I have a series of oil well point that are located in England. They are cast on the ED1950 datum using the International 1924 reference ellipsoid but are not projected. I need to transform the values to WGS84. So, I used the ED_1950_To_WGS_1984_18 datum transformation to transform the ED1950 point file to WGS84. After the transformation, the new points are shifted about 100m to the south and west of the original. This amount of shift seems reasonable given that the location of the ED1950's center is in Potsdam and the oil well points are in England.
Here is my question. I know that the International 1924 reference ellipsoid is larger along the semi-major axis than the GRS80 reference ellipsoid. The semi-minor axis of International 1924 reference ellipsoid is also greater than the semi-minor axis of GRS80 reference ellipsoid. Thus, one would expect that longitude values for ED1950 should be further west when matched to their corresponding WGS84 longitude values and ED1950 latitude values are further south than WGS84 latitude values. If this is correct, then my question is why doesn't the ED_1950_To_WGS_1984_18 transformation shift the ED1950 coordinates to the north and east of the original rather than south and west as shown in the screen capture?
I don't know a lot about the European datums, but going from an old local to a recent world datum will offer many variations. Generally when going from a larger datum to a small datum the positions will shift towards 0,0. This may be more coincidental, I haven't worked out the math myself. Also, the terms GRS80 and WGS84 are not interchangeable, though the variations are insignificant for most applications.
You can look up the details of the 7-parameter transformation in your documentation if you want some hard numbers. The XYZ shift for your transformation is this -89.5, -93.8, -123.1.