in additional tables www.aaup.org/sites/defaul...
Table from www.aaup.org/article/unde... with more details available in additional tables www.aaup.org/sites/defaul...
I’ve always liked that they also asked about quantities! The title of the plot is off, but directly from the same repo and study:
raw.githubusercontent.com/zonination/p...
raw.githubusercontent.com/zonination/p...
May 3, 2025 at 7:58 PM
I’ve always liked that they also asked about quantities! The title of the plot is off, but directly from the same repo and study:
raw.githubusercontent.com/zonination/p...
raw.githubusercontent.com/zonination/p...
Ooof. Yes. And the dedication, which I skipped right over the first time, is so beautiful:
April 3, 2025 at 6:30 PM
Ooof. Yes. And the dedication, which I skipped right over the first time, is so beautiful:
CDC's going to have to launch a new journal...
March 26, 2025 at 9:52 PM
CDC's going to have to launch a new journal...
While we're swapping, here's a one-page description of roughly when to use which type of colormap, screencapped from the Data Viz slides on larremorelab.github.io/slides/. Hope this is useful!
February 12, 2025 at 3:57 AM
While we're swapping, here's a one-page description of roughly when to use which type of colormap, screencapped from the Data Viz slides on larremorelab.github.io/slides/. Hope this is useful!
Here's a crack at it! Thanks for posting the data and calculations. And for the tip on Datawrapper. That's a nice tool!
February 12, 2025 at 3:37 AM
Here's a crack at it! Thanks for posting the data and calculations. And for the tip on Datawrapper. That's a nice tool!
😅 Apologies for being flippant! The red-blue "diverging" colormap across the US states threw me, as the values are sequential. During Datawrapper Step 3 (Visualize) they have some nice options for sequential colormaps.
February 12, 2025 at 3:33 AM
😅 Apologies for being flippant! The red-blue "diverging" colormap across the US states threw me, as the values are sequential. During Datawrapper Step 3 (Visualize) they have some nice options for sequential colormaps.
Blatant image duplication and manipulation in Fig. 5 of one of my son's favorite publications. And yet, no calls for retraction!?
@retractionwatch.com @elisabethbik.bsky.social
@retractionwatch.com @elisabethbik.bsky.social
January 5, 2025 at 9:52 PM
Blatant image duplication and manipulation in Fig. 5 of one of my son's favorite publications. And yet, no calls for retraction!?
@retractionwatch.com @elisabethbik.bsky.social
@retractionwatch.com @elisabethbik.bsky.social
A 1985 editorial position from @science.org introduces the Board of Reviewing Editors. Goals:
1. Help w submission volume.
2. Outside opinions to help return 60% of papers to authors without review within 10d. Review the rest.
3. P(accept | top BORE rating)≈50%
Initial BORE was really something! 👇
1. Help w submission volume.
2. Outside opinions to help return 60% of papers to authors without review within 10d. Review the rest.
3. P(accept | top BORE rating)≈50%
Initial BORE was really something! 👇
December 17, 2024 at 4:42 PM
A 1985 editorial position from @science.org introduces the Board of Reviewing Editors. Goals:
1. Help w submission volume.
2. Outside opinions to help return 60% of papers to authors without review within 10d. Review the rest.
3. P(accept | top BORE rating)≈50%
Initial BORE was really something! 👇
1. Help w submission volume.
2. Outside opinions to help return 60% of papers to authors without review within 10d. Review the rest.
3. P(accept | top BORE rating)≈50%
Initial BORE was really something! 👇
For me, footnote 7 sums up our process *over the past seven years* best:
October 10, 2024 at 5:39 PM
For me, footnote 7 sums up our process *over the past seven years* best:
The fig above (w odd) and the fig below (w even) are used to show that an explicit strategy for X (top flows, orange) can avoid danger, no matter what O does (bottom flows, blue).
As you can see, even and odd are important here, too. (See paper for definitions of these automata.)
As you can see, even and odd are important here, too. (See paper for definitions of these automata.)
October 10, 2024 at 5:35 PM
The fig above (w odd) and the fig below (w even) are used to show that an explicit strategy for X (top flows, orange) can avoid danger, no matter what O does (bottom flows, blue).
As you can see, even and odd are important here, too. (See paper for definitions of these automata.)
As you can see, even and odd are important here, too. (See paper for definitions of these automata.)
The above result is cool because it says that X must avoid connecting 3, but O can do whatever they want. Still, X can avoid three-in-a-row!
Our analysis relies on describing the board through a set of canonical sub-boards as automata, and analyzing flows thereof. Here's a fig of flows for odd w.
Our analysis relies on describing the board through a set of canonical sub-boards as automata, and analyzing flows thereof. Here's a fig of flows for odd w.
