Measuring Drama in Goose-like Games

This study proposes objective, simulation-based criteria for measuring "drama" in chance-driven race games, extending a concept originally defined for skill-based abstract games such as Chess. Drama is characterized as the ability of a trailing player to recover and win, which prevents early advantages from making a match uninteresting. The authors argue this quality matters even in games where randomness replaces skill, and they test their approach on the Game of Goose family, comparing the classic game against two pedagogical historical variants and an abstract variant with no special rules. The methodology relies on over 10 million computer-simulated matches written in R, with each variant tested across three to eight players.

Three criteria are defined and measured. The WIN criterion captures how balanced winning chances are across seating positions; the Goose game performed best, and all historical variants showed small deviations, though the second player held a consistent advantage that was statistically significant, with the first player generally disadvantaged. The LEAD criterion measures how late the eventual winner takes the lead, with roughly 74 to 86 percent of matches decided within five turns of the end across variants, indicating strong suspense. The ANP criterion counts how many players lead at some point, finding that in typical matches with four or more players about half of all players take the lead, which the authors regard as a balanced, engaging outcome.

The study also reports game length, which follows a log-normal distribution with a heavy right tail, and the distribution's parameters behave predictably as player count changes. The authors present the methodology as general enough to apply to other stochastic race games, while acknowledging that the link to log-normal behavior remains empirical and unexplained analytically.