I am glad to announce that I will be committing to
@Princeton
as part of the Class of 2028! Thanks to everyone who has gotten me to this point and I cannot wait for an exciting four years in the beautiful city of Princeton, New Jersey!
@ApplyPrinceton
#college
#commitment
I made a ShinyApp relating to how "true" of a fan you are based on distance to teams you support!
ShinyApp link:
GitHub link:
Please repost for reach, thank you!
(Below is based on where I grew up - Lawrenceville, NJ)
AYOE: Air Yards Over Expected
Filtered to QBs With 100+ Pass Attempts In 2023 From Week 1 to 11; Trained Data from 2014 - 2020; Tested Data from 2021 - 2023
Essential Part Of This Is That The Receiver Skill Is Accounted For
#sportsanalytics
#footballanalytics
#nfl
Details ↓
NBA FGPOE (Field Goal Points Per Game Over Expected)
Trained Data from 2021 - 2023 to Test 2024 So Far. XGBoost With Situational Factors. Tells A Lot About Scoring Efficiency.
#basketball
#sportsanalytics
#nba
#datascience
#statistics
Please repost
Details ↓
After November 3, we see Luka Doncic leading the pack for the NBA MVP award as he has led the Mavericks to a 4-2 record with great base and advanced statistics! This model is based on training data from 1980 - 2023 with the correlation between statistics and the award reception.
Was happy to present at the OSU Sports Analytics Conference earlier today by
@OhioState
! Thanks to
@RyanRuddy6
,
@JohnDDraper
, and other members of the team once again!
Thought it would be fun to make a reverse MVP ranking.
Players considered played at least (65/82) of the max games played by a team player with at least 20 minutes in them
#nba
#basketball
#sportsanalytics
#datasci
#stats
Visualization of 2023 QBs with >= 250 Pass Attempts Between Their Efficiency (EPA/Play) and Aggressiveness (aDOT Over Expected)
aDOT over Expected itself incorporates receiving epa for targeted receiver so the negative relationship makes sense
#nfl
#football
#sportsanalytics
Always great to see our former
@sabr
interns return to present their own research at
#SABRanalytics
. Congrats to
@reece_ca1vin
on a terrific presentation today on optimizing lineups with ball tracking data!
If the Knicks can afford him, Paul Reed suddenly becomes an excellent young backup center option.
He fits into what the Knicks ostensibly want from their centers on both ends, >95th %ile in OReb rate each of his 4 seasons, and an 99th %ile rim deterrer.
#NewYorkForever
It's funny to see how the people who hate on analytics in sports don't know anything about how it works. "It doesn't consider context" - yes it does. Analytics people understand that it isn't perfect - it also has become the one to target as a result of hindsight bias
#sports
Creating a Deviation From Expected Pitch (DFEP) Metric For 2024 MLB Qualified Pitchers Based On Situational Factors To Identify Unpredictable and Predictable Pitchers
Write-Up On The Links Below ⬇️
#mlb
#baseball
#sportsanalytics
#sabermetrics
Looking at a new "expected points" metric in the NHL
This includes also looking at individual progressions through the season.
Write up will be in the LinkedIn link below which I will post as soon as possible
#hockey
#nhl
#sports
#analytics
#datascience
#statistics
There is a misconception about statistics in sports.
You MUST pair the statistics with context when evaluating an entity in sports. You also do need to add the factor of "eye test" even though it has a subjective nature.
#statistics
#sportsanalytics
#sports
#datascience
Predicted vs Actual NBA Positions
Trained Data From 2015 to 2021, Tested from 2022 to 2024. Used Primary Position Only. Few Interesting but Not Surprising Results Outlined in Details
#basketball
#nba
#basketballanalytics
#sportsanalytics
Please repost and details below ↓
Was an unbelievable experience at the CMU Sports Analytics Conference headed by Carnegie Mellon University Statistics & Data Science,
@Stat_Ron
,
@stat_sam
, others. Was able to talk to the most successful gambler of all time, Bill Benter.
#sportsanalytics
#cmsac
Panthers plan to hire SumerSports VP Eric Eager in an analytics role, per league source.
Eager worked for Pro Football Focus prior to SumerSports, a startup in 2022 (that also employs Thomas Dimitroff) aimed at helping NFL teams in player acquisition and roster management.
I have somewhat of an idea of evaluating offensive line analytically without using tracking data.
In short, it involves using a non-tracking version of RYOE and looking at trades and signings to see how RYOE changes as a result.
Detailed version⬇️
#nfl
#football
#analytics
2023 Formula One TSOE (Turn Speed Over Expected) in KM/H with Turn Situational Factors
Just want to preface possible lack of accuracy (you cannot directly associate turns with telemetry data)
Thanks to
@BrotoFFCasanova
and
@pbulsink
for f1dataR
#f1
#formulaone
#sportsanalytics
OUR SCHEDULE IS AVAILABLE!
