Anfonwch hwn fel neges destun: Robust versus optimal strategies for two-alternative forced choice tasks