The Secret to Grading

Alon Levy has offered a helpful algorithm for grading:

First, you assign to each student a likability score between 0 and 100. This will bias the program in favor of students you like. If you want the grades to have a mean of m and a standard deviation of s, and a student’s likability score deviates from the mean likability by d standard deviations, then every paper the student submits should be drawn from a normal distribution (m + d/2 + k, (s/2)^2) where k is a constant for a given assignment that corrects for rounding grades higher than 100 down. The chance the students will find out that effort doesn’t pay off is minute; most of them make no effort anyway.

I'm going to do some monte carlo testing to see if it works.

See, this is why I awarded Alon the title Smartest Blogger in North America™.