The HRV signals of 16 HCM patients are analyzed. Eight of them died or had a history of aborted sudden cardiac death, forming the high-risk group. The other eight patients form the low-risk group. Stationarity analysis is applied in order to avoid nonstationarities and to reject those data from further analysis. The surrogate data method is used to test nonlinear determinism on the RR series. Correlation dimension is calculated for all of the signals. An exponential fit to obtain the Dc is introduced, and a new index, Dck, derived from the proposed methodology is presented. This new phenomenological index will have the ability to stratify HCM patients with low and high risk of sudden cardiac death.