The programmability of FrACT₁₀ using HTML messages allows to code an “ideal observer” (→blog entry). Here is a simple one which responds according to a psychometric function of given flex (flex = inflection point ≙ observer acuity), slope and lapse rate. Guessing probability (= chance level) depends on the number of possible alternatives. [Code: sim.js.]
Note: the lower the slope, the more the outcome is likely to deviate from the correct acuity. Well seen also in the “All Trials Plot”. Shading around final estimate represents the bootstrapped CI95.