| A | B (Calculation)      | B (Value) |
|---|----------------------|-----------|
| 1 |                      |           |
| 2 |                      |           |
| 3 |                      |           |
| 4 | =SUM(A1:A4)/4        | 2.5       |
| 5 | =(1/4)*A5 + (3/4)*B4 | 3.125     |
| 6 | =(1/4)*A6 + (3/4)*B5 | 3.84375   |
| 7 | =(1/4)*A7 + (3/4)*B6 | 4.6328125 |

# represent row numbers
r = np.arange(len(df))
# create same sized array filled with 3 (the fourth index)
o = np.empty_like(r)
o.fill(3)
# take pair wise maximum.  will look like [3 3 3 3 4 5 6]
# this allows me to take the mean of the first 4 and the
# existing values elsewehre
np.stack([r, o]).max(0)
# perform the groupby then ewma
df.A.groupby(np.stack([r, o]).max(0)).mean().ewm(adjust=False, alpha=.25).mean()

3    2.500000
4    3.125000
5    3.843750
6    4.632812
Name: A, dtype: float64