A high pressure automated lag time apparatus (HP-ALTA) was used for the investigation of the controversial memory effect in methane propane mixed gas hydrates. The instrument can apply a large number of linear cooling ramps to a small volume of sample water under an isobaric condition of up to 15 MPa and record the maximum achievable subcooling for each cooling ramp. Over a hundred nucleation events were recorded for each of the several superheating temperatures used for the dissociation of the gas hydrate in a sample. In total, four different sample cells were used, and the effect of heating time was also studied for two of the four sample cells. A difference between two stochastic nucleation probability distributions was systematically and unambiguously quantified in terms of the most probable difference in the maximum achievable subcoolings. The protocol offers by far the most statistically robust method of quantification of the magnitude of the memory effect in each sample. From the analysis of several thousands of nucleation events, the following conclusions were made: (1) Even though the nucleation phenomena were intrinsically stochastic, a clear bias was observed which supported the existence of the memory effect. In particular, a reduction in the most probable subcooling of at least 4 K was required for positive identification of the memory effect for one of the sample cells. (2) The reduction increased as the superheating temperature was lowered. (3) The magnitude of the memory effect varied substantially among the sample cells used. (4) No significant effect of the heating time was observed in the range studied.