Scalability is an important issue in the design of networked virtual environments (NVEs). In order to achieve scalability, it is essential to minimise the network traffic required to maintain overall consistency in the NVE. A popular method of achieving this is via entity behaviour prediction mechanisms, such as dead reckoning and the hybrid strategy model (HSM). Typically, the performance of such mechanisms is rated by the number of network packets they generate. However, it is also important that their impact on overall consistency is investigated. Absolute consistency is the degree to which different views of a NVE on remote hosts correspond. In previous work, it was shown that the use of a spatial threshold with dead reckoning can result in unbounded local absolute consistency. A solution that employed a time-space error threshold measure was shown to remedy this issue. In this paper, the scope of the time-space measure is extended to include the HSM. It is shown how the HSM can also result in unbounded local absolute inconsistency. A solution that once again incorporates the time-space threshold is examined. However, this approach results in a significant increase in network traffic. To resolve this, a novel extension to the HSM algorithm is presented, which is demonstrated to reduce network traffic, whilst still maintaining a low level of local absolute inconsistency.