Period bouncers are cataclysmic variables (CVs) that have evolved past their orbital period minimum. The strong disagreement between theory and observations of the relative fraction of period bouncers is a severe shortcoming in the understanding of CV evolution.

We test the implications of the hypothesis that magnetic braking (MB), which is suggested to be an additional angular momentum loss (AML) mechanism for CVs below the period gap (P_orb ≲ 120 min), weakens around their period minimum.

We computed the evolution of CV donors below the period gap using the MESA code, assuming that the evolution of the system is driven by AML due to gravitational wave radiation (GWR) and MB. We parametrised the MB strength as AML_MB = κAML_GWR. We computed two qualitatively different sets of models, one in which κ is a constant and another in which κ depends on stellar parameters in such a way that the value of κ decreases as the CV approaches the period minimum (P_orb ≈ 80 min), beyond which κ ≈ 0.

We find that two crucial effects drive the latter set of models. (1) A decrease in κ as CVs approach the period minimum stalls their evolution so that they spend a long time in the observed period minimum spike (80 ≲ P_orb/min ≲ 86). Here, they become difficult to distinguish from pre-bounce systems in the spike. (2) A strong decrease in the mass-transfer rate makes them virtually undetectable as they evolve further. So, the CV stalls around the period minimum and then “disappears”. This reduces the number of detectable bouncers. Physical processes, such as dynamo action, white dwarf magnetism, and dead zones, may cause such a weakening of MB at short orbital periods.

The weakening MB formalism provides a possible solution to the problem of the dearth of detectable period bouncers in CV observational surveys.