Impact of the pocket effect photometric nonlinearity affecting ZTF data since November 2019. Top panel: Magnitude bias caused by the pocket effect (model) as a function of (input) magnitude. Lower panels: Difference between SN Ia parameters recovered from realistic light curve simulations affected by the pocket effect as a function of the input parameter (truth). Top to bottom: SN Ia peak magnitude in b band (from x₀), the SN Ia stretch, color, and Hubble residual (non-standardized). Orange markers (errors) show the median (nmad) difference per bin of input parameters. The horizontal orange lines show zero. The histograms on the right display parameter pulls ((sim.-truth) or error), which should follow a 𝒩(0, 1) distribution (dashed gray). Best fit normal distributions on pulls are shown in orange. In the m_b panel, the top gray ticks show the redshift corresponding to the peak magnitude of a typical M_B = −19.3 SN Ia.