Valve's upcoming MOBA-hero shooter, Deadlock, recently overhauled its matchmaking system, employing an algorithm surprisingly sourced from ChatGPT. This revelation, shared by Valve engineer Fletcher Dunn on X (formerly Twitter), highlights the evolving role of AI in game development.
ChatGPT's Unexpected Contribution to Deadlock's Matchmaking
Dunn's tweets detail how ChatGPT suggested the Hungarian algorithm as a solution to Deadlock's matchmaking challenges. This followed significant player criticism of the previous MMR system, with complaints focusing on unevenly skilled teams and consistently facing superior opponents. Players voiced frustration on Reddit, describing experiences where teammates appeared significantly less experienced than their adversaries.
The Deadlock team acknowledged these concerns, previously announcing a complete matchmaking system rewrite on their Discord server. Dunn's use of ChatGPT, therefore, represents a successful application of AI in resolving this critical issue.
Dunn's enthusiastic embrace of ChatGPT, even maintaining a dedicated browser tab for it, underscores the tool's growing utility. He's actively sharing his "ChatGPT wins," aiming to showcase the AI's capabilities to skeptics. However, he also acknowledges a potential downside: the replacement of human interaction and collaboration, raising ethical questions about AI's role in software development. This sentiment was echoed by a social media user who pointed out concerns about AI replacing programmers.
The Hungarian algorithm, as explained by the article, addresses the problem of finding optimal matches in a bipartite system (two parties involved), particularly useful in scenarios where only one side (e.g., player preferences) needs to be considered.
Despite the improved algorithm, some players remain unconvinced, expressing continued dissatisfaction with Deadlock's matchmaking performance and criticizing Dunn's reliance on ChatGPT.
Game8, however, maintains a positive outlook on Deadlock's future, anticipating a successful launch. Further details on the game and its playtest can be found in a linked article (not included here).