Zach
@zachnielsen.bsky.social
Analytics, sports, trying to figure out Bayesian math one step at a time
Playing around a bit on visualizing predictions. One thing that's been pretty challenging is the over under variation. Turns out, every hockey game lands at 6, so it's real easy to have a model just predict 6 for every game.
January 7, 2025 at 12:50 AM
Playing around a bit on visualizing predictions. One thing that's been pretty challenging is the over under variation. Turns out, every hockey game lands at 6, so it's real easy to have a model just predict 6 for every game.
Excellent point, as I look to the soy sauce I have for exactly one recipe
i think some of the Food Discourse comes down to limited pantries. if you have to buy a $5 bag of flour, $10 bottle of vanilla, $3.50 jar of baking powder, etc to make cookies, the alternatives start to look cheaper (especially when you want 1 cookie now and not 100 over a year or two)
sorry, but “it is more cost effective to prepare food at home vs ordering delivery” is just true? you can do whatever makes sense for your life but that fact remains.
January 6, 2025 at 1:50 AM
Excellent point, as I look to the soy sauce I have for exactly one recipe
and, projected division winners! Not sure how accurate it is, the points seem quite low to me.
Projected Division Winners:
Atlantic: TBL (70.2 pts)
Metropolitan: WSH (65.9 pts)
Central: DAL (72.4 pts)
Pacific: VGK (72.1 pts)
Projected Division Winners:
Atlantic: TBL (70.2 pts)
Metropolitan: WSH (65.9 pts)
Central: DAL (72.4 pts)
Pacific: VGK (72.1 pts)
January 5, 2025 at 6:21 AM
and, projected division winners! Not sure how accurate it is, the points seem quite low to me.
Projected Division Winners:
Atlantic: TBL (70.2 pts)
Metropolitan: WSH (65.9 pts)
Central: DAL (72.4 pts)
Pacific: VGK (72.1 pts)
Projected Division Winners:
Atlantic: TBL (70.2 pts)
Metropolitan: WSH (65.9 pts)
Central: DAL (72.4 pts)
Pacific: VGK (72.1 pts)
Quick update on hockey modeling, with more to come. Had a not interesting refactor coming in, along with LLMs leading me astray and causing complication by overcomplicating the code.
January 5, 2025 at 6:19 AM
Quick update on hockey modeling, with more to come. Had a not interesting refactor coming in, along with LLMs leading me astray and causing complication by overcomplicating the code.
Day two of NHL analysis - so fun fact, predictions from yesterday were based on simple Elo calculations. TOR did cash (didn't check on the two underdogs). Also missed a few teams in the final upload, we regret the error!
January 1, 2025 at 6:00 AM
Day two of NHL analysis - so fun fact, predictions from yesterday were based on simple Elo calculations. TOR did cash (didn't check on the two underdogs). Also missed a few teams in the final upload, we regret the error!
Great link from FiveThirtyEight, smart folks
https://fivethirtyeight.com/methodology/how-our-nhl-predictions-work/
https://fivethirtyeight.com/methodology/how-our-nhl-predictions-work/
How Our NHL Predictions Work
References Hockey-Reference.com / Elo ratings / Autocorrelation / Monte Carlo simulations / Logistic regression / Poisson distribution The Details In 2021, we a…
fivethirtyeight.com
December 31, 2024 at 4:58 AM
Great link from FiveThirtyEight, smart folks
https://fivethirtyeight.com/methodology/how-our-nhl-predictions-work/
https://fivethirtyeight.com/methodology/how-our-nhl-predictions-work/
Having lived in Las Vegas with our own NHL team, I've decided to get into hockey for the first time.
Specifically, I want to explore the NHL through analytics first principles, primarily through an analytics and sports betting lens (two interests, for fun).
Specifically, I want to explore the NHL through analytics first principles, primarily through an analytics and sports betting lens (two interests, for fun).
December 31, 2024 at 4:56 AM
Having lived in Las Vegas with our own NHL team, I've decided to get into hockey for the first time.
Specifically, I want to explore the NHL through analytics first principles, primarily through an analytics and sports betting lens (two interests, for fun).
Specifically, I want to explore the NHL through analytics first principles, primarily through an analytics and sports betting lens (two interests, for fun).
Really cool use of LLMs to capture sponsorship dollars - a hard problem, fun to see innovation here
https://www.kaggle.com/code/oleksandrarsentiev/exploring-sports-sponsorship-ad-exposure-analysis#F1:-Grand-Prix-Las-Vegas
https://www.kaggle.com/code/oleksandrarsentiev/exploring-sports-sponsorship-ad-exposure-analysis#F1:-Grand-Prix-Las-Vegas
Exploring Sports Sponsorship: Ad Exposure Analysis
Explore and run machine learning code with Kaggle Notebooks | Using data from multiple data sources
kaggle.com
December 20, 2024 at 11:42 PM
Really cool use of LLMs to capture sponsorship dollars - a hard problem, fun to see innovation here
https://www.kaggle.com/code/oleksandrarsentiev/exploring-sports-sponsorship-ad-exposure-analysis#F1:-Grand-Prix-Las-Vegas
https://www.kaggle.com/code/oleksandrarsentiev/exploring-sports-sponsorship-ad-exposure-analysis#F1:-Grand-Prix-Las-Vegas
Had a chance to rewrite JobDescriptionAI.com, which I'm super proud of! AI is getting better writing HTML and CSS. I think it looks great
December 13, 2024 at 4:04 AM
Had a chance to rewrite JobDescriptionAI.com, which I'm super proud of! AI is getting better writing HTML and CSS. I think it looks great
Nothing like better tooling to get in the way of better analytics!