October 10, 2024 at 5:33 PM
The above result is cool because it says that X must avoid connecting 3, but O can do whatever they want. Still, X can avoid three-in-a-row!
Our analysis relies on describing the board through a set of canonical sub-boards as automata, and analyzing flows thereof. Here's a fig of flows for odd w.
Our analysis relies on describing the board through a set of canonical sub-boards as automata, and analyzing flows thereof. Here's a fig of flows for odd w.
For Misère Connect k≥3 on a w-by-h board, we have a wide variety of results, and same for k=2.
So it turns out that the space of optimal play depends on the interaction between the size of the board and the number to connect... but in a way that reduces to whether h and w and odd or even!
So it turns out that the space of optimal play depends on the interaction between the size of the board and the number to connect... but in a way that reduces to whether h and w and odd or even!
October 10, 2024 at 5:26 PM
For Misère Connect k≥3 on a w-by-h board, we have a wide variety of results, and same for k=2.
So it turns out that the space of optimal play depends on the interaction between the size of the board and the number to connect... but in a way that reduces to whether h and w and odd or even!
So it turns out that the space of optimal play depends on the interaction between the size of the board and the number to connect... but in a way that reduces to whether h and w and odd or even!
It is with great pride that I announce:
Misère Connect Four is Solved
Misère games flip the typical win condition, so here we ask: Who wins when both you and your opponent try to avoid connecting four?
arxiv.org/abs/2410.05551
Misère Connect Four is Solved
Misère games flip the typical win condition, so here we ask: Who wins when both you and your opponent try to avoid connecting four?
arxiv.org/abs/2410.05551
October 10, 2024 at 5:15 PM
It is with great pride that I announce:
Misère Connect Four is Solved
Misère games flip the typical win condition, so here we ask: Who wins when both you and your opponent try to avoid connecting four?
arxiv.org/abs/2410.05551
Misère Connect Four is Solved
Misère games flip the typical win condition, so here we ask: Who wins when both you and your opponent try to avoid connecting four?
arxiv.org/abs/2410.05551
Academics! Where do you prefer or hope to publish?
Have a few minute to tell us? My group at CU Boulder is surveying the publication preferences of academics in a pairwise comparison survey called Publish or Comparish.
→ tinyurl.com/PubPrefs
Thanks for considering & pls share with colleagues!
Have a few minute to tell us? My group at CU Boulder is surveying the publication preferences of academics in a pairwise comparison survey called Publish or Comparish.
→ tinyurl.com/PubPrefs
Thanks for considering & pls share with colleagues!
September 13, 2024 at 6:53 PM
Academics! Where do you prefer or hope to publish?
Have a few minute to tell us? My group at CU Boulder is surveying the publication preferences of academics in a pairwise comparison survey called Publish or Comparish.
→ tinyurl.com/PubPrefs
Thanks for considering & pls share with colleagues!
Have a few minute to tell us? My group at CU Boulder is surveying the publication preferences of academics in a pairwise comparison survey called Publish or Comparish.
→ tinyurl.com/PubPrefs
Thanks for considering & pls share with colleagues!
Join our growing, multi-disciplinary community for an amazing science+policy line-up at the International Conference on the Science of Science and Innovation (ICSSI) at the National Academy of Sciences in D.C. July 1-3! 🔬
Abstracts due March 24! 👉 www.icssi.org.
Abstracts due March 24! 👉 www.icssi.org.
February 7, 2024 at 8:11 PM
Join our growing, multi-disciplinary community for an amazing science+policy line-up at the International Conference on the Science of Science and Innovation (ICSSI) at the National Academy of Sciences in D.C. July 1-3! 🔬
Abstracts due March 24! 👉 www.icssi.org.
Abstracts due March 24! 👉 www.icssi.org.
In fact, Nobelists' frustrations with the dynamics of funding, prestige, and discovery are basically an annual tradition!
From our 2017 essay in Science, we quote Ohsumi (2016, Medicine) on funders' preference for low-risk proposals.
www.science.org/doi/full/10....
From our 2017 essay in Science, we quote Ohsumi (2016, Medicine) on funders' preference for low-risk proposals.
www.science.org/doi/full/10....
October 3, 2023 at 6:14 PM
In fact, Nobelists' frustrations with the dynamics of funding, prestige, and discovery are basically an annual tradition!
From our 2017 essay in Science, we quote Ohsumi (2016, Medicine) on funders' preference for low-risk proposals.
www.science.org/doi/full/10....
From our 2017 essay in Science, we quote Ohsumi (2016, Medicine) on funders' preference for low-risk proposals.
www.science.org/doi/full/10....
Lovely skies last night in Boulder, CO. ⛰️
August 29, 2023 at 2:20 PM
Lovely skies last night in Boulder, CO. ⛰️