2022 FALL EDITION

The projects included in this issue include research from University of Pennsylvania Ph.D. student Ryan Brill in collaboration with Penn undergraduates and high school students from Moneyball Academy, as well as projects completed by rising high school students who completed our summer high school programs Moneyball Academy and Moneyball Academy Training Camp in 2022. We hope you enjoy reading them!

Table of Contents

Introducing Grid WAR: Rethinking WAR for Starting Pitchers
Ryan S. Brill, Justin Lipitz, Emma Segerman, Ezra Troy, Abraham J. Wyner 

Predicting the Quarterback-MVP in the NFL
Ryan S. Brill, Ryan Weisman 

Zero Running Backs or Zero Points: An Analysis on the Optimal Fantasy Football Draft Strategy
By Andrew Heller, James Sohigian, Henry Bartz, Ethan Kawahara, Felix Wang

Predicting Career Wins Above Replacement from Rookie Stats in Baseball: Addressing the Esteban German Dilemma
By Jack Blumenstein, Josh Braverman, Lekh Murthy, Wilson Wendt

Stars Matter: an Analysis of College Football Recruiting, Development, and Draft Success
By Naya Kessman, Ellery Axel, Cotton Snoddy, Will Hoey

MLB Umpire Pitch Evaluation Declines with Age; Implications for World Series Selection
By Brooks Fischer, Ethan Schulman, Hayden Lee, Ethan Jin

The 2021-2022 Liaoning Flying Leopards: The Best Championship Team in Franchise History
By Yecheng Yue, Jason Shi, Sirius Xiao

Is Everton’s One-Season Drop Off Due to Random or Structural Factors?
By Sammit Bal, Sanjay Bharadwaj, Carlo Sebastiani, Kevin Yu

Whether Weather, Wind Speed and Temperature, Impacts Offensive Success in the NFL
By Hudson Parks, Cole Ceklosky, Jake Roseman, Jonah Awad, Jimmy Yan

The Importance of Break Point Percentage
By Jenny Yu, Joshua Kwong, Amar Kumar, Jeffen Zhang

Running to Runs: the Importance of Baserunning
By Jaden Patel, Irene Bian, Matthew Ladden, Jackson Krump


Introducing Grid WAR: Rethinking WAR for Starting Pitchers 
Ryan S. Brill, Justin Lipitz, Emma Segerman, Ezra Troy, Abraham J. Wyner 

Traditional methods of computing WAR (wins above replacement) for pitchers are based on an invalid mathematical foundation. Consequently, these metrics, which produce reasonable values for many pitchers, can be substantially inaccurate for some. Specifically, Fangraphs and Baseball Reference compute a pitcher’s WAR as a function of his performance averaged over the entire season. This is wrong because not all runs allowed have the same impact in determining the outcome of a game: for instance, the difference in impact between allowing one run in a game instead of zero runs is much greater than the difference in impact between allowing six runs in a game instead of five. Hence we propose a new way to compute WAR for starting pitchers: Grid WAR (gWAR). The idea is to compute a starter’s gWAR for each of his individual games and then define a starter’s seasonal gWAR as the sum of the gWAR of each of his games. We find that gWAR highly values games in which a pitcher allows few runs (0 or 1).


Predicting the Quarterback-MVP in the NFL 
Ryan S. Brill, Ryan Weisman 

The NFL MVP award is chosen each year by a panel of 50 sportswriters who are selected by the Associated Press. As the MVP is chosen by humans who do not necessarily base their decision on statistics, but on watching all the games and talking to coaches and players, it is natural to wonder whether there is a mathematical rule that can describe the MVP selection process. Hence in this paper, we create a logistic regression model to predict the NFL Quarterback-MVP. Our algorithm predicts the correct Quarterback MVP in each year since 2003 except 2009 and 2015. In 2009, our model chooses Brees over Manning, and in 2015, our model chooses Palmer over Newton. Perusing articles from 2009 and 2015 reveals that many people in the media agreed with our model that Brees and Palmer were snubbed.


Zero Running Backs or Zero Points: An Analysis on the Optimal Fantasy Football Draft Strategy
By Andrew Heller, James Sohigian, Henry Bartz, Ethan Kawahara, Felix Wang

The Zero-Running Back (RB) draft strategy is to ignore RB until around the 6th round, and to draft the leftover RBs (usually consisting of high upside players or backups that might play or can be used for trade value). Instead focus high draft rounds on WRs, TEs, and QBs in the early rounds of the draft. We simulated fantasy football drafts over the 2010 to 2021 seasons to determine the efficacy of the zero-running back draft strategy.  From our data, the Zero – RB doesn’t work optimally and it is suggested that the Fantasy Football draft process is based primarily on luck; random year to year variation in player productivity.

