The conditioning of room air in summer is widely practiced. However many operators do not have an overall quantitative strategy for minimizing energy. Widespread practice is to simply use on-off, that is, have air conditioning-on when a room is occupied, and conditioning-off, when un-occupied. Here we apply the probabilistic Fr 13 risk framework (Chem. Eng. Sci. 152 (2016) 213-226) for the first time to take account of naturally occurring fluctuations in daily ambient peak temperature (T-o), occupancy rates (L-T) and room traffic inflow and outflows (n) in a typical (hotel) room fitted with heat-attenuating curtains in the hot and dry clime of South Eastern Australia (latitude -37.819708, longitude 144.959936). We use this to synthesise an extended steady-state unit-operations model and assess the two energy strategies for conditioning of the room air to a widely used auto-set bulk temperature of 22 degrees C. It was assumed the impact on overall energy demand from room lighting and refrigerator and occupant metabolism was negligible. We show incident radiative energy will be significant, and using historical ambient temperature fluctuations and occupancy rates, that adoption of the alternative on-only continuous conditioning would use less energy in 93.6% of summer days, based on an overall commercially viable L-T = 75% with the minimum possible room traffic flow of n = 1. Practically, this equates to six (6) only failures of this alternative energy strategy in the 90 days of summer. Importantly, the probabilistic Fr 13 framework permits a practical design for larger-scale quantitative testing of the proposed strategy. A preliminary 10-day 'proof-of-concept' trial showed energy savings of 18.9% (AUD $2.23 per suite (paired-room) per day) with a concomitant 20.7% reduction in GHG. Findings will be of immediate interest and benefit to operators and managers of large buildings that rely on conditioning of dry air in summer. Crown Copyright (C) 2018 Published by Elsevier Ltd. All rights reserved.