Bamber and Aspinall 2013 published a formal expert elicitation on the future of our two ice sheets, and found a large spread between individual expert answers. It has been argued by them, and others that there are outlier experts in the data, and that they drive the long-tails of the pooled uncertainty distribution. I wonder if that is really true or whether it could be a visual artifact from drawing samples from a long tailed distribution. When you plot samples from a long tailed distribution then of course the tail values will appear to stand out, but that does not necessarily make them outliers. The question is really whether there are two groups of experts optimists vs. pessimists or whether they all have been sampled from a continuum of experts.