Thanks for the interest in the problem, Gail!

The table has about 160,000 rows right now; the number of rows will vary on any given day, as they're added and deleted when users and projects are created and removed.

I'll be out of the office until Monday, so I won't be able to provide an actual execution plan until then; I really should have done that initially. Bad omission on my part, but I'll supply it as soon as I can. The actual plan for a query looking for a Name with k_id BETWEEN 20000 and 20002, or a k_id with Name = @Name ends up doing an index seek on k_id or Name, but the estimated number of rows for the seek always has a single row, whereas the actual number of rows will end up being whatever number of rows is actually returned by the query. Seeing the actual plan will undoubtedly be more useful than just this clunky summary, I'd imagine 😛