Accuracy from the slot: Evaluating draft selections in the National Hockey League

The National Hockey League (NHL) entry draft is a process wherein teams make sequential selections from a pool of eligible players. Given the young age of prospects, drafting requires long-term forecasting of future performance under a high level of uncertainty. This study assessed the selection accuracy across all seven rounds of the draft, as well as between lottery and non-lottery picks within the first round. NHL performance data were collected for all forwards (N = 956) and defensemen (N = 558) drafted between 2007 and 2014. In both groups, Kruskal-Wallis H tests conducted between draft rounds revealed a significant, relatively strong, overall effect of draft order on future performance. However, Mann-Whitney U post-hoc tests showed projecting future performance of forwards was only accurate in the first two rounds, while for defensemen, selection was only accurate in the first round. Moreover, forwards selected with lottery picks in the first round outperformed their non-lottery peers offensively but not defensively. As for defensemen, those selected with lottery picks did not differ from their non-lottery peers in offensive or defensive performance. Our findings highlight substantial inaccuracies in the NHL draft, particularly past the first two rounds of selection. We offer multiple possible explanations driving such inaccuracies that could form the basis for further work in this area.