Gravitational wave emission is considered to be the driving force for the evolution of short-period cataclysmic binary stars , making them a potential test for the validity of General Relativity .
In spite of continuous refinements of the physical description , a 10 % mismatch exists between the theoretical minimum period ( P _ { turn } \simeq 70 min ) and the short-period cut-off ( P _ { min } \simeq 80 min ) observed in the period distribution for cataclysmic variable binaries .
A possible explanation for this mismatch was associated with the use of the Roche model .
We here present a systematic comparison between self-consistent , numerically constructed sequences of hydrostatic models of binary stars and Roche models of semi-detached binaries .
On the basis of our approach , we also derive a value for the minimum period of cataclysmic variable binaries .
The results obtained through the comparison indicate that the Roche model is indeed very good , with deviations from the numerical solution which are of a few percent at most .
Our results therefore suggest that additional sources of angular momentum loss or alternative explanations need to be considered in order to justify the mismatch .