I've just completed a package for generating numbers from the paraNormal distribution. This is the distribution of numbers that you are currently thinking of. Sample usage: > rParaNorm(1) [1] 3 - which is what I was thinking of when I pressed 'Return'. for generating multiple numbers, a confidence measure is returned that is a function of your uncertainty in your future choices: [here I'm thinking of 5, then 7, but then I'm not sure] > rParaNorm(4) [1] 5 7 1 5 Confidences: [1] 100% 98% 2% 1% Code for the quantiles of this distribution can easily be derived, but it is worth noting that the density is a function of the person running the code. Package download is here: http://tinyurl.com/7xet5 Baz