Re: [math-fun] NYTimes: How to Fix Our Math Education

I once took an epidemiology course from Olli Miettinen at Harvard School of Public Health, where tests were not only true-false +-1, but each question also had a space where you were supposed to enter how sure you are that your answer is right -- a number between 0 and 1.

An interesting idea would be a test where your certainty must be in the *open* interval (0,1). This way, you can't attain the supremum score (100%), but can get arbitrarily close to it. Also, it means that no-one can be confident of having the highest score. (I would probably write tanh(BB(BB(99))) to maximise my expected score.)