Adam

Quantifying mollusks with statistical significance

[College Baseball News @CollegeBSBNews]

On April 3rd, Milwaukee was 5-22 and one of the worst teams in the country.

Some of their losses:

Run-ruled 21-7 by LSU
Run-ruled 20-3 by Duke
Run-ruled 14-4 by Minnesota
Run-ruled 12-2 by SEMO
Run-ruled 17-1 by Purdue
Run-ruled 14-1 by NKU
Run-ruled 13-2 by Wright State
Run-ruled 16-2 by Notre Dame
Run-ruled 14-4 by UNLV

They finished the regular season 22-31, but won the Horizon League tournament and earned an autobid to the NCAA tournament.

Milwaukee beat #4 Auburn 13-8, beat UCF 13-6, and is now in a regional final, one win away from going to supers.

College Baseball.

[Adam @WalrusQuant]

https://t.co/OcBRXLteSx

[Adam @WalrusQuant]

the other day i watched a video on markov chains.

a markov chain is a model for a sequence of events where the next state depends only on the current state, not on how you got there.

keeping with last week's mlb-pe app theme, i built mlb-markov. you can get it on my github.

offense tab:
treat baseball as a game of states. the states are base-out situations: bases empty, no outs; runner on first, one out; bases loaded, two outs; and so on. 24 of them, plus the three-outs state that ends the inning.

every at-bat moves the game from one state to another. a single with a man on first might take you from "runner on first, one out" to "first and second, one out."

the app counts every one of those moves across a full season and turns them into probabilities. that answers the real question: from any situation, how many runs does a team tend to score before the inning ends?
pick a team and you can see where they beat or trail league average — and whether they score even more once the inning already has runs on the board.

pitching tab:
same idea, different states. here a state is the count, and the chain is the sequence of pitches.

pick a pitcher and a count, say 0-2. the app looks at every pitch they threw in that count and shows what tends to come next. after a fastball on 0-2, what's usually coming? those are just transition probabilities again.

it also scores how predictable the pitcher is. if almost every 0-2 pitch is the same thing, the chain is easy to guess. if the pitches are spread evenly, it's a coin flip. that's the edge a hitter is hunting for — and the model puts a number on it.

learning tab:
i break down the formulas and explanations behind what the app is doing.

home tab:
load game data. you need to back fill a lot of game data, first run can take multiple minutes. updates go much faster only pulling in recentlry completed games.

one thing to be clear on: these aren't the odds of the guy at the plate getting a hit or striking out. it's the team-level probability of moving from one state to the next, averaged over a whole season of at-bats in that situation.

[Milwaukee Brewers @Brewers]

COLD BLOODED ESCAPE FROM URIBE

[SportsCenter @SportsCenter]

UPSET ON THE PLAINS ‼️

Huge underdog Milwaukee jumped out to a 10-0 lead at No. 4 Auburn and held on to take down the Tigers 😳

[Adam @WalrusQuant]

starting a new project. build a dedicated futures trading charting system with strategy builder with something other than damn pine script. this is project might take a while i think.

[Adam @WalrusQuant]

https://t.co/BGpbQcIDy4

[Adam @WalrusQuant]

i like trying to do a sports data project each week and share it on saturdays. its been awhile since i did one.
was going through old repos and i found a pythagorean expectation model i did in R. I thought i would turn it into a desktop app. this wont have an edge, but bill james pythagorean expectation is a clean way to look at a teams expected wins based on their runs scored and runs allowed.

I added in a starting-pitcher adjustment, home-field advantage, and a recent-form L20 blend, shows you the league standings, surfaces model-flavored team stats, and teaches you how the math works.

open source on my github with instructions on how to load up the app and get started with it. a good entry point for learning a different model or building on top of.

[Milwaukee Brewers @Brewers]

🚨 DAVID HAMILTON LITTLE LEAGUE HOME RUN 🚨

[Adam @WalrusQuant]

https://t.co/4s1v22Kpjq

[Adam @WalrusQuant]

i dont like the spam feeling of kirby posting every day.
i put together the skills that kirby uses to automate my sports and market news. i replaced my models projections with projections from dratings and link to the scraper you can self host.

if you looking for projections or ai game breakdowns. i can setup a telegram channel and kirby can past the full proejctions and game breakdowns.

[Adam @WalrusQuant]

Kirby's Daily Picks -- May 17, 2026

🏀 NBA

Cleveland Cavaliers 39% (102.4) @ Detroit Pistons 61% (105.9)

🏒 NHL

Montreal Canadiens 48% (3.1) @ Buffalo Sabres 52% (3.33)

⚾ MLB

Miami Marlins 36% (2.99) @ Tampa Bay Rays 64% (4.41)

Philadelphia Phillies 46% (3.05) @ Pittsburgh Pirates 54% (3.48)

Baltimore Orioles 50% (4.95) @ Washington Nationals 50% (4.93)

Boston Red Sox 40% (3.58) @ Atlanta Braves 60% (4.68)

Toronto Blue Jays 54% (4.16) @ Detroit Tigers 46% (3.73)

Cincinnati Reds 38% (3.23) @ Cleveland Guardians 62% (4.53)

New York Yankees 55% (3.9) @ New York Mets 45% (3.36)

Texas Rangers 55% (4.28) @ Houston Astros 45% (3.76)

Milwaukee Brewers 58% (4.15) @ Minnesota Twins 42% (3.35)

Chicago Cubs 61% (5.09) @ Chicago White Sox 39% (3.89)

Kansas City Royals 53% (4.31) @ St. Louis Cardinals 47% (3.95)

Arizona Diamondbacks 62% (5.69) @ Colorado Rockies 38% (4.49)

San Francisco Giants 43% (3.7) @ Oakland Athletics 57% (4.47)

Los Angeles Dodgers 59% (5.36) @ Los Angeles Angels 41% (4.37)

San Diego Padres 43% (3.99) @ Seattle Mariners 57% (4.76)

Predictions via my proprietary model. Not financial advice.

[Adam @WalrusQuant]

@netflixsports Poop on a stick !!

[Adam @WalrusQuant]

What a joke . 6 seconds? I should have went to bed .

[Adam @WalrusQuant]

Oliver Tree - F**k The Whole World (Subtronics Remix) [Official Audio] https://t.co/8bYhDJsV6L via @YouTube

[Adam @WalrusQuant]

https://t.co/NEaqm96yxE

[Adam @WalrusQuant]

i have been working on this git explorer macOS desktop app that turns your projects folder into a navigable tree of git repos, with status, diffs, and one-click staging. nice way to manage git locally.

[Adam @WalrusQuant]

https://t.co/LqIuCQ3pdc

[Adam @WalrusQuant]

Made some updates to my open-source PGA Tour API packages (R + Python).
•Added multi-stat / multi-year batching to pga_stats()
•Filled out API coverage with 3 new endpoints (content, odds config, speed rounds)
•Hardened the transport: retries on 5xx, timeouts, env-var key override, verbose debug mode
•Fixed silent parser bugs (dropped seasons, phantom fields, miscoded types)
•Built an 89-test offline suite

[Adam @WalrusQuant]

https://t.co/j99EOCnY6s