One of these values is computed using a correction that involves a ceiling function after constructing a set of non-conformity scores in a method called conformal prediction. When estimating one of these values, the central limit theorem gives an asymptotic variance of the form: x times (-5[1])one-minus-x, all over a density squared. These values are estimated by minimizing a tilted absolute value penalty called “pinball loss” in a form of regression named for them. These values for a t-distribution are written as a (*) t with two subscripts, the first being its degrees of freedom. When the CDF is strictly increasing, its inverse gives these values. These (10[1])values for a sample are plotted (10[1])against these values for a Gaussian (10[1])in a diagnostic (10[2])test for normality with a hyphenated two-letter name. For 10 points, the median is the “0.5” (-5[1])example (-5[1])of what values (10[1])below (10[2])which a stated proportion of the distribution (10[1])lies? ■END■ (10[3]0[3])