Two of the primary factors in the development of networked multiplayer computer games are network latency and network bandwidth. Reducing the effects of network latency helps maintain game-state fidelity, while reducing network bandwidth usage increases the scalability of the game to support more players. The current technique to address these issues is to have each player locally simulate remote objects (e.g. other players). This is known as dead reckoning. Provided the local simulations are accurate to within a given tolerance, dead reckoning reduces the amount of information required to be transmitted between players. This paper presents an extension to the recently proposed Hybrid Strategy Model (HSM) technique, known as the Dynamic Hybrid Strategy Model (DHSM). By dynamically switching between models of user behaviour, the DHSM attempts to improve the prediction capability of the local simulations, allowing them to stay within a given tolerance for a longer amount of time. This can lead to further reductions in the amount of information required to be transmitted. Presented results for the case of a simple first-person shooter (FPS) game demonstrate the validity of the DHSM approach over dead reckoning, leading to a reduction in the number of state update packets sent and indicating significant potential for network traffic reduction in various multiplayer games/simulations.