I have had that problem, and it bothered me also. For me it happens only when the podcast publisher changes the Episode URL for a podcast. The new URL makes Media Center think it is a new episode, so it downloads it as such. When this happens, the subject hasn't changed, so when the "new" episode gets stored, has the same name as the original, suffixed by a (1) or (2) etc.
At the time this decision was made to match on the URL, the designer worried that some podcasts would be published with the same subject, so that matching subject would cause some episodes not to be downloaded. On the other hand the URL is guaranteed to be unique. Detecting duplicates matching on the URL means it is guaranteed you will get all the episodes, but sometimes you will get duplicates.
I've personally dealt with the problem by looking at my complete list of podcasts (e.g. Downloaded Podcast playlist) and then deleting all the episodes with a (1) or (2) etc at the end of the subject name. This is a lousy kludge, I know, but it has worked for me for some years now.
When you get duplicates, do the duplicates all have different Episode URLs than the originals?
You mention that some other podcast programs don't have this problem. Should Media Center be detecting duplicates based on the Episode Name rather than the URL? Perhaps Media Center should be matching on the Episode Name and date. What do you think?