I'm not getting a very good match with the Fluent data, but I'm more worried about k starting to decrease much sooner