Several prototypes of improved biomass cookstoves have been designed to enhance the energy performance in households. In this work, a repeatability statistical analysis of a designed and built biomass cookstove based on Top-lit Updraft (TLUD) gasification is presented. The variability of the cookstove was assessed through the analysis of variance (ANOVA). The stove was tested under a modified Water Boiling Test (WBT) protocol with cold and hot starts, for this aim, three experiments were carried out, and three observations of parameters in each phase were measured or calculated. The stove operated under controlled conditions: secondary/primary airflows ratio, combustion chamber, and fuel type (pine wood pellets). Furthermore, the air velocity (≤0.25 m/s) on the hood work surface allowed having a controlled environment. The energy efficiency (~31.4 ± 0.7%), power (~9.9 ± 1.1 kW), fuel consumption (~33.7 ± 3.7 g/min), and specific fuel consumption per unit of time (~134 ± 15.81 kJ/L/min) were the response variables analyzed. It was obtained that the two phases of each stage of the modified protocol are statistically repeatable in this experimental facility leading to reliable performance of the cookstove with a confidence level of 95%.