Marvin's changes tool reports on five shot categorizations:
Shots that appear in the "new" EDL, which were not in the "old" EDL, are considered new. This typically indicates that a new shot was added to the cut.
Shots that appeared in the "old" EDL, but that are not in the "new" EDL, are considered an omit. This typically indicates that an existing shot was removed from the cut. However, it could also mean that the marker containing the shot id was accidentally deleted or trimmed off the clip. Further investigation is required to confirm.
Marvin analyzes multiple aspects of every cut in an EDL. The changes tool primarily compares the cut in and cut out timecodes and considers a shot to be a change if these values changed inside of the handles. It may also indicate that the shot's marker was modified or that the shot has moved acts/reels. Further investigation is required to determine how the change impacts any work.
Shots where the cut in and cut out timecode values change outside of the handles are considered a rescan. This may also indicate that the clip name or tape name was changed, typically due to a take or camera angle switch. Further investigation is required to determine how the change impacts any work.
Marvin can only compare two cuts using the provided handles in the DCUT file (from the shots tool). If the "old" sequence contains changes since the shots were turned over, it's possible that a rescan may be missed. For example, consider a shot that was already slipped 8 frames in a previous cut with 10 frame handles. If it was slipped again in the same direction by 3 frames in the "new" sequence, now putting it outside of handles, then Marvin would not report it as a rescan during a default comparison because it is unaware of the cumulative effect of the previous changes. To prevent this from happening, you should update the handles for all affected shots prior to exporting the DCUT file and importing it into the changes tool.
Shots that did not change in any way during the comparison are considered unchanged.