I tried a manual IVTC (simple 3:2 pulldown removal) on the sample, and although it seemed to reconstruct the original frames successfully and eliminate the combing, there were dupe frames appearing every so often. It's possible that there is a more complex pulldown pattern going on here.