December 11, 2024 at 5:50 AM
Nothing like better tooling to get in the way of better analytics!
I've been exploring better ways to handle context in AI/LLM chats instead of starting fresh each time - nothing worse than having to restart a chat with the classic "Think Step by Step, don't make up numbers, here's the background" etc.
December 11, 2024 at 5:49 AM
I've been exploring better ways to handle context in AI/LLM chats instead of starting fresh each time - nothing worse than having to restart a chat with the classic "Think Step by Step, don't make up numbers, here's the background" etc.
As a UNLV fan it’s been fun to be nationally relevant for football! For the first time…ever? (Please ignore our 3rd down play calling)
It’s fine and 100% justified to have complaints about the current state of college football but we are about to watch a UNLV-Boise State playoff play-in game on the blue turf and that kicks 10,000 tons of ass
December 7, 2024 at 2:44 AM
As a UNLV fan it’s been fun to be nationally relevant for football! For the first time…ever? (Please ignore our 3rd down play calling)
Love this. It also supports my hot take that everyone loves data, people just need it visualized appropriately
If you teach a basic data analysis class, Spotify Wrapped is a fun way to show your students you can create billions in PR value with what amounts to a filter, a group by, a count and a sort.
December 4, 2024 at 5:09 PM
Love this. It also supports my hot take that everyone loves data, people just need it visualized appropriately
VERY thankful for Croissant being available on Mac! Super great app for posting
November 29, 2024 at 5:24 PM
VERY thankful for Croissant being available on Mac! Super great app for posting
Having to worry about regular maintence will be the thing that drives me to an electric car my goodness
November 28, 2024 at 12:46 AM
Having to worry about regular maintence will be the thing that drives me to an electric car my goodness
Working on Bluesky Analytics, and it's coming together! Very excited.
November 27, 2024 at 8:27 PM
Working on Bluesky Analytics, and it's coming together! Very excited.
Hard part about building an analytics tool is when you have no test data because you never post. Hope my calculations are all correct!!
November 27, 2024 at 6:57 PM
Hard part about building an analytics tool is when you have no test data because you never post. Hope my calculations are all correct!!
Me, working on NFL big data bowl 2025: "Claude, what's your POV here on this data."
Claude: "solid reasoning, gets some good stuff, also here's a stat that doesn't make sense."
Me: "Claude, did you make that up?"
Claude: "You're right! I did!"
Claude: "solid reasoning, gets some good stuff, also here's a stat that doesn't make sense."
Me: "Claude, did you make that up?"
Claude: "You're right! I did!"
November 22, 2024 at 7:00 PM
Me, working on NFL big data bowl 2025: "Claude, what's your POV here on this data."
Claude: "solid reasoning, gets some good stuff, also here's a stat that doesn't make sense."
Me: "Claude, did you make that up?"
Claude: "You're right! I did!"
Claude: "solid reasoning, gets some good stuff, also here's a stat that doesn't make sense."
Me: "Claude, did you make that up?"
Claude: "You're right! I did!"
November 21, 2024 at 12:44 AM
@evanmiya.bsky.social welcome! Excited to have you here for some basketball content
October 18, 2023 at 2:27 AM
@evanmiya.bsky.social welcome! Excited to have you here for some basketball content
Could you imagine owning threads.com right now?
July 7, 2023 at 7:29 AM
Could you imagine owning threads.com right now?
Shoutout to all of the brand managers having to write style guidelines for Threads. Best of luck!
July 7, 2023 at 7:25 AM
Shoutout to all of the brand managers having to write style guidelines for Threads. Best of luck!
While I was Flipping through some old photos, I had forgotten about the absolute stranglehold Snapchat had on my life in college. Just crazy seeing how social media changes over time
July 7, 2023 at 7:13 AM
While I was Flipping through some old photos, I had forgotten about the absolute stranglehold Snapchat had on my life in college. Just crazy seeing how social media changes over time
Reposted by Zach
WE WILL RELEASE A NEW TWITTER CLONE EVERY HOUR UNTIL OUR DEMANDS ARE MET
July 6, 2023 at 12:39 AM
WE WILL RELEASE A NEW TWITTER CLONE EVERY HOUR UNTIL OUR DEMANDS ARE MET
ChatGPT’s best feature is giving me the vocab so I can at least begin to google efficiently
June 29, 2023 at 6:16 AM
ChatGPT’s best feature is giving me the vocab so I can at least begin to google efficiently