Fall 2024 Archives

Expected Value Curves Don’t Tell the Full Story: Exploring NFL Draft Position Trade Value Curves

Wharton Sports Analytics Journal Open Access · ISSN 3070-4065 FALL 2024 Expected Value Curves Don't Tell the Full Story: Exploring NFL Draft Position Trade Value Curves AUTHOR Blake Zilberman, University of Pennsylvania Ryan S. Brill, University of Pennsylvania ABSTRACT Football analysts traditionally value a future draft pick position by its…Read More

College Basketball: An In-depth Study of the “Foul Up 3” Dilemma

Within college basketball, the decision to foul or not with a 3-point lead (the “foul up 3” dilemma) significantly impacts teams. A single fouling decision can directly determine the outcome of a game or even an entire season. Within this research project, I take a novel look at the “foul up 3” dilemma. As opposed to using time as a blocking factor, as previous studies have, my research focuses on generating a coachable strategy that outlines the superior fouling decision for each moment in time.…Read More

To Go for Two or Not to Go for Two? A Statistical Analysis of the Biggest Prisoner’s Dilemma in the National Football League

In the National Football League, after a team scores a touchdown, they can kick an extra point or attempt a two-point conversion. Historically, teams have opted for the former, as it had a much higher likelihood of success. In 2015, however, the NFL instituted a rule to move the extra point distance back 13 yards, making it a more difficult kick. This paper will analyze the effect of this rule on the extra point rate, statistically analyzing both strategies and offering recommendations to both NFL teams and the league itself.…Read More

Beyond the Boundary: Revolutionizing the IPL MVP Index

Since its inception in 2008, the Indian Premier League (IPL) has attracted many of the world’s most skilled cricket players, offering a highly competitive arena for them to showcase their talents. Each season, the IPL awards the Most Valuable Player (MVP) title to the player who achieves the highest rating on the league’s MVP metric. Ideally, this award recognizes the top performer of the season, with high rankings indicating outstanding achievement among elite players. However, the calculation used by the IPL to assess player performance lack consistency, are limited in scope, and rely on arbitrary criteria. This paper employs a multivariate regression model to develop a more robust formula, assigning mathematically optimized weights to devised metrics that better capture player contributions. With an R² value of 0.80—compared to the existing system’s 0.66—this new formula provides a more accurate and comprehensive evaluation of player performance.…Read More

Examining regional and familiarity bias of referees in USA Fencing Division I bouts

This study examines whether USA Fencing referees exhibit favoritism toward competitors from their designated regions, as frequent officiating within local areas could establish familiarity and potentially lead to unconscious bias. This issue is particularly relevant in foil and saber, where referees make subjective judgments on “right-of-way” during simultaneous hits, and in epee, where penalty decisions hold significant weight. Unlike other sports with subjective scoring that utilize referee panels, fencing relies on a single referee, granting them considerable control over bout outcomes. Despite the opportunity for prejudice, there are no studies of referee bias in fencing; this study aims to be the first to do so. Utilizing a substantial dataset of 35,111 Division I pool bouts from 2012 to 2019, I applied linear and logistic regression models to analyze the effect of regionality on score differentials and bout outcomes.…Read More

Clutch Matches Are In The Middle: Optimizing OPR with Weighted Least Squares

In this paper, we present an improvement to Offensive Power Rating (OPR), a popular linear regression model for assessing team performance at a given event. One key assumption of linear regression is the independence of the errors, but in the FIRST® Robotics Competition (FRC) context, this assumption is not exactly true. Using data from all district events between 2009 and 2024, we model the unweighted errors as a function of tournament progression and generate weightings to improve the regression fit through Weighted Least Squares (WLS). The best weightings show that the most representative matches for a team’s overall performance are midway through the tournament. That is, the real clutch matches are in the middle.…Read More

Exploring various NBA draft value curves

NBA teams often trade draft picks. We are interested in the relative value of draft positions, which determines whether a team should accept or reject a trade, or which draft picks it should offer in a trade to make it fair. In this work, we explore various NBA draft value curves. We introduce a novel method, mapping player performance measures (e.g., WAR, RAPTOR, BPM) to salary using Gamma regression in order to constrain draft value curves to be positive. We find that, depending on the measure of performance value and the method of aggregation (e.g., mean or median), draft value curves are wildly different.…Read More

Calculating Win Probabilities of Any Matchup of Soccer Teams: A Whole-History Rating Approach for the Wharton High School Data Science Competition

This study presents a novel approach to evaluating soccer team strength and calculating win probabilities given any two matchups of soccer teams in the fictional North American Soccer League (NSL), whose data is provided for the Wharton High School Data Science Competition. Traditional win-loss metrics have many limitations, like failing to account for the strength of schedule, margin of victory, and home-field advantage. Thus, we developed a Whole-History Rating (WHR) model adapted from Rémi Coulom that incorporates expected goals (xG) and accounts for home-field advantage, overcoming the sequencing issues inherent in single-pass Elo systems. Using a dataset of 476 NSL games, we implemented a multi-pass system for iterative rating adjustments and employed Bayesian inference to develop probability distributions for team ratings.…Read More