Net intensities: Accuracy improvement through a Bayesian perspective on the measuring strategy and their persistent lack of precision. An illustration

2,3-Diketopiperazine (2,3-piperazinedione), C4H6N2O2, crystallizes in the m onoclinic space group P2(1)/c with a = 5.941 (3), b = 10.080 (3), c = 8.282 (2) Angstrom, beta = 95.57 (3)degrees, V = 493.4 (5) Angstrom(3), Z = 4, D -x = 1.536 Mg m(-3) and M-r = 114.1. The six-membered ring adopts a skew-bo at conformation with Q = 0.467 (3) Angstrom, theta = 64.6(3)degrees and phi = 269.8 (4)degrees. The C-2 symmetry, which typifies the free molecule, is broken by the formation of two intermolecular N-H O bonds involving only o ne of the C=O groups of the 2,3-diketopiperazine molecule. The intensity da ta typical for a time series are summarized in traditional statistics by B = sigma(2)(B) = Sigma B-i and R = sigma(2)(R) = CR,, where B is the backgro und and R the raw intensity. Exploitation of the same data using Bayesian m ethodology leads to similar values for B and R, but the variances for these signals are significantly smaller. This reduction in variance is dictated by the length N of the time series. With 25 observations in each hkl time s eries we arrive at variances that are 25% of their classical values. So, a measuring strategy in which a single similar to 2% of the maximum Bragg int ensity found at lambda 0.71 Angstrom.