That's a bug.

We're using an english language analyzer that pulls out root words and searches for those, but we shouldn't prioritize that for exact matches.