currently: data science / optical @ NFL prev: data science @ GSW sometimes bball analytics moreso on bsky: https://teapawt.net/t.co/h0HHMOZYwJ
Joined November 2009
- Tweets50.9K
- Following4.4K
- Followers8K
- Likes64.9K
Sonnet 5.5 is free, and with my AI plus (or whatever) I can't even get Gemini 3.7 flash, not to mention 3.8 or Argon
Practically speaking, these models don't even exist!
there is no stat less informative or helpful than single game plus minus. a statistical fact and a logical one!
Great way of explaining small sample size effects
The undefeated Wolves are outscoring their opponents by 23+ per game, with +18 of that coming from their performance on 3s.
For the season last year, the best team gained +3.3 per game on 3s, so this is definitely small sample size theater.
espnanalytics.com/nba-team-4…
QTd here teapawt.net/KlawByke/status/211419…
Small Sample Size Alert:
way-too-early OpenRAPM 2027 numbers:
1. Dyson
2. Kessler
3. Jalen Johnson
4. Flagg
5. Giannis
6. Moussa
7. Booker
8. PG
9. Sabonis
10. Chet
openrapm.com/#season=2027
#7stealspergame #lol
in a quite small sample, PHI looks pretty meh, except for Maxey who looks pretty MVP
openrapm.com
reminder that "shot quality adjusted final scores" are an incomplete picture
single game plus minus is the single least useful box score stat in basketball. It is almost never indicative of anything other than confirmation bias.
this is an indisputable fact as real as gravity, the sun coming up, and water being wet.
also - Boozer is ridiculously good
