A Computational Approach to Test Spatial Competition

The Spring Meeting of Turkish Economic Association, İstanbul, Türkiye, 29 - 31 Mayıs 2025, ss.137, (Özet Bildiri)

Location games typically lack a pure strategy Nash equilibrium. Such cases are commonly addressed by

turning to mixed strategies as a potential solution. In this study, we develop a novel method to empirically

test the fit of mixed strategy Nash equilibria in the context of location games. The core idea involves

interpreting decision-makers’ theoretical mixed strategies as cumulative distribution functions (CDFs). We

then use the Kolmogorov-Smirnov statistic to test whether empirically observed real-world location choices

can be interpreted as random draws from these CDFs. To operationalize our approach, we develop a

general-purpose computational toolbox applicable to any location game. Finally, to demonstrate the

practicality of our method, we apply this toolbox to analyze location choice patterns among pharmacies in

selected regions of Türkiye. We hope this study will inspire future research on strategic settings where pure

strategy Nash equilibria are not guaranteed, including Bertrand price competition and political elections.