An Nguyen

Here's a cool write-up that uses a slightly different method to fit a model I experimented with back in 2022, but goes into much more detail.

Thanks! Yeah someone mentioned this, and it definitely was not something I picked up but it is super cool and makes So much sense to find out. i had put together the typeline connection with Kaheera the Orphanguard, but not that they were color-pie related (and that Kaheera follows the same pattern)

joke unassembled, free to a good home: hockey, mtg cube, ice cube

but also hell yeah, i remember y'all posting what must be years ago now that y'all had assembled all the og dual lands for something but forgot y'all played cube until getting into cube myself recently

Ageless Sentinels: An Elo Ratings Approach to Magic: The Gathering's Creature Type Order | An Odd and An End only one of each, though

you ever have a really stupid idea and then a week later you've written 6,000 words and created a whole blog:

anoddandanend.com/posts/2025/m...

it features #magicthegathering, linguistics, and questionable applications of statistical methods. all entirely on brand for me

Ageless Sentinels: An Elo Ratings Approach to Magic: The Gathering's Creature Type Order | An Odd and An End only one of each, though

you ever have a really stupid idea and then a week later you've written 6,000 words and created a whole blog:

anoddandanend.com/posts/2025/m...

it features #magicthegathering, linguistics, and questionable applications of statistical methods. all entirely on brand for me

i just realized the Go-Shintai creatures in #magicthegathering don't have creature types? They all have subtype Shrine, but that's an enchantment type, not a creature type.

(Also, I'm working on writing something (very unusual) Very silly (very on par for me), can you tell.)

hi i am Not back on social media but i tell people enough that something shared and bad >>>> perfect and never done/public, so!

score-charts-only.netlify.app

this was assembled from scraps over ~10 hours across the past 2 days, and is kinda a disaster on every axis, but! it is out in the world now

hi i am Not back on social media but i tell people enough that something shared and bad >>>> perfect and never done/public, so!

score-charts-only.netlify.app

this was assembled from scraps over ~10 hours across the past 2 days, and is kinda a disaster on every axis, but! it is out in the world now

and that wraps #ASIS2025 ! i had an absolute blast; i hope y'all got some measure of enjoyment from my random musings throughout

hearing jen cooper talk about archival and her work is a) incredibly incredibly cool and (b) just makes me think of the iihf flood story, which has haunted me forever since @quarkyhockey.bsky.social and @alyssalongmuir.bsky.social told me about it #ASIS2025

and the close of the conference on a more fun lighthearted note: a live podcast ft. the unparalleled @arielledror.bsky.social and the incredibly cool jen cooper! #ASIS2025

this seems to be a partial call to action for nwsl league translation factors, but with a more personal edge too (players who have already moved to an international league before making a similar significant move to the nwsl, players who speak english) #ASIS2025

both the opening and closing industry keynote very focused on recruitment and roster building, with an emphasis on the money/salary cap side of things for the nwsl+mls #ASIS2025

closing industry keynote time #ASIS2025

maybe the sports analytics conference tracking data bingo should include convex hulls #ASIS2025

the final skillcorner presentation! again a defensive bent, focusing specifically on transition #ASIS2025

another v important and not v talked about thing (normconf-esque): logging #ASIS2025

nevermind we've Now checked the strained backronym box on the sports analytics conference bingo (ty @tanho.ca ) #ASIS2025

the data ingestion monitoring dashboard and slack integration is super smart and extremely useful and v reminsicent to monitoring stuff at my work, but is not necessarily the kind of dashboard someone starting would think about making #ASIS2025

hell yeah the least talked about part of dashboards/any sort of web app in these sports analytics contexts: deployment (esp production deployment) #ASIS2025

give a person a report, you can answer their question. give a person a dashboard, you can make them answer their questions (until things change with the data or the question) #ASIS2025

last block of presentations! intro of the wonderful Lydia starts with a pun, to applause (its me, i applauded) #ASIS2025

ah the classic: distance to goal in some form is always Far and Away the most important contributor to an xG model #ASIS2025

wrapping up before the next break with a general summary of various trends in women's soccer. very simple, but well-explained and useful for me in particular. transitioning to the applicability of men's soccer models to women's soccer #ASIS2025

also i don't think i've emphasized it enough how good i think the scheduling of this conference is. frequent breaks, more practical fun presentations to start then more stats-dense presentations concentrated later in the conference, a light closer, forced collaboration, the list goes on #ASIS2025

spicy: opting for expected in "expected possession value prevented" abbreviated as EPVp rather than something like xPVp (though, typing it out, I can see why) #ASIS2025

another skillcorner one. this one's on 1v1 defending. defending in general seems to be a hot topic for these. not sure if that's based on the data provided (they mention two ways to think about defense as winning the ball (event) or positioning (tracking)) or if that was a comp theme #ASIS2025

there's an LLM talk where one of the AI use suggestions is supplementing referees, and i was simply haunted by the echoes of all the Real Bad rulebooks i have read in the name of shot-plotter. and i only read maybe ~5 pages of each of them. #ASIS2025

i'm glad you are finding valued and also impressed, bc i feel like often my comments have Very little to do of the actual content of the presentations i'm seeing. would highly recommend reaching out to The presenters too/instead for actual information

this one is using Both tracking data and the extra data! high line based on y position of centerbacks from the tracking data, and then evaluating it based on the provided, derived xThreat metric #ASIS2025