Near-extreme outside dry- and wet-bulb temperatures are essential for the design of building HVAC systems. Coincident design temperatures currently defined by ASHRAE and CIBSE may not fully allow engineers to design an HVAC system that can match a risk level desired for the problem at hand. A rational method is therefore presented in this paper for the proper selection of coincident climate design conditions, which is based on the theory of probability. An average statistic risk level has been proved to be approximately equal to the annual percentile of coincident dry- and wet-bulb temperatures for HVAC systems whose thermal time constant is less than I h. An algorithm is also developed to determine the contours of equal risk. Representative coincident design dry- and wet-bulb temperatures are obtained through statistic analysis of the 15 years of weather records in Hong Kong. These rationally derived outside design temperatures are compared with the coincident dry- and wet-bulb temperatures currently defined by ASHRAE and CIBSE. (c) 2004 Elsevier B.V. All rights reserved.