As part of the following commit: https://github.com/pydata/pandas/commit/fbb110247ea81e980bf9265cc1b0b26c8eb549be#L5L799 Series.std and Series.var no longer pass along ddof to the nanops.nanvar function. Also since ddof defaults to 1 in pandas and 0 in numpy. ``` python arr = np.arange(10) s = Series(arr) # False np.std(s) == np.std(arr) # True np.std(s) == np.std(arr, ddof=1) ``` Which I don't know if it's a bug but it's non-intuitive.