Simulated Hα emission line as a function of stellar mass for the EDS-SF_sim selection is shown for the lowest redshift range: 0.9 ≤ z_meas ≤ 1.2. Top panels present the relation at the flux limit of F(Hα)≥2 × 10⁻¹⁷ erg s⁻¹ cm⁻² and with no cut in z_prob, whilst the bottom panels present the improvement obtained with our fiducial selection that also includes a cut in redshift probability z_prob ≥ 0.4. The left panels show the distributions of the different subsamples: the ℬℳ sample (grey points; Eq. (7)), the observationally selected galaxies with accurate redshifts (green points), and the redshift contaminants (red points). Note that the green and red points together constitute the observationally selected sample, as defined by the denominator of Eq. (6). Black and cyan contours and dashed lines show the ground-truth distributions for the full EDS parent sample (H_E ≤ 26) and a subsample selected with F(Hα)≥2 × 10⁻¹⁷ erg s⁻¹ cm⁻², respectively. These contours serve as a reference to illustrate the intrinsic (ground-truth) distributions, allowing for a qualitative comparison with the observation-like measurements, and to highlight the intrinsic selection bias introduced by the flux limit, which would persist even without redshift measurement errors. The right panels show galaxy number densities per deg² in bins of stellar mass and Hα flux, with 𝒮ℛ (top number) and 𝒞 (bottom number) indicated. Numbers are colour-coded: black-blue for reliable bins (𝒮ℛ > 45%, 𝒞 < 20%) and red and magenta for others. The numbers reported above each mass bin represent the 𝒮ℛ (top number) and 𝒞 (bottom number) for the entire mass bin.