Compilations of measured energy savings have shown that engineering calculations do not always correlate well with actual performance. One important difference between engineering calculations and real world performance is the effect of weather. Energy service companies, whose profits are a function of energy savings, and building energy researchers have developed weather-normalization formulas or techniques. True tests to determine the adequacy of these methods, however, require careful control of other determinants of building energy use. This paper describes results obtained by using a building energy simulation tool to evaluate some of these methods for commercial buildings. Degree-day-based normalization techniques designed to account for the effects of weather on commercial building energy use are identified. The normalization techniques are compared using the results of DOE-2 simulations for two office building prototypes using many years of actual weather data for a single location. We conclude that, for the prototypes and location examined, the techniques performed reasonably well, and the sophisticated techniques did not perform noticeably better than the simpler ones.