2014/03/24 by Luciano Gualà, Gualà, Luciano, Stefano Leucci +3 · 1 voice · 1 citation
Computer Science · Social Sciences · #Artificial Intelligence in Games #Digital Games and Media #Peer-to-Peer Network Technologies #cs.CC
paper · pdf · doi:10.48550/arxiv.1403.5830
openalex publication_date 2014/03/24 · openalex created_date 2022/10/01 · openalex updated_date 2026/07/28
The twentieth century has seen the rise of a new type of video games targeted at a mass audience of "casual" gamers. Many of these games require the player to swap items in order to form matches of three and are collectively known as tile-matching match-three games. Among these, the most influential one is arguably Bejeweled in which the matched items (gems) pop and the above gems fall in their place. Bejeweled has been ported to many different platforms and influenced an incredible number of similar games. Very recently one of them, named Candy Crush Saga enjoyed a huge popularity and quickly went viral on social networks. We generalize this kind of games by only parameterizing the size of the board, while all the other elements (such as the rules or the number of gems) remain unchanged. Then, we prove that answering many natural questions regarding such games is actually \NP-Hard. These questions include determining if the player can reach a certain score, play for a certain number of turns, and others. We also \hrefhttp://candycrush.isnphard.comprovide a playable web-based implementation of our reduction.