The effective survey volume, *V*_{eff}(*M*_{G}), which for our worked example is only a function of each object’s estimated absolute magnitude *M*_{G}. The figure shows three regimes (thick black line) where different terms in ﹩{S}_{{ \mathcal C }}({\boldsymbol{q}})﹩ limit *V*_{eff}(*M*_{G}): for the most luminous objects (*M*_{G} < 11), *V*_{eff} is simply limited by the initial selection *ϖ* > 3 mas; for the least luminous objects in the volume, it is limited by the initial cut apparent magnitude, *G* < 20. In the intermediate regime, the volume is limited by the (subsequently) chosen cut in expected parallax S/N, ﹩{\left\langle \tfrac{\varpi }{{\sigma }_{\varpi }}\right\rangle }_{\min }﹩ (Equation (11)). For very demanding choices in ﹩{\left\langle \tfrac{\varpi }{{\sigma }_{\varpi }}\right\rangle }_{\min }﹩, this cut may dominate for all *M*_{G} (green line); if such a cut is omitted or very lenient (blue line), this regime may disappear. This figure can also serve to illustrate why volume-limited samples are generally very suboptimal: if we wanted to construct a volume-limited sample of WDs covering 7 < *M*_{G} < 15, it would have a volume of only *V*_{eff} = 10^{−3} kpc^{−3}, and we would have to discard 90% (99%) of the accessible sample members at *M*_{G} = 13 (10).