When conducting building performance simulation (BPS) at the early design stage, potential impact on design decisions is the greatest. However, the reliability of evaluations resulting from early design stage BPS models may suffer given that building attributes likely to be influential on a neighbourhood project's performance are often undecided. In this paper, we investigated the risk of making incorrect early design decisions by comparing performance estimations of daylight potential (sDA) and energy demand for heating/cooling of a range of neighbourhood massing schemes at low and high level of design development. The approach consisted of isolating and measuring the risk of performance loss (i.e. of being wrong in "ranking" the neighbourhood's overall daylight or energy performance) due to unknown building facade attributes. Cooling demand evaluations were found to be most reliable (92% cases resulting in low risk of performance loss due to unknown facade attributes). Spatial Daylight Autonomy (sDA) and heating demand based assessments resulted in lower reliability (78%, 85% low risk cases respectively).