<oai_dc:dc xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd"><dc:title>Market Efficiency in Sports Betting Markets: A Comparative Analysis of the NBA, MLB, and NFL Before and After Legalization</dc:title><dc:creator>Makwana, Om </dc:creator><dc:subject>Moneylines</dc:subject><dc:subject>Sports Betting</dc:subject><dc:subject>PASPA</dc:subject><dc:subject>Murphy v. NCAA</dc:subject><dc:subject>NFL</dc:subject><dc:subject>NBA</dc:subject><dc:subject>MLB</dc:subject><dc:coverage>Economics</dc:coverage><dc:relation>B A</dc:relation><dc:description>This thesis explores the informational efficiency of North American sports betting markets to determine if bookmaker-implied probabilities derived from moneyline odds serve as well-calibrated
predictors of game outcomes, and whether the 2018 legalization of sports betting in the United
States—following the Supreme Court’s decision in Murphy v. NCAA—improved market efficiency.
Using a comprehensive dataset of 63,731 regular-season games spanning three major professional
leagues—the National Football League (NFL, 2007–2025), the National Basketball Association
(NBA, 2008–2025), and Major League Baseball (MLB, 2010–2025)—this analysis employs Ordinary Least Squares (OLS) Linear Probability Models incorporating post-legalization and COVID19 pandemic dummy variables alongside their interactions with market-implied probabilities. Results indicate that betting markets across all three sports are strongly predictive (p &lt; 0.001), and
slope coefficients are statistically indistinguishable from unity (NFL: βˆ
1 = 1.044, p = 0.165 for
H0 : β1 = 1; NBA: 1.005, p = 0.738; MLB: 1.014, p = 0.608), indicating that moneyline-derived
probabilities are well-calibrated predictors of game outcomes. The NBA exhibits the highest predictive accuracy (Brier Skill Score = 0.193), followed by the NFL (0.165) and MLB (0.039).
Critically, the legalization interaction term is not statistically significant for any individual sport
(p &gt; 0.05), suggesting that the structural shift in the legal environment has not fundamentally
altered the calibrative properties of these markets. COVID-19 pandemic effects are similarly insignificant. In the pooled cross-sport specification, neither NBA nor MLB markets are statistically
distinct from the NFL baseline, suggesting broadly similar calibration properties across leagues.
These findings contribute to the literature on prediction market efficiency and have implications
for regulatory policy, behavioral economics, and sports analytics</dc:description><dc:contributor>Jorge Enrique Perilla Garcia, Thesis Supervisor</dc:contributor><dc:contributor>Sung Jae Jun, Thesis Honors Advisor</dc:contributor><dc:rights>open_access</dc:rights><dc:date>2026-04-08T03:50:56Z</dc:date><dc:identifier>https://honors.libraries.psu.edu/catalog/10296odm5057</dc:identifier></oai_dc:dc>