I have some unproven factors for predicting certain popular low-scoring questions: https://meta.stackoverflow.com/questions/373412/require-confirmation-before-a

I have some unproven factors for predicting certain popular low-scoring questions: https://meta.stackoverflow.com/questions/373412/require-confirmation-before-allowing-mostly-image-or-mostly-link-posts https://meta.stackoverflow.com/questions/373412/require-conf... A lot of questions that should have a very low score get fixed by people other than the original author. This may make your modeling more difficult because not only are there multiple authors, but the expected score of a question is not time-invariant. In fact, the final score of a question is a path dependent composite score of potentially several intermediate formulations.