[dang]

While I wish your argument were correct, it's not. The flaw is in the inference voted for → was intellectually interested.

A model that appears to fit the data [1] is that multiple kinds of interest drive voting, and intellectual interest is not the most powerful one. This is the weakness of the story-voting mechanism. It is why, on a site that values intellectual interest, the vote signal must pass through other filters, like flagging and moderation.

1. Maybe we can test this someday.