De-duplicating incrementally refreshed data

Obviously, I can extend the window, but I'm finding this doesn't really solve the problem 

You may be able to find a balance there that works good enough.

In any case, you are hitting on the true core of the issue here - incremental refresh is exactly that - incremental.  It expects data to be immutable once it has been added.  What you expect to get (differential refresh) is not what incremental refresh is about, and is not what it can deliver.

There is a small remedy in using the "detect data changes"  feature but it comes with a gigantic price tag. I'll let you find out what that is.

What you want to do is continue with your incremental refresh, and  occasionally (once a month for example) force a reprocess of a couple of the "older" partitions as well.  Doesn't even have to be a complete flush and fill of the dataset , you can use the XMLA endpoint to target individual partitions.