I judge expertise by the ability to answer questions or solve problems.
The statistics you request are altogether useless for that purpose, since they are mere aggregates of how others (mostly random and unknown voters) perceive a contributor's answers. In order to evaluate a contributors' expertise, you need to (1) have a good understanding of the answers, and (2) compare them with reliable sources you either know already or are capable of ascertaining. Sheer numbers are not an adequate substitute for what it takes to be able to conduct that evaluation.
I believe a person whose reputation is built by primarily by asking questions only is not an expert in philosophy.
That assumption is unreliable. It fails to distinguish between questions where asker's research & knowledge are evident, and those which happened to make it to the Hot Network Questions or were otherwise about something trendy at some point in time. Furthermore, a user's question could be so advanced or profound that very few (if any) users might be able to grasp it, let alone match or top the asker's expertise.
what percentage of a users asked questions have been marked as answered. This tells me how sincere that user is and whether they are acquiring knowledge from their questions.
That percentage is not indicative of asker's learning. It fails to distinguish between genuine learners, and biased askers who accept the answer --whether accurate or not-- that matches their preconceived notions. The latter approach is the opposite of sincerity.
Moreover, you implicitly assume that if an asker learned from an answer, that answer inevitably will be marked as Accepted. That presumption ignores the very likely scenario that an asker might learn from multiple answers, in part because answers hardly ever are mutually exclusive.
The usage of statistics as a tool for evaluating users' expertise entails too many flaws.
