Good stuff, but the thought occurs to me that what you really want to know is when the user has stopped learning - i.e., when proficiency stops increasing as a

Good stuff, but the thought occurs to me that what you really want to know is when the user has stopped learning - i.e., when proficiency stops increasing as a result of doing more problems - and how much each individual problem increases proficiency. But that would undoubtedly be more complicated.