Predicting Career Wins Above Replacement from Rookie Stats in Baseball: Addressing the Esteban German Dilemma
By Jack Blumenstein, Josh Braverman, Lekh Murthy, Wilson Wendt

Esteban German had a great rookie season, which would seem to predict future success, however it often does not. Data was obtained from FanGraphs, and a model was created to predict rest of career WAR using rookie metrics and a multivariate linear regression model. Important features in our model included: power, strikeouts, stolen bases, quality of contact, defensive ability, age. We found that career WAR can be projected well from rookie metrics using our model. Some players that have a strong start to their careers usually regress because of unsustainable measures in their underlying statistics.

Stars Matter: an Analysis of College Football Recruiting, Development, and Draft Success
By Naya Kessman, Ellery Axel, Cotton Snoddy, Will Hoey
Successful NCAA Division I Football (FBS) programs can recruit better high school players and they also produce players that are highly drafted into the NFL. This work explores which FBS programs are better at developing their players for the NFL draft, factoring in player recruiting quality. Recruiting class rating data from 247Sports.com was combine with Draft data from CollegeFootballData.com. Ohio State, Alabama and Penn State were found to be the best programs at developing their players for the NFL draft.


MLB Umpire Pitch Evaluation Declines with Age; Implications for World Series Selection
By Brooks Fischer, Ethan Schulman, Hayden Lee, Ethan Jin

In contrast to other sports, Major League Baseball (MLB) has maintained a strong human element in officiating, especially in home plate umpire pitch evaluation. With improved pitch tracking technology, umpire consistency and accuracy is now more evident. This project examines how umpire pitch calling changes with age. A strong negative relationship is observed between umpire age and umpire pitch call accuracy. Umpire pitch consistency is less age dependent. Reasons that umpire pitch call accuracy could worsen with age include: eyesight, reaction times, and selection bias; poor older umpires are harder to replace than the poor younger ones. MLB selects more experience and senior umpires for the WS, those with relatively worse accuracy in pitch calling.

The 2021-2022 Liaoning Flying Leopards: The Best Championship Team in Franchise History
By Yecheng Yue, Jason Shi, Sirius Xiao

The Liaoning Flying Leopards won the Chinese Basketball Association championships in the 2017 and 2021 seasons. In both 2018 and 2019, the Leopards kept their 2017 lineup but lost to the Southern Tigers. In the past, people thought the 2017 team was better because the team was unstoppable on offense. Comparing the 2017-18 Flying Leopards to 2021-22 – we are analyzing whether the basketball league CBA’s 2022-21 season Flying Leopards is the best championship team in their franchise history using three distinct metrics: turnover ratio, TS%, efficiency differential. The 2020-21 Leopards had a lower offensive efficiency, but their defensive efficiency got better, thereby, having a higher efficiency differential. The 2022 Leopards have a more flexible squad and they are the best team in franchise history.

Is Everton’s One-Season Drop Off Due to Random or Structural Factors?
By Sammit Bal, Sanjay Bharadwaj, Carlo Sebastiani, Kevin Yu
As a low scoring sport, soccer has relatively more random outcomes. This project examines the English Premier League team Everton’s results from the 2020-21 and 2021-22 seasons to determine the cause in the observed drop in performance. We find that Everton’s drop off was due mainly to structural factors. Factors which include the departures of the manager and three star players that were worth a combined $50 million.

Whether Weather, Wind Speed and Temperature, Impacts Offensive Success in the NFL
By Hudson Parks, Cole Ceklosky, Jake Roseman, Jonah Awad, Jimmy Yan

In the NFL, when playing in poor weather (cold and/or windy), teams often choose relatively run-heavy gameplans. We examine how different NFL weather conditions impacted the effectiveness of both run and pass plays. Using data from nflfastR, the Expected Points Added (EPA) per play from all plays between 2014-2021, was assessed over different temperatures and wind speeds.

As the wind speed increases, passing plays become less effective while rushing play remain roughly constant. As temp increases, passing becomes more effective, while rushing gets worse.


The Importance of Break Point Percentage
By Jenny Yu, Joshua Kwong, Amar Kumar, Jeffen Zhang

In tennis, players alternate who serves in each game. Often players can hold their serving game while breaking your opponent’s serve is far more difficult. Our project examines how important break points are and how break point percentages affect the outcome of a match. Winners were found to have a higher break point percentage by 14.0%, and there is a slight variation across surfaces.


Running to Runs: the Importance of Baserunning
By Jaden Patel, Irene Bian, Matthew Ladden, Jackson Krump

Over the past 20 years, baseball has reduced the rate of stolen base attempts and devalued baserunning. Our project examines team level baserunning statistics to discern if there is a correlation between a team’s baserunning and its winning percentage. A factor, BsR, was created that includes all plays on the basepaths. BsR was found to correlate with team winning percentage; better baserunning results in a slight advantage.