Competitive programming participation rates: an examination of trends in U.S. ICPC regional contests

Abstract

A wide range of benefits have been posited from participation in competitive programming contests. However, an analysis of participation in north American regional contests in the International Collegiate Programming Contest (ICPC) shows that participation in these contests is sharply declining, coinciding with the COVID-19 pandemic. Moreover, prior to the pandemic, while the number of teams participating in regional contests was increasing, the number of institutions sending teams to these contests was declining. We find several statistically significant correlations that may underscore structural reasons for this trend. Consistent participation in contests and the number of teams per institution sent to a contest both are correlated with likely participation in future contests. On the other end of the spectrum, institutions sending a team to a contest for the first time in 3 years were much less likely to return in the next year. For this category of teams, if a team is unable to solve any problems in the contest, the institution is significantly less likely to send a team in the next year. Many of these contests have very challenging problem sets, and consequently, have many teams that fail to solve any problems. This result suggests that structuring the problem sets to increase the likelihood that most teams successfully complete problems would broaden participation in these contests.