At the beginning of an exam, you try to predict whether each problem is easy or difficult $($D $=+$ if it is difficult and $$ if it is easy$).$ Let us assume that you use two observable problem attributes $($predicates$)$:

$$ The text length L of the problem $($L $=1$ if it is long, $0$ otherwise$)$

$$ The amount M of math in the text $($M $=1$ if there is a lot of math, $0$ otherwise$)$