Extended Data Fig. 5: Comparison of prediction accuracies of the statistical model and 1D simulations.

a, b, Predictions for the areal density (a) and neutron yield (b) including the second phase of the Optimization Campaign. The predictions from the statistical model (blue) remain accurate for implosions from both phases of the Optimization Campaign, whereas the corresponding simulated 1D quantities from LILAC (red) remain inaccurate and overpredict potential increases in performance. Horizontal error bars and centre values represent one standard deviation and mean, respectively, for 500 draws from the posterior distribution of the statistical model. Vertical error bars represent one standard deviation for the neutron yield and areal density detectors. The dashed line is the y = x line on which the data points would lie if the prediction perfectly matched the measurement.