The up and down method of Dixon and Mood (1948), is also called the Bruceton test in certain military fuze explosive experiment. The testing procedure of a typical Bruceton run can be described as follows. Starting from an initial dosage level, a 0-1 type experiment is performed based on this given dosage. If the response is 0 (non-explosive), we will increase dosage level by one unit in the next experiment; while if the response is 1 (explosive), we will decrease the dosage level by one unit for the next experiment. The process continues until certain fixed observations are obtained. A salient feature for this method is that the observations are concentrated around the dosage levels that will produce the median value for the response variable. Normally, a parametric family is assumed, e.g., probit or logit models are the most popular ones. The classical method essentially tried to estimate the parameters, and then to estimate 99% quantile, for instance, by extrapolation. In this talk, we first concern the data analysis and inference based on given data as they are, then investigate the way to suggest more meaningful experiment. Some further works related to fuze explosive-train are mentioned.