In the stats windows „Repeated segments“ is exactly what it is meant to be.
In the filter „Repeated segments“ includes all segments CT is able to fill in automatically.
Example: „Figure 1“ and „Figure 2“ (or even whole sentences with only a changing number) are repetitions in this sense, so CT filters them as „Repeated segments“ and is able to fill it automatically.
Note: This works better than in any other tool, by the way.
I really appreciate this filter, and it would be absolutely wrong to include the repeated segments from the filter function to the repeated segments of the stats. And it would be wrong the handle the filter in a stricter way.
But wouldn't it be more logical to rename the filter „(Pseudo-) Repeated segments“?
This is a good remark. The "Repeated segments" in the Filter menu may be renamed to "Repeated and propagated" segments as it includes the segments which are to be auto-propagated.