Local multiplayer social party games are intended for social engagement purposes, but large skill differences between participants can cause problems of decreased player agency. Existing Dynamic Difficulty Adjustment (DDA) systems have largely been optimized for single-player environments, which cannot accommodate the complex interactivity and shifting power dynamics of asymmetrical multiplayer games. This paper introduces the use of a trajectory-aware multi-channel adaptive difficulty framework integrated into an asymmetrical local multiplayer racing game prototype, titled Drive Me Crazy. Player skill is continuously estimated online by combining discrete game event data (obstacle collisions, minigame success) with continuous trajectory telemetry data, specifically corner smoothness, lane jitter, and overspeeding ratios. Using these online skill estimates, a multi-channel controller dynamically adjusts spawn pressure, precision demands, penalty severity, and reward bias. Furthermore, a driver tenure-based rubber banding approach is introduced to actively combat the runaway leader problem, which is present by design in the game’s asymmetrical scoring model. Preliminary baseline testing via a controlled A/B study (N = 20) confirms the framework’s effectiveness. The system reduced the mean final score spread significantly from 1675.00 ± 672.16 to 911.00 ± 385.99 (p = 0.0210) and increased lead changes from 0.40 to 2.40 per session, while crucially maintaining a stable natural driver swap frequency (p = 0.7848). Overall, the proposed framework proves to be an effective, non-intrusive tool for utilizing continuous behavioral telemetry data to ensure balanced competition and sustained social engagement in mixed-skill multiplayer environments.
Kovács et al. (Sat,) studied this question.