When modelling the environmental and economic aspects of meeting a given heat and power demand with a combination of combined heat and power (CHP) and grid power, it is common to use a coarse temporal precision such as 1-h demand blocks in heat and power demand data. This may be appropriate for larger applications where demand is reasonably smooth, but becomes questionable for applications where demand exhibits substantial volatility such. as for a single residential dwelling-an important potential market for the commercialisation of small-scale fuel cells and other micro-CHP. Choice of temporal precision is also influenced by the relative ease in obtaining coarse data, their compatibility with available energy price data, and avoidance of computational overheads when data sets expand. The thesis of this paper is that use of such coarse temporal precision leads to averaging effects that result in misleading environmental and economic outcomes for cost-optimal micro-CHP systems. Much finer temporal precision is required to capture adequately the specific characteristics of residential energy demand and the technical qualities of solid oxide fuel cell and stirling engine micro-CHP systems. This thesis is generally supported by the results of analysis, which shows that in some cases optimal design generation capacity of the CHP system is reduced by more than half between analyses using 1-h precision and 5-min precision energy demand data. When optimal dispatch of given generator and boiler capacities is considered, the quantities of energy delivered by the various components of the energy provision system (i.e. generation from CHP, heat from CHP, heat from an additional boiler, electricity from grid) varied by up to 40% between precisions analysed. Total CO2 emissions reduction is overestimated by up to 40% by the analyses completed using coarse demand data for a given micro-CHP generator capacity. The economic difference is also significant at up to 8% of lifetime costs for a given micro-CHP generator capacity. © 2005 Elsevier Ltd. All rights reserved.