Comment by eli

I question whether relative change in popular sentiment is a good way to evaluate a prediction.