Evan

@pickleo7 Is defensive situation accounted in this in any way? Is an uncontested fast break layup the same xP as a contested put back attempt from the same exact spot?

Introducing my model for Expected Points (xP) in College Basketball: As an avid soccer fan (#COYG) and data nerd, expected goals (xG) has fascinated me in the way it can tell the underlying story in such a binary, low scoring sport. I think there’s value for this analysis in CBB. The concept is pretty simple: how many points do we expect a given shot to be worth from a specific location on the floor, accounting for the distance and angle. Think of it like this: if a random player takes a shot from 2 feet behind the 3PT line straight in line with the basket, we might expect that shot to be worth 1.1 points (which would translate to a 36.67% rate from there). xP is a way of looking at shot value at an individual level to tune out some of the noise of results and see what underlying trends we can find. Scoring more points than your xP over the course of the season would be a reliable indicator of above average shot makers, while scoring less could mean that your primary scorers may be in good positions to score but aren’t good shooters. The inverse could be the case for your defense. This is not a one stop shop metric, but is a different way of evaluating the most important part of basketball: scoring. On an individual level: Koby Brea over the course of our 2 year sample consistently outperforms his xP by .444 points a shot (2nd only to Mason Gillis at .450 points a shot). So, this matches up with our knowledge that Brea was an elite shooter in college and was much more effective than the average player would be in the same positions he got into. On a game by game level: xP can make you not feel crazy about your rival’s worst 3PT shooter randomly making 7 3s to beat you on your home floor. It also may show that a wild game result (like 2025 champs Florida shooting 4-27 from 3, which led to me tweeting I was “out on Florida”) really can just be shooting variance. I plan to delve deeper into using the data to create other xP based metrics that can be used for evaluation. I created my model with the wonderful CBB data API from @CFB_Data, and trained the LGBM on close to 1.7 million shots from the last 2 seasons of CBB shot location data. Bill does phenomenal work and is a must follow for anyone interested in learning more about the sports they watch week in and out. I plan to run game by game analysis and keep some running leaderboards that show offensive and defensive efficiency for teams and some more player based ratings as well. Here is a sample game recap on Expected Points from Alabama Purdue back in November, as well as a “Deserve to win O Meter” Monte Carlo simulation of xP, an idea borrowed from the great work of Arsenal data commentator @scottjwillis Enjoy and feel free to ask any questions!

@ADavidHaleJoint The last 24 years would say otherwise.

I think Miami gets in because the universe cannot possibly hate the Canes enough to have FSU’s awful conference record keep Miami out of the champ game, then Manny Diaz win the league, then Canes miss the playoff too.

@AriWasserman Which one?

Malachi Toney is so freaking good. Did you guys see that?

@ADavidHaleJoint @ESPNStatsInfo 2017 and 2007

@2stePHenPaul @SickosCommittee @SickosCommittee ZOOM OUT

@2stePHenPaul @SickosCommittee They're literally above the top line. 204 2nd half pts.

Is your team a first half or second half team? Below the line = more points scored in the first. Above the line = more points scored in the second.

@RedditCFB *again

In this week’s edition of referees completely botching big games, Miami has had a safety stolen due to incompetence.

@Buccigross @FlaPanthers @MapleLeafs #bucciovertimechallenge Lundell / Rielly

#bucciovertimechallenge #18 of this year’s #StanleyCupPlayoffs…@FlaPanthers / @MapleLeafs

@suhcyut Are we doing amongst now?

What’s a random/underrated Happy Endings quote you say often? For me, whenever I get excited about something dumb or totally normal, I say, “Act like ya been there!” (Quote, Penny, Boys to Menorah.)

@ESPN_BillC Thanks @JDunFunny

Just got my first "Hey, uh, [Tyrone Biggums scratch], you got any of those Week 0 SP+ projections?" message of 2022. The season is officially upon us. But that reminds me: Final SP+ projections are going up on Sunday. And then Week 0 picks'll go up Wednesdayish.

@suhcyut *26th

“I’m so forgetful. I’m such an Aries!” “Aren’t you on the cusp of Aries because you were born on the 22nd?” 💞 Happy 41st birthday, Penny Hartz! 💞

Let’s show @univmiami what the power of the people looks like… I’ll donate $250 for every RT up to $3,000,000 to rev this engine up…. #TrustTheProcess #UM

Spring Smackdown is over. After one month and 7,059 total votes cast, your winner is: Cocktails & Dreams

@rogbennett "Football is a simple game, 22 Players kick a ball for 90 minutes and at the end, the North Macedonians win"

@SpringSmackdown YO GUYS!!!

The FINAL! I don't believe it! Seed 10: Season 2 Episode 16 COCKTAILS AND DREAMS vs Seed 17: Season 2 Episode 15: THE BUTTERFLY EFFECT EFFECT

ROOT FOR CHAOS

@ESPN_BillC Do you keep any type of results spreadsheet/database that tracks SP+ success by team/conference/line/margin between line and predicted score etc.? If so, any chance that will ever be publicly available?

Hey, so I think I'm gonna do a Twitter Live Q&A thing tomorrow afternoon, say, 2pm ET? I miss talking to you guys! (Well, *some* of you.) If you're interested, shoot me a question on this thread, and I'll pick some out to answer tomorrow.

@RandallOtisTV youtube.com/watch?v=Ga9UPY…

What’s the best fake song? Like a song that exists in the world of a show/movie but not our world