Thermal building simulation currently uses Typical Meteorological Year (TMY) data to guide the design decision-making process or for compliance with energy standards. TMY data usually excludes extremes and in many cases are gathered from microclimatic contexts that are not sufficiently representative of the project sites (e.g., airports), adding uncertainty in the analyses. To enable a quantification of uncertainty due to weather by exploring a wide variety of atypical weather conditions, the authors have previously proposed synthetic weather data for building simulation. This is a suite of weather time series, generated from typical weather data that includes heat waves and atypical peak temperatures. In this paper, we used this synthetic weather to examine the effect of considering atypical conditions on design decisions. We also compared the impact of including ‘city-modified’ weather data on retrofit decisions using urban microclimate simulation. We found that it may not be viable to pre-select a subset of weather data for all buildings at a given location. Rather, multiple weather data sets may be simulated based on the design strategies and performance criteria of importance. In other words, an extreme condition/year for one building isn’t necessarily the same for another. For example, in the case study presented, heat island effect was found to be a likely hindrance to night time cooling. This paper informs the debate on the necessity of expanding the current energy building analyses to a broader consideration of weather variability and more realistic urban microclimate characterization.