The dynamic-clamp technique has been recognized and used by electrophysiologists for over 15 years. Nevertheless, only a small number of papers have been written focusing on the performance and reliability of this protocol and how the accuracy of a dynamic-clamp system can be assessed. Here we review the published literature to date, focusing on how experimental, computational, and algorithmic factors contribute to the reliability of the dynamic-clamp protocol. Several of these papers point towards a common and technologically realizable solution – the need for dynamic-clamp systems that run at computational rates much faster (100–200 kHz) than currently available. At present, dynamic-clamp rates are limited by the use of desktop computers and rationalized by the kinetics of the model simulated. Recent results show that faster and lower latency systems would result in a greater range of conductances that could be utilized, improved stability, and more accurate real-time model simulations.