I will definitely take a look. Anecdote but I’m familiar with one multi-year example of a small casual prediction market that seemed like a very good predictor and another one that I don’t have the data from that seemed effective as well over time. I’ve hypothesized why that might be the case but never came to firm conclusions.