Leveling up fun: Learning progress, achievement, and expectations influence enjoyment in video games

What factors influence how much fun people have when engaging in inherently enjoyable tasks? The theory of learning progress predicts that people will have the most fun in environments of intermediate difficulty because these environments offer the most progress in learning about the world. Past studies have frequently focused on simple experimental paradigms in which learning was still instrumental for later tasks. Here, we put the theory of fun as learning progress to a test in three large and realistic video game data sets: a puzzle game (with 7,994 levels and 376,341 votes), a racing game (138,662 levels and 614,770 votes), and a platformer game (115,032 levels and 795,313 votes). As predicted, people preferred levels of intermediate difficulty in all games. Yet, additional factors influencing people's enjoyment also emerged: players preferred levels that matched closely with their prior expectations of difficulty and were also motivated by achievement. We further confirmed these factors in two precisely controlled experiments. Taken together, these results advance our understanding of the dynamics of fun in realistic environments and emphasize the importance of using both realistic, game-like environments and highly controlled experiments to refine theories of human learning and decision-making.