The original complaint document are available here.
The news reported about "verbatim, note-for-note copying".
There were at least two independent musicologists both of whom argued that this case is obviously an infringement.
These opinions mention the 39 coinciding notes out of 64 total notes - taken from the original complaint. This ratio was a key point for both of the independent experts judging this case to be an infringement indeed.
My remark on this:
The two independent musicologists did not point out that the compared 16 bars consist of 8 phrases, most of which are close variants. These variants have two major types A and B. The sequence of these variants: ABBB ABBB.
Now let's think of this:
There are huge amount of songs with the same progression of four chords repeated long and where these chord are based on the same simple bassline of four looped notes. If we compare only these bass notes, then we can obtain another 16 bar or even much longer sections where (most of) the notes coincide. We know that this is a cheat and will not convince us about song level substantial similarity. Those four bass notes can be considered as one cycle, then consider the repetitions for a certain extent (weight).
Considering the repetition the case still shows a strong similarity. See results in the "similarity test algoritm".