Mason Veilleux @mveilleux.com · 09/09/2026This is a good litmus test to tell if someone is Bayesian enough: do you know who this is? 000
Mason Veilleux @mveilleux.com · 06/09/2026I just added CFB a few days ago, but NFL is more mature. You can see how weird points posteriors look in football for this upcoming game: gameplan-science.com/nfl/game/202... 110
Mason Veilleux @mveilleux.com · 06/09/2026I don’t have points posterior distributions yet for CFB but per drive basis, it was a coin flip match. (and looking at this I think I fucked up and diffed by log scale) 110
Mason Veilleux @mveilleux.com · 09/08/2026Okay so with the prior that scales with the size of reviews, I think I got something going. 100
Mason Veilleux @mveilleux.com · 06/08/2026My wife says to stop giving him dreams, but he needs to be punished for his crimes 000
Mason Veilleux @mveilleux.com · 04/08/2026I made a family chat bot for my wife and I (using Hermes) and by far the most fun part has been developing the voice and lore of the bot. I have the bot say good morning and recount the dream it had, example: 000
Mason Veilleux @mveilleux.com · 22/07/2026Some friends to moved to Germany (German citizens) so why not. 000
Mason Veilleux @mveilleux.com · 22/07/2026FG model is looking money. Here is kicker skill distributions over time and centered. 100
Mason Veilleux @mveilleux.com · 17/07/2026Hear me out. A website of a Bradley-Terry model for every sport. 010
Mason Veilleux @mveilleux.com · 14/07/2026Learned something new in Stan where you can have two priors — one for regularizing and the other for the actual data you want to inform the model with. 000
Mason Veilleux @mveilleux.com · 02/07/2026The test run looked good for 2023-2025 seasons. This is 2025 offensive strength estimates against PPG 000
Mason Veilleux @mveilleux.com · 02/07/2026I bet I’m going to find the dumbest bug in here and I’m going to have to re-run this 000
Mason Veilleux @mveilleux.com · 23/06/2026Also league-wide trend of fewer drives per game. Not sure exactly what’s driving, but sure would think it has to do with QB’s /offensive playstyle mostly.. 100
Mason Veilleux @mveilleux.com · 23/06/2026This plot was made out of spite for all the minutes (maybe hours?) playing the Colts in Madden throughout my childhood. 200
Mason Veilleux @mveilleux.com · 23/06/2026On the frontier of scientific research laying next to my sleeping toddler on my phone running an LLM on Chinese GPUs just like the good lord intended 000
Mason Veilleux @mveilleux.com · 22/06/2026I’ve been working on a geometry-based model for NFL kicker latent skill (much like Gelman’s putting model) .. maybe I’m pooling too much but I’ve got very weak priors on the hierarchy relationship? Still working out the kinks 130
Mason Veilleux @mveilleux.com · 06/06/2026Waiting for the network engineers to wrap up so I can deploy the EOD Friday hot fix 010
Mason Veilleux @mveilleux.com · 25/05/2026This is what Alaska rewards you with when you don’t let the voices win during the winter. 000
Mason Veilleux @mveilleux.com · 28/04/2026I will say again that if your AI does a bad thing, you are at fault. 000
Mason Veilleux @mveilleux.com · 23/04/2026I used AI to make what I thought would be how PewDiePie does his family blogs but instead it looks like toddler is an Elden Ring boss. 100
Mason Veilleux @mveilleux.com · 03/04/2026It only ends one way for evil pikachu in Pokémon: The Kanto Convention 010
Mason Veilleux @mveilleux.com · 27/11/2025Stuffing done, roasted veggies done, I’m locked in boys 000
Mason Veilleux @mveilleux.com · 26/07/2025Logged onto FB and now I’ve been staring at my hand like I’m ngmi 010
Mason Veilleux @mveilleux.com · 13/07/2025I spun up an Instant SQL Neovim plugin last night. Ever since watching @hamilton.bsky.social 's presentation, I feel like a caveman doing the old "code→run->debug" loop 100
Mason Veilleux @mveilleux.com · 25/05/2025Okay I needed to look this up because I didn’t think the price was right. It’s 4.25 per person 010
