The existence of a home-court advantage is one of the most durable empirical patterns in all of sports. Yet, the mechanisms explaining its strength and persistence remain a mystery in large part because of well-known challenges with statistical identification. We use attendance restrictions in place during the 2020–2021 National Basketball Association regular season as an instrument in order to identify the effect of fans and crowd size on home-court advantage. We show that home teams win by [Formula: see text] points, on average, when fans are present at games compared with [Formula: see text] points when no fans are present. This equates to winning approximately 2.2 additional home games over the course of a regular season. In fixed effects instrumental variables regression models, we estimate that the marginal effect of an additional one thousand fans on home-court advantage is [Formula: see text] points. We conclude that the mere presence of home fans, on its own, explains a larger share of home-court advantage than previously thought.