Nonlinear power spectrum at various redshifts compared to linear theory from CLASS (Lesgourgues 2011). The use of the forwardscaling method in the N-body gauge (Fidler et al. 2019) allows the N-body simulation to agree with linear theory at all redshifts including the effects from weak field general relativity, radiation, and massive neutrinos. The narrow ±5% band is plotted using a linear scale, making those fluctuations more visible and allowing also for negative values. Fluctuations at low k are statistical and reflect the sample variance of the particular realisation. At high k the enhanced power is due to nonlinear growth of structure not captured by the linear theory. This nonlinear contribution also includes the slight −2% dip at k ≈ 10⁻¹ h Mpc⁻¹, corresponding precisely to the first BAO peak in the power spectrum. The forward-scaling method used for the Flagship simulation guarantees a match to the linear theory of CLASS at all redshifts, which was not possible to achieve with the traditional back-scaling technique (which only guarantees this at z = 0 in the presence of massive neutrinos and radiation). The bottom panel compares the power spectra to the Euclid Emulator nonlinear power spectrum (Euclid Collaboration: Knabenhans et al. 2021).