Paul, have you considered re-implementing the orange-names experiment as a gradient of gray-to-orange, to avoid a tribal sense of "grays versus oranges"? Or is

Paul, have you considered re-implementing the orange-names experiment as a gradient of gray-to-orange, to avoid a tribal sense of "grays versus oranges"? Or is that not-obviously-scalable?