calc_rank_of_first_retained_disparity_between

calc_rank_of_first_retained_disparity_between(first: Union[HereditaryStratigraphicColumn, HereditaryStratigraphicSpecimen], second: Union[HereditaryStratigraphicColumn, HereditaryStratigraphicSpecimen], confidence_level: float = 0.95) → int | None

Determine upper bound on MRCA generation at given confidence.

How many depositions elapsed along the columns’ lines of descent before the first mismatching strata at the same rank between first and second?

Parameters

confidence_levelfloat, optional

With what probability should the true rank of the first disparity between first and second fall at or after the returned rank? Default 0.95.

Returns

int, optional

The number of depositions elapsed or None if no disparity (i.e., both columns have same number of strata deposited and the most recent stratum is common between first and second).

Notes

If no mismatching strata are found but first and second have different numbers of strata deposited, this method returns one greater than the lesser of the columns’ deposition counts.

The true rank of the first disparity with second is guaranteed to never be after the returned rank when confidence_level < 0.5.

If the differentia width and confidence level are configured such that one or more spurious differentia collisions is plausible, this method will never return None.

Consider two columns that compare identical at all common ranks. If these columns have few enough common ranks that all these collisions could plausibly be spurious then there is not yet enough to support common lineage between the columns so rank 0 must be selected as the estimated site of first retained disparity. If these columns have enough common ranks that all collisions could not plausibly be spurious then the largest rank that could not conceivably be spurious must be selected (and we are guaranteed such a rank exists). It is only when not even one collision could plausibly be spurious that strong enough evidence exists to conclude that there is no disparity.

For example, the method may return None with 64-bit differentia and 95% confidence level but will never return None with 1-bit differentia and 95% confidence level.