Would Chuck Norris certainly win the Hunger Games? Simulating the result reliability of Battle Royale games through agent-based models

  • Hannes Rosenbusch (Contributor)
  • Jonas Röttger (Contributor)
  • David Rosenbusch (Contributor)




The amount of random chance involved in multiplayer competitions varies depending on the structure of the competition. A hugely popular mode of competition is the Battle Royale, in which a group of participants semi-randomly encounters and battles each other until only one player is left. We utilized computer simulations to investigate how strongly the outcomes of different Battle Royale formats reflect the actual distribution of skill among the players. Further, we investigated which game features affect the congruency between skills and results.


We build agent-based models in NETLOGO to simulate the outcomes of different Battle Royale formats.


Our results consistently forwarded that outcomes from Battle Royale are only weakly aligned with actual player skills.


Our findings suggest that many forms of Battle Royale involve substantially more luck than other sports or e-sports competitions. Further, game developers should note that introducing new game features can entail both higher and lower involvement of skill over luck, and that these features can be tuned to obtain a desired level of uncertainty.
Date made available2019
Date of data production11 Sept 2019

Cite this