Bayeswatch 7: Wildfire — LessWrong