We are BEYOND EXCITED to share the program for
#MSAM2023
and once you decide you LOVE the schedule of talks/posters, be sure to register at .
For anyone interested in what goes into the MBAPPE model, I'm going to echo the about section here!
MBAPPE (Multi-League Bayesian Adjusted Penalized Plus-Minus Estimate) is a Bayesian RAPM with SPM prior. What is that? Thread Below📝 (1/10)
I made a plot to show the general relationship between pitcher predictability and proficiency!
It does represent the ideal of less predictability translating to more proficiency but its not a strong relationship.
#baseball
#sabermetrics
#sportsanalytics
#mlb
Creating a Deviation From Expected Pitch (DFEP) Metric For 2024 MLB Qualified Pitchers Based On Situational Factors To Identify Unpredictable and Predictable Pitchers
Write-Up On The Links Below ⬇️
#mlb
#baseball
#sportsanalytics
#sabermetrics
For reference, this model prioritizes wins a lot becuase I used PCA (principal component analysis) to reduce the dimensionality of most of the metrics into one value except things like wins
#football
#nfl
#americanfootball
#sportsanalytics
Our predictive analyst
@avsportsanalyst
, has updated his
#NFL
MVP award rankings…
As it stands,
#Eagles
quarterback Jalen Hurts took over 1st place followed by
#Lions
Jared Goff, and
#Chiefs
Patrick Mahomes… this model is based off “basic” advanced statistics from 2007!
To be fair, footy statistics suck in terms of fully capturing impact. Clearly did not capture foden’s impact on city (or maybe my boy Cole Palmer was robbed)
#premierleague
#soccer
#football
#analytics
@NathanG1904
@ftblnl
@4lex_mcfc
It’s essentially based on values on a pitch which is generated based on past data. So essentially, if you move from one section of pitch to another through a pass or carry you’re generating a higher probability of the sequence ending in a goal
Episode 15 is out now! Joined by a rising star in the analytics community
@avsportsanalyst
check out the episode for a great conversation discussing the new free agency signings, analytics in new sports, and a fun Baker Mayfield debate!
Creating GOPV for NHL Goalies: Goalie OverPerformance Value
Basically represents a linear bayesian adjustment based on strength of offensive team faced of the ratio between goals against and xG against
Write-Up On LinkedIn + Medium Below⬇️
#hockey
#nhl
#sportsanalytics
#data
Similarity metric describing how similar players on the "most used" lineup per NBA team are to each other as a whole team
The GitHub will have a plots folder where you can analyze the visualization for all NBA teams
#nba
#basketball
#sports
#sportsanalytics
For reference, the teams and players I support are:
NFL: Eagles
NBA: 76ers and KD
MLB: Phillies
NHL: Flyers
Footy: Chelsea/Barca and Messi
Cricket: India/CSK
Tennis: Djokovic
F1: Bottas
However, sports analytics is all about objectivity so no bias
#sportsanalytics
#sports
His rise to the league is unbelievable, leading in so many advanced statistics. Think I will need to update my MVP model because he has to be up there probably top 3
Hali is averaging 25-4-12
Shooting 52/46/92
He ranks 1st in APG, AST%
4th in PER, 9th in TS%, eFG%
4th in WS, 1st in oWS, 3rd in WS/48
3rd in BPM, 1st in oBPM, 5th in VORP
4th in EPM, 1st in oEPM
+21.6 on/off diff.
In terms of statistical ridiculousness, he’s guard Jokic
@mplsoccer_dev
At least for in game penalties, its supposed to be an advantage because they got fouled in the penalty box. But if it changes for penalty shootouts could be interesting
This 2024 Celtics team was just amazing all-around. As a Sixers fan I am absolutely gutted but from what I have been hearing analytics helped the team immensely so I am all in
Two different NFL QB-MVP models. One uses pure statistical inputs with CJ predicted as MVP, while the other condenses most of the statistics using principal component analysis (prioritizing wins) and has Hurts as MVP.
#football
#nfl
#sportsanalytics
#footballanalytics
#datasci
As the 2024 version of the Indian Premier League starts soon, just wanted to repost my results for my Win Probability Added model to measure impact in last year's IPL!
(Credit to
@TAlbTree
for the headshot + logo data)
#cricket
#ipl
#analytics
#sports
I analyzed midfielder performance in the 2022-23 season for the Big 5 Leagues comparative to all other midfielders in five different categories using PCA. I showed 5 midfielders from each league for demonstration, thread is below, Prem first
#football
#sportsanalytics
#soccer
@Hoops_Nerd
And yeah, eye test shouldn’t be used when you’re making sports analytics projects it’s more about using it when you evaluate players if you do it for fun
Introducing SMILE+
This was a silly project which I had a blast doing, and helped me learn a lot about computer vision
Check out my article covering my methodology and code on Medium