Pinned post

I wrote a little post about the war of attrition (game theory) with two players. It's intended mostly for a graduate audience:
mattwthomas.com/blog/war-of-at

All of this was in other places, but it took me a while to understand this fully and felt it was good to share it all in one place since it would have saved me a lot of time.

Pinned post

I just finished a post about QF (quadratic finance) which some are pushing as the "mathematically optimal" way to fund open source.

mattwthomas.com/blog/fund-open

I hope to clear up some of the confusion that I've heard about it and point out some problems with the way that people outside of Economics are interpreting the original paper.

I decided at start hosting a US CTAN (TeX) mirror. I get about 350 GB of download traffic per day. Though, when I started, I got hit with 2TB per day! I think it was just a lot of projects benchmarking the new location.

Anyway, I freaked out and upgraded the hardware a lot. So, it should be pretty fast.

I'll probably write an article about this since I couldn't find any info about the bandwidth required.

Me explaining ounces:

"They are usually a unit of mass -- unless it's a liquid. In that case, it's a unit of volume." 🤷

I figured it out! This is the survival function of the binomial distribution, and the survival function of binomial distribution is strictly increasing in p (except in the case of minimum).

Therefore, a unique inverse exists, but there is no closed form solution except in the special case of minimum and maximum.

Show thread

Does anyone know how to retrieve the distribution of a random variable from the distribution of it's order statistic? The CDF of the order statistic is

$$
F_{X_{(r)}}(x) = \sum_{j=r}^{n} \binom nj [ F_{X}(x) ]^{j} [ 1 - F_{X}(x) ]^{n-j}
$$

I know $F_{X_{(r)}}(x)$, but want to get $F_{X}(x)$.

Latex rendered: ttm.sh/uf8

I'm giving hosting my own (pleroma) instance a try! For now, it's a single user instance.

I'm interested in advice that people might have on this. There's a bunch that I haven't setup yet (like blocklists).

Every working paper I write should exist within the same academic universe with consistent notation and characters (in the examples)

I have tried displaying abstracts via use of checkboxes and anchor links on my site. Which is better?

Checkbox:
mattwthomas.com

Anchor:
mattwthomas-com-git-patch-1-mw

Do people recommend using `rel="shortlink"` in 2021? I'm confused about whether this is still a pending standard that people like or if it's abandoned.

I set up a Google routine to turn my lights off when no one is in the house. It's nice because it saves electricity.

What's less nice is that I now get a notification every time my partner leaves the house (because the lights are turning off) *creepy*

Me, early this morning: curses, why is this bag of coffee so hard to open?

Me, a short while later: why is the writing on this bag of coffee upside-down?

I need to solve the problem of a minimal graph coloring with at most k vertices of one color and you can "split" vertices such that each vertex can be split into any number of vertices each with a partition of the original's edges.

This is hard yet surprisingly useful.

Profanity, Freenode. 

You have got to be fucking kidding me.

% host irc.libera.net
irc.libera.net is an alias for chat.freenode.net.
chat.freenode.net has address 104.237.198.130
chat.freenode.net has address 64.44.40.114
chat.freenode.net has address 54.37.136.225

Very funny.

I always think that original tex code (or any code in fact) is a bit private, mostly because I usually don’t want to expose my stupidity and inefficiency in coding. For example, I wouldn’t want anyone to see my “\newcommand{\w}{\sigma}” under any circumstance. (Now you are seeing it only because I didn’t write it—I got it in a template from someone else who’s kind enough to share their tex files.)

An old April Fool's day article I found amusing:

"On the Ratio of Circumference to Diameter for the Largest Observable Circles: An Empirical Approach"

arxiv.org/abs/1204.0298

Finally, an applied approach to Euclidean geometry.

Show older
Mathstodon

